University of Southern California Dissertations and Theses

Levels of interest: the effects of teachers' unions on Congress, statehouses, and schools

(USC Thesis Other)

Levels of interest: the effects of teachers' unions on Congress, statehouses, and schools

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Running Head: The Effects of Teachers' Unions on Congress, Statehouses, and Schools 1

Levels of Interest:
The Effects of Teachers' Unions on Congress, Statehouses, and Schools

by

Bradley D. Marianno

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(EDUCATION)

August 2018

Copyright 2018 Bradley D. Marianno
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 2

DEDICATION
To my immediate and extended family, who are some of best professional educators I know.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 3

ACKNOWLEDGEMENTS
There are several individuals that have played a decisive role in my education journey,
from my earliest experiences in elementary school to the advanced stages of my doctorate
degree. Without their patience, kindness, and encouragement, it is extremely unlikely that I
would have made it to the finish line, let alone even considered engaging in the journey in the
first place. The pages of this dissertation could be filled with important family, friends, and
colleagues whose life path crossed my own in meaningful ways. Here I acknowledge just a few
who made important and lasting contributions to my academic and personal character.
First, I say thank you to Katharine Strunk who fielded a phone call from an unknown
prospective student finishing a graduate program at Brigham Young University. You took a
chance on me and remained invested until the end. Your critical feedback, encouragement, and
concern for my professional development made an indelible imprint on my academic trajectory.
But very few students are also lucky enough to have an advisor that cares about them and their
family personally. You truly are one of a kind.
To Julie, Josh, and Gary—thank you for your willingness to advise and support me
through this dissertation. Your insights, comments, and suggestions have made me a better
scholar.
To my wife, Holly, and my daughter, Maddie, who took this journey with no
complaints—a thank you here does not even begin to compensate for the many hours of support,
encouragement, and love you provided. Holly, you tolerated my home office, proofread papers,
drove to vacation destinations while I worked in the passenger seat, and provided hours-worth of
meal preparation, house cleaning, and psychological counseling. Maddie, you were always game
to play, and were patient when I had to close the bedroom door to finish a paper. You both
remain the most important thing in my life.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 4

To Boy Scout Troop 516 in South Pasadena—we spent many weekends hiking and
camping in the mountains, deserts, and on the beaches of Southern California. These trips
rejuvenated me and made the rigors of a PhD program much more tolerable. Thank you for
letting me serve as your leader. You impacted my life more than I ever imagined.
To Tracey and Ayesha—you left a clear path for me to follow. Your encouragement,
guidance and hard work made my road to this dissertation much easier and more enjoyable.
More importantly, thank you for playing both the role of colleague and friend.
Finally, to my family by birth—You have been with me from the beginning. To my mom
and dad—I owe you more than I could ever repay. Mom, you taught me to choose to make every
day a good day and to keep my academic work in proper balance. Dad, you are my educator
hero. Thanks for cheering for me, talking education policy with me, and most importantly
demonstrating how to make family life and service the highest priority above all else. To Matt,
Eric, and Scott—the best brothers a guy can ask for—thank you for the running text message
stream that provided daily laughs and debates. Thanks for visiting and providing needed
distractions. Scott, thanks for being my last-minute proofreading historian. You guys are simply
the best.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 5

TABLE OF CONTENTS
DEDICATION…………………………………………………………………………………...2
ACKNOWLEDGEMENTS……………………………………………………………………..3
TABLE OF CONTENTS………………………………………………………………………..5
CHAPTER ONE- Teachers’ Unions in a Changing Policy and Political Context
A Brief Review of the Policy and Political Context Surrounding Teachers’ Unions………….7
Introduction to the Dissertation…………………………………………………………...........8
CHAPTER TWO- Down but Not Out: The National Education Association in Federal
Politics
Introduction…………………………………………………………………………………...12
Data…………………………………………………………………………………………...14
Methods……………………………………………………………………………………….16
Results..……………………………………………………………………………………….19
How has the Distribution of NEA Allies and Foes Changed Over Time? …………..........19
Who Did the NEA (and Their Opponents) Target with Contributions During the 2010,
2012, and 2014 Elections and to What Extent Did Their Favored Candidates Win?..........21
What Federal Legislation (e.g. School Choice) has the NEA Officially Supported and
Opposed from 2009-2016 and What are the Outcomes of this Respective Legislation?.....25
Conclusion……………………………………………………………………………………27
CHAPTER THREE- Preserving the Status Quo?: The Relationship between Teachers’
Union Strength and State-Level Agenda- and Policy-Setting Influence
Introduction…………………………………………………………………………………...30
Teachers’ Union Power in an Era of Institutional Change in Education Policymaking……...33
Data…………………………………………………………………………………………...41
Analytic Strategy……………………………………………………………………………...57
Results...………………………………………………………………………………………58
Conclusion……………………………………………………………………………………61

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 6

CHAPTER FOUR- Where Have All the Senior Teachers Gone?: Teacher Transfer
Provisions and Between-School Gaps in Teacher Experience and Student Achievement
Introduction…………………………………………………………………………………...65
Literature Review……………………………………………………………………………..69
The Debate Regarding Seniority-Based Teacher Transfer Provisions in CBAs…………..69
Prior Studies……………………………………………………………………………….71
Data…………………………………………………………………………………………...75
Analytic Strategy……………………………………………………………………………...85
Results..……………………………………………………………………………………….89
How Much Do Teacher Transfer Provisions Change Over Time?………………………..89
Are Changes in Transfer Provisions Associated with the Way Experienced Teachers are
Distributed Across Schools?………………………………………………………………90
Are Changes in Transfer Provisions Associated with Between-School Gaps in Student
Achievement? …………………………………………………………………………….91
Conclusion……………………………………………………………………………………94
CHAPTER FIVE- The Future of Teachers’ Unions in Public Education: Cross-Cutting
Themes from Three Papers on Teachers’ Union Influence….................................................98
REFERENCES ………………………………………………………………………………..103
FIGURES AND TABLES…………………………………………………………………….126
APPENDIX…………………………………………………………………………………….147

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 7

CHAPTER ONE- Teachers’ Unions in a Changing Policy and Political Context
Boasting a membership well over four million (American Federation of Teachers, 2016;
Sawchuk, 2016), and revenues that total over 500 million dollars (Foundation Center, 2016),
teachers’ unions are generally regarded as influential, well-funded associations that strive to
protect and serve the interests of teachers at the national, state, and local levels of government
(Bascia & Rottmann, 2011; Moe, 2011). Teachers’ unions primarily affect public education by
exercising influence over a wide swath of policy governing the most important school-based
input in determining short-term and long-term student outcomes—teachers (e.g. Rivkin,
Hanushek, & Kain, 2005; Rockoff, 2004; Chetty, Friedman and Rockoff, 2014). At the local
level, teachers’ unions negotiate collective bargaining agreements (CBAs) in the 43 states that
permit or require district administrators to bargain contracts with teachers’ unions (National
Council on Teacher Quality, 2016). These legally binding documents govern nearly every aspect
of teachers’ work, from the way they are evaluated, disciplined, and professionally developed, to
how they are assigned to schools and classrooms, and how they interface with parents and school
administration (e.g. Ballou, 2000a; Goldschmidt & Stuart, 1986; Hess & Loup, 2008). In the
states that permit collective bargaining, these agreements work in concert with state and federal
policies governing teachers’ work (e.g. teacher tenure, teacher due process procedures). In an
effort to impact new legislation, teachers’ unions also maintain election and lobbying presences,
donating nearly 200 million dollars since 2011 to advocate for teachers’ interests at all levels of
government (National Institute on Money in State Politics, 2016).
Nonetheless, now over 150 years removed from their early beginnings in the 1850’s,
teachers’ unions have just recently experienced some of the most significant state legislative and
court challenges in their history in public education. Since 2011, every state in the nation has
proposed, and in some cases, enacted reforms to unions’ rights to negotiate collective bargaining
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 8

agreements (CBAs) with local school districts (e.g. Wisconsin Budget Repair Bill; Freeman &
Han, 2012; Marianno, 2015). In a six-year span (2012 to 2017), six states enacted Right-to-Work
legislation prohibiting unions from collecting membership dues paid to support union
professional activities (only four states enacted such legislation in the 50 years prior to 2012).
The U.S. Supreme Court revisited union’s long-held right to collect membership dues for
collective bargaining and other union services (Friedrichs v. California Teachers’ Association;
Janus v. AFSCME), and state courts have considered reforms to tenure rights, seniority-based
layoff provisions, teacher discipline procedures, and evaluation policies (e.g. Doe v. Antioch,
Vergara v. the State of California; Wright v. New York; Forslund v. Minnesota). This is all to say
that the legal rights of teachers’ unions are changing, and quickly.
But even as state lawmakers reform codified protections for teachers’ unions, the small
but growing extant literature on unions’ impacts on the U.S. education system provides very few
definitive conclusions to guide policymakers’ judgements on whether to reduce, expand, or
preserve the role of unions in public education (see Cowen & Strunk, 2015, for a review). As the
reforms have outpaced the evidence, it is entirely possible, even probable, that new reforms are
steeped mostly in political rhetoric and anecdote and may overstate or also, possibly,
underestimate, the impact of teachers’ unions when determining how to alter union rights. This
dissertation seeks to add to the research literature and inform new policy proposals on teachers’
unions by answering the critical question: How do teachers’ unions influence congress,
statehouses, and schools?
This dissertation, is intended to tackle this broader question through three different
empirical studies, each exploring a largely unaddressed or unanswered issue in the extant
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 9

literature on teachers’ unions. The first two papers explore the influence of teachers’ unions on
Congress and state policy-setting.
1. Congress—How has the influence of teachers’ unions changed over time in
federal elections and in congressional policy-setting?
2. States—What is the relationship between state teachers’ union influence and state
agenda- and policy-setting?
The third paper explores the impact of a specific high-profile CBA policy—seniority-based
teacher transfers—on the distribution of teachers in schools.
3. Schools—Do seniority-based transfer provisions exacerbate within-district gaps
in teacher quality and student achievement?
Paper one (chapter two) provides new evidence on the political activity and policy-setting
agenda of the largest national teachers’ union during a time of rapid political change. I rely on
rich descriptive time-series data on NEA federal policy positions, campaign contributions, and
election successes from the NEA’s annual Legislative Report Card, the National Institute on
Money in State Politics’ campaign contribution database, and the Federal Election Commission
(FEC) databases on candidates and election outcomes. I find that NEA Democrat allies have
decreased precipitously over time with the election of a Republican majority in Congress.
Nonetheless, the NEA still experiences considerable success in congressional roll call votes
partly because of the election of a growing contingent of Republican allies in the House and
Senate.
While the amount of research on teacher collective bargaining continues to grow, few
researchers have paid attention to the role teachers’ unions play in state-level politics. In paper
two (chapter three), I provide new evidence on the agenda- and policy- setting roles of state
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 10

teachers’ unions. Using novel measures of state teachers’ union influence and a database of
proposed and enacted state teacher policies from 2011 and 2015, I estimate a series of state fixed
effects models that predict the proposal and enactment of state legislation related to teachers’
unions’ interests. I find that state teachers’ unions are more successful at protecting and
promoting their legislative interests in states where they are more powerful.
Finally, as policymakers search for solutions to address the inequitable distribution of
experienced and quality teachers across schools, seniority-based transfer provisions in teacher
collective bargaining agreements (CBAs) are receiving more attention. Nevertheless, even as
new reforms to teacher transfer rights are being proposed, there is little consistent evidence that
transfer provisions significantly contribute to gaps in teacher experience, teacher quality, or
student achievement across disadvantaged student subgroups. In the third paper (chapter four), I
draw upon a longitudinal database of teacher CBAs and school and district-level data in
California to estimate a series of district and year and district-by-year fixed effects models that
gain identification from the within-district changes in teacher transfer and vacancy policies over
time. Although I find significant gaps in teacher experience and teacher education between
advantaged and disadvantaged schools in the sample, my results suggest that seniority-based
transfer policies do not make the situation worse. However, I do find evidence that increases in
the overall restrictiveness of contract language, beyond just provisions governing teacher hiring
and placements, disproportionately harms the performance of students in disadvantaged schools
over time, thus potentially worsening within-district gaps in student achievement.
Each of the three papers in this dissertation adds to the literature on teachers’ unions in an
important way. The first paper of this dissertation generates new conclusions on how the shifting
political structure in Washington influences the federal policy-setting power of teachers’ unions.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 11

The second paper is the first to estimate the influence of teachers’ unions on state policy utilizing
more than one year of data and across several different measures of state teachers’ union
influence. The panel dataset enables the use of more rigorous identification strategies that can
control for different sources of omitted variables, and thereby generate more definitive
conclusions on the policy-setting power of state teachers’ associations. Finally, the third paper of
this dissertation adds to a small body of literature on teacher transfer policies by undertaking the
first longitudinal study of how these policies influence the distribution of teacher experience and
student achievement in schools with higher and lower percentages of minority students.
Consequently, all three studies work towards expanding the breadth of empirical research on the
impact of teachers’ unions—and the policies they espouse—in Congress, statehouses, and
schools. It is the type of evidence marshalled by this dissertation that policymakers need in order
to inform legal changes to the codified protections for teachers’ unions and collective bargaining.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 12

CHAPTER TWO- Down but Not Out: The National Education Association in Federal
Politics
1

Introduction

The federal political environment surrounding teachers’ unions is becoming increasingly
complex and contentious (e.g. Johnson, 2017; Kelley & Mead, 2017; Young, 2011). Spurred by a
new presidential administration and secretary of education who favor privatized education
(Maxwell, 2017) and a Republican majority in both chambers of Congress, teachers’ unions face
renewed political opposition from school choice proponents and other education foundations and
advocacy groups (Reckhow & Snyder, 2014). The National Education Association (NEA)
declared President Trump’s 2018 budget, which would direct approximately 1.4 billion dollars to
school choice options and cut education funding overall by 9.2 billion dollars, a “wrecking ball
aimed at public schools” (NEA, 2017a). The NEA president recently pledged that the union
would “not find common ground” with the new president and with Secretary of Education Betsy
DeVos, who she said “has made a career trying to destroy neighborhood public schools (Strauss,
2017).” Even under the previous presidential administration, the NEA publicly called for the
resignation of President Obama’s first education secretary and openly battled the implementation
of new compensation schemes and evaluation systems incentivized under federal Race to the
Top, Teacher Incentive Fund, and No Child Left Behind Waiver programs (Grunwald, 2015;
Loewus, 2014; Smith, 2009).
The incentive for teachers’ unions’ participation in federal politics is quite clear: many of
the laws and programs that govern the aspects of education that they care about originate with
and are funded by federal lawmakers (Lott & Kenny, 2013; Moe, 2011; Winkler, Scull, &

1
This paper was published in Educational Policy volume 32(2), pgs. 234-254 in January 2018.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 13

Zeehandelaar, 2012). Not surprisingly, teachers’ unions contribute a significant amount of
money to federal election campaigns. With expenditures totaling over 59 million dollars in
federal elections between 1989 and 2010, teachers’ unions typically spend more than most, if not
all, interest groups. (Moe, 2011; Winkler, Scull, & Zeehandelaar, 2012). In last year’s election
cycle (2016) alone, the National Education Association (NEA) disbursed approximately 22
million dollars in state and federal campaigns, which places them as the 10
th
largest spender
among the more than 18,000 Super Political Action Committees (PACs) in American politics
(Burnette, 2016b).
While there is some evidence to substantiate the idea that teachers’ union wield influence
in federal policy-making (Manna, 2006; Moe, 2011; Winkler, Scull, & Zeehandelaar, 2012),
beyond documenting their campaign expenditures, the extant literature is surprisingly thin in
regard to teachers’ union activity in federal politics (Cowen & Strunk, 2015). In particular, we
know little about the nuances of teachers’ union political action, including how often their
favored candidates win elections, how often their favored policies are enacted, and the degree to
which these outcomes change over time with changes in the political environment in
Washington. This paper begins to address these gaps in the literature using a dataset comprised
of the National Education Association official positions on federal policy, NEA published grades
for Congress members, and information on federal campaign contributions over time. I focus on
describing the NEA’s election and policy involvement and outcomes over a period where
massive shifts occurred in the partisan control of the House, Senate, and presidency. In
particular, I ask and answer the following research questions: 1) How has the distribution of
NEA’s political allies and foes changed over time in the House and the Senate?; 2) Given the
observed changes to the distribution of NEA allies and foes over time, who did the NEA (and
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 14

their opponents) target with campaign contributions during the 2010, 2012, and 2014 elections
and to what extent did their favored candidates win?; and 3) What legislation (e.g. school
choice) has the NEA officially supported and opposed and how has the success rate of this
legislation changed over time?
The results show that the proportion of NEA allies (or legislators granted A or B ratings
by the NEA) has decreased over time in both chambers; However, even with the loss of the
Democrat majority in the Senate and House, the NEA still maintains a majority of A/B-rated
allies in the Senate and regained a slight majority in the House after the 2014 midterm elections.
Moreover, even though the NEA has lost a significant number of allied Democrat seats in the
House and Senate, they have gained important Republican allies, particularly in traditional
Democrat states, and they are increasingly donating in larger amounts to these Republican allies’
campaigns. I find that NEA-supported candidates win more often than they lose in federal
elections and NEA election success is strongly correlated with the voting behavior of Congress.
Overall, the results suggest that the new Republican majority in Congress has perhaps weakened
but not altogether undermined NEA influence in federal politics and policy-setting.
In the remainder of this chapter, I first overview my data and research methods. I then
answer each research question and conclude with implications for the federal education agenda.
Data

To explore patterns in the political activity of the National Education Association, I
combine data from three sources: 1) the NEA’s Legislative Report Card; 2) the National Institute
on Money in State Politics’ campaign contribution database; and 3) the Federal Election
Commission (FEC) databases on candidates and election outcomes.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 15

Data on NEA “allies” and “foes” are derived from the NEA’s official Legislative Report
Card. Updated annually on their website and grouped by two-year congressional sessions (e.g.
2009-2010), the dataset contains letter grades on an A through F scale for all members of the
Senate and House who participated in Congress during the session. Grades are determined
primarily from how lawmakers voted on key legislation “critical to advancing the NEA’s
identified legislative priorities” (NEA, 2015). Congressional members are additionally scored for
cosponsoring bills that align with the NEA’s priorities, for performing “behind-the-scenes-work
to advance or impede NEA priority issues,” and for their accessibility to NEA leadership in
Washington, D.C. and in their home state or district. These data are available from the 2009-
2010 to the 2015-2016 congressional session.
Data on campaign contributions are derived from the National Institute on Money in
State Politics, which tracks campaign contributions for all federal candidates for public office
beginning in 2010 by gathering information from campaign financial disclosure reports required
by federal law. I pull data on all contributions from the National Education Association as well
as those from three opponent groups—school choice advocates, anti-labor, and business
organizations during the 2010, 2012, and 2014 elections. Following the methodology of
Mulvihill (2017), I define school choice advocates as individuals and groups that donated over
100,000 dollars to pro-school choice ballot initiatives from 2009 to 2017. This list, which
includes 66 different individuals and organizations, represents some of the top funders of school
choice causes around the nation including billionaire education reform philanthropists like Bill
Gates and Alice Walton as well as pro-charter organizations like Families for Excellent Schools,
Education Reform Now, and K12 Inc. To track anti-labor contributions, I rely on the
classification of the National Institute on Money in State Politics, which classifies contributions
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 16

from the National Right-to-Work Committee as distinctly anti-labor. Consequently, all anti-labor
contributions in this study are representative of this group. General business association
contributions are also aggregated by the institute and include donations from over 500 different
corporations.
Data on election outcomes are publicly available from the FEC through the 2014 election
cycle. The election outcomes data contains the primary and general election vote totals and
election outcome status for each candidate for federal office. In particular, the campaign
summary file contains information on all candidates who filed for candidacy for federal office,
including their name, representative chamber, district, state, incumbency status, and political
party affiliation.
Finally, data on NEA policy positions are also derived from the NEA’s Legislative
Report Card. While the calculations that result in a lawmaker’s final grade are not made public,
the House and Senate votes contributing to the grades are published in the official report card. I
use these votes to determine the NEA’s legislative agenda during the 2009-2010, 2011-2012,
2013-2014, and 2015-2016 congressional sessions (what types of legislation they supported or
opposed) and to determine the outcome of this legislation.
Methods

The purpose of this paper is to extend the research evidence on teachers’ unions’ federal
political participation and influence by 1) revealing trends in the distribution of NEA political
allies and foes over time; 2) showing patterns in the amount and success of NEA and NEA
opponent campaign contributions to federal lawmakers; and 3) describing the success of the
NEA’s favored policies. I compare differences in NEA and opponent campaign contributions
and candidate win percentage across key state and legislator characteristics, including NEA
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 17

Legislative Report Card grade, election year, and partisan affiliation. It is important to note that
this analysis is not meant to be causal nor is it meant to infer a direct relationship between
teachers’ union contributions, candidate election success, and the enactment of union-supported
policies. The findings should be interpreted as descriptive evidence of the political participation
of the NEA and other groups in federal politics.
To answer research question 1 on changes to the distribution of NEA “allies” and “foes”
across the Senate and House over time, I first explore NEA legislator grades by congressional
session and chamber. NEA allies are conceptualized as those who receive an A or B grade on the
annual report card and NEA foes are constituted as those who receive a D or F. To further
explore how the distribution of NEA allies and foes changed with shifts in the partisan
composition of the legislature over time, I calculate the change in NEA ally and foe
congressional seats from the 2009-2010 to the 2015-2016 congressional session and between
adjoining sessions (e.g. from 2009-2010 to 2011-2012). I explore changes in the number of ally
and foe seats by the political affiliation of the legislator occupying the seat and by the partisan
political affiliation of the state. Because the NEA grade database does not independently report
legislator party affiliation, I linked the NEA grade data with the FEC campaign summary data
from the election just immediately after the two-year congressional legislative session (e.g. the
2010 election was matched with the 2009-2010 legislative session). Due to the lack of a common
numeric identifier for each lawmaker between the two files, candidates were linked by first and
last name, election year, election jurisdiction, legislative chamber, and legislative district (for
House candidates), which resulted in a complete one-to-one match. The partisan political
affiliation of the state was determined by the partisan vote share received by Democrat and
Republican presidential candidates since the 2000 election. Republican states are defined as
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 18

those that consistently voted majority Republican in presidential elections from 2000 to 2012,
and Democrat states are defined as those that voted Democrat in the same elections. Swing states
are those that voted for both Democrat and Republican presidential candidates from 2000 to
2012.
To answer research question 2 regarding patterns in NEA and NEA opponent campaign
contributions, I calculate NEA, school choice advocate, anti-labor, and general business
campaign contributions during the 2010, 2012, and 2014 federal elections by legislator party and
NEA grade. Again, due to the lack of a common number identifier across the National Institute
on Money in State Politics’ campaign contribution database, the FEC election outcomes and
campaign summary database, and the NEA Legislative Report Card database, I linked candidates
by first and last name, election year, election jurisdiction, legislative chamber, and legislative
district (for House candidates), which again resulted in a complete one-to-one match across the
three datasets. After linking the three databases, I now observe every candidate that ran for an
open Senate or House seat in the 2010, 2012, and 2014 elections. To control for differences in
the size of legislatures across states and for the number of races in each election year, NEA,
school choice advocate, anti-labor, and general business contributions are reported per candidate
by dividing the total number of contributions from the interest group by the total number of
candidates running for office in each cross-tab cell. To explore changes in the type and success
of NEA-supported candidates over time, I generated both contribution and win percentages (i.e.
election success rate) for NEA candidates in each cross-tab cell. The contribution percentage
should be interpreted as the proportion of candidates receiving a contribution from the NEA. The
win percentage should be interpreted as the proportion of NEA candidates elected out of the total
number of candidates who received NEA contributions.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 19

Finally, to answer research question 3, I performed a content analysis on the individual
roll call votes in the NEA Legislative Report Card database by congressional session and
chamber. An analysis of the topic of the 148 House and Senate roll call votes available in the
database resulted in 13 main content areas (i.e. Budget/Funding, Higher Education, Health Care,
Social Issues, School Choice, Union Rights/Dues, Tax Cuts, Compensation/Benefits, Wages,
Education Job Creation, School Modernization, Gun Violence Prevention, and Immigration
Reform) that received lawmaker attention during the 2009-10, 2011-12, 2013-14, and 2015-2016
congressional sessions. Included in the Report Card database are roll call votes on specific bill
amendments as well as votes on final passage. Across the topic areas, I generated statistics for
the number of votes supported/opposed by the NEA and the NEA policy success rate (defined as
the number of vote outcomes (pass/fail) that supported the NEA official position divided by the
total number of votes taken).
Results
How has the Distribution of NEA Allies and Foes Changed Over Time?
Figure 1 displays how the distribution of NEA allies and foes (determined from the
grades assigned to lawmakers on the annual NEA Legislative Report Card) has changed over
time. The proportion of A and B members (allies) is shown in green and the proportion of D and
F members (foes) is shown in grey. The figure also plots the partisan distribution of Congress
where any point above black dotted line at 50 percent represents partisan majority control. The
proportion of Republicans is shown in red and the proportion of Democrats is shown in blue.
Figure 1 demonstrates that during the 2009-2010 congressional session, when Democrats
maintained a majority of 58 percent in the Senate and 59 percent in the House, the NEA held a
majority of A/B allies in both chambers at 61 percent and 59 percent respectively.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 20

Comparatively, the proportion of D/F foe members was much lower at 36 percent in the Senate
and 38 percent in the House.
The figures make clear that the proportion of A/B members in both chambers has
decreased over time. The decrease was quite sharp in the House, where the proportion of A/B
members dropped to 48 percent in the 2011-2012 congressional session, which largely tracked
the loss of the Democrat majority in the 2010 mid-term elections (shown by the dip in the blue
line in the 2011-2012 congressional session). Nonetheless, even as Democrats have yet to regain
control of the House, the proportion of A/B House members continues to increase over time, to
the point where the NEA allies again hold the majority (albeit by a small amount) in the 2015-
2016 congressional session. Comparatively, change in the Senate composition of allies and foes
was more gradual. Although the change in the proportion of A/B members in the Senate largely
trends in the same direction as the change in proportion of Democrat senators, Democrats
surrendered their control of the Senate during the 2014 mid-term elections (shown by the dip in
the blue line in the 2015-2016 congressional session) whereas NEA allies still maintain majority
control. Overall, the proportion of A/B members in the Senate decreased from 58 percent in
2010-2011 to 53 percent in 2015-2016.
Table 1 shows more detail on the extent of the changes in NEA ally and foe Senate and
House composition by showing the overall change in the number of seats broken out by political
party, NEA grade, and state political party preference. Panel A displays the changes for the
Senate, Panel B displays the changes for the House, Column 1 displays the overall changes,
Column 2 displays the changes for firmly Democrat states, Column 3 displays the changes for
firmly Republican states, and Column 4 displays the changes for swing states. The table reveals
that much of the loss of A/B members observed in Figure 1 can be attributed to a loss of
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 21

Democrat A/B members across all three types of partisan states (firmly Democrat, firmly
Republican, and swing states). Panel A, Column 1 shows that between 2010 and 2016, the NEA
experienced a net defeat of 12 Democrat A/B members in the Senate, which occurred after the
2010 and 2014 mid-term elections. Column 3 of Panel A demonstrates that most of the loss of
A/B Democrat allies occurred in firmly Republican states (-8 seats), but the NEA also lost allies
in firmly Democrat (-1 seat) and swing states (-3 seats). Nonetheless, as shown in Figure 1, the
NEA did not forfeit their majority in the Senate with the loss of the Democrat majority during
the 2014 mid-terms. This can be partly attributed to an overall increase in the number of A/B
allied Republicans (+4 seats) as show in Column 1 of Panel A, which occurred mainly in Firmly
Republican states (+3 seats) (Panel A, Column 3).
This pattern of Democrat ally loss but increase in Republican allies is even more
pronounced in the House. As shown in Panel B, Column 1, the NEA lost 61 A/B Democrat allies
but gained 26 Republican allies. Much of the defeat of Democrat allies occurred after the 2010
mid-term elections (-57 seats) but also took place after the 2014 mid-terms (-12 seats). However,
the NEA added 12 Republican allies after the 2010 mid-terms and 18 Republican after the 2014
mid-terms which helped stem the tide of ally-loss that occurred in 2010. This partly explains why
the NEA succeeded in regaining majority-ally control of the House after the 2014 mid-terms
despite the Republicans still maintaining majority control (as shown in Figure 1).
Who Did the NEA (and Their Opponents) Target with Contributions During the 2010,
2012, and 2014 Elections and to What Extent Did Their Favored Candidates Win?
Given the importance of allies to the success of the NEA’s legislative agenda, we might
expect the NEA to strategically target their campaign contributions to these allies with hopes of
improving their favored candidates’ election prospects. Similarly, their opponents may direct
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 22

contributions to NEA foes, with hopes of electing lawmakers favorable to their legislative
interests. Important to note from this analysis is that I am unable to disentangle whether the
NEA’s campaign contributions (or those of their opponents) cause candidates to win. It is
entirely possible that these interest groups strategically donate to candidates that are likely to
achieve office because the prospects of receiving a return on their campaign investment is
higher. Nonetheless, the observed patterns are still instructive for understanding the NEA’s
contribution strategy in a time of political change in Congress.
Tables 2 explores where the NEA and their opponents strategically target their campaign
expenditures to potentially shape the distribution of their political allies and foes in Congress.
Table 2 shows the amount of NEA, school choice advocate, anti-labor group, and general
business campaign contributions per candidate for the Senate and House for three types of races:
1) when the incumbent is a Democrat A/B ally and the challenger is a Republican; 2) when the
incumbent is a Republican A/B ally and the challenger is a Democrat; and 3) when the
incumbent is a Republican D/F incumbent and the challenger is a Democrat.
2
Recall that NEA
legislator grades are matched to the election year directly following the legislative cycle (e.g.
2009-2010 grades are matched with the 2010 election cycle) such that the NEA’s contributions
in the election cycle should reflect a current knowledge of how an incumbent lawmaker voted in
the last legislative session.
A few patterns emerge from an analysis of NEA and opponent group contributions in the
Senate and House. First, Table 2 reveals that the NEA never contributes in opposition to a

2
It is important to note that to receive a grade in the NEA Legislative Report Card, a lawmaker must have served for
a least one session. Consequently, this analysis excludes open races between Democrats and Republicans where no
graded incumbent exists. Furthermore, I exclude a small number of uncontested races where the incumbent goes
unchallenged. Finally, races with a D/F Democrat incumbent are excluded because no D/F Democrats exist in the
Senate and very few exist in the House.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 23

Democrat “A/B” candidate and rarely donates against a Republican “A/B” candidate. Instead the
organization prefers to maintain support for their allies in the House and the Senate regardless of
party preference when they already hold congressional seats. For example, the NEA contributed
5,788 dollars to Democrat A/B candidates and zero dollars to Republican challengers in the
House across all election cycles (Table 2, Column 5). Similarly, they donated 3,500 dollars per
Republican A/B candidates in comparison to only 517 dollars per Democrat challenger. The
patterns are less clear in the Senate where there are fewer races, but even so, the NEA directed
8,726 dollars per Democrat A/B candidates and zero dollars to their challengers. They actually
contributed more per candidate towards Democrat challengers (3,333 dollars) versus Republican
A/B candidates (1,000 dollars) (Table 2, Column 1) though these types of races were rare,
occurring only once in each election year.
In some cases, the NEA faces opposition from school choice advocates and general
business interests in these races, although school choice advocates’ support for NEA allies often
far outweighs their contributions to NEA opponents. For example, school choice advocates
donated 5,301 dollars per candidate to Republicans challenging A/B Democrat incumbents in the
Senate, yet they directed 13,319 dollars per candidate to the Democrat A/B incumbents in these
races, exceeding the NEA’s 8,726 dollars to these same politicians (Table 2, Column 1)
3
.
General business groups donated far more to Senate Democrat A/B incumbents at 334,933
dollars per candidate compared to their support to Republican challengers at 154,920 dollars. In
comparison, the National Right to Work Committee (anti-labor) donates exclusively in
opposition of the NEA’s favored candidates but does so in much smaller amounts on a per-

3
This may represent a strategic effort by school choice advocates to target incumbent lawmakers who are engaged
on education issues. It appears that school choice advocates donate much more to incumbents versus non-
incumbents and do so regardless of party preference and legislator grade. School choice advocates may be making
efforts to sway key incumbent lawmakers towards more favorable school choice positions.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 24

candidate basis
4
. Consequently, while general business and school choice advocates often donate
more to candidates than the NEA, their contribution strategy is less cohesive and they frequently
donate in favor of incumbent union-supported candidates. While anti-labor frequently opposes
union candidates, their resources are far less substantial when compared to the donations
commanded by the NEA.
Finally, given the increasing importance of A/B Republicans to the maintenance of an
NEA allied majority in the House, it is important to note that NEA contributions to Republican
A/B candidates per candidate nearly doubled over time to House A/B Republican lawmakers.
The NEA contributed 2,125 dollars to Republican A/B candidates in 2010 compared to 5,000
dollars per candidate in 2014 (Table 2, Columns 6-8). The same trend is not found in the Senate,
though again there are altogether fewer Republican A/B races during the observation time period
and thus it is more difficult to detect a general trend.
Figure 2 shows the contribution rates from the NEA to candidates in the key races shown
in Table 2. The solid lines represent the Senate contribution rates and the dashed lines show the
House contribution rates. The important takeaway from this figure is that while the NEA
continues to contribute to Democrat A/B candidates running against Republicans at similar (and
very high) rates over time (as shown by the blue lines) their support to Republican A/B
candidates running against Democrats has increased quite substantially, particularly from the
2012 to the 2014 election years (as shown by the red line).
5

4
When they did donate, anti-labor groups contributed an average amount of between 596 dollars and 2602 dollars to
their supported candidates, depending on the election cycle. This is in contrast to the average NEA contribution,
which ranged from 1262 dollars to 1341 dollars, depending on the election cycle. Consequently, the average
donation size for anti-labor groups is similar to that of the NEA but they donate to a much smaller number of races.
5
The NEA’s contribution rate to Senate A/B Republican candidates is not shown because of the small number of
races.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 25

Figure 3 shows the election success rates of NEA supported candidates in these same
races. The NEA’s overall election success rate (or the proportion of NEA financially supported
candidates elected to public office) is shown by the yellow line. The success rate for specific
types of candidates are demarcated by the blue (A/B Democrat incumbent against a Republican
challenger), red (A/B Republican incumbent against a Democrat challenger), and light blue
(Democrat challenger against a D/F Republican incumbent) lines, identical to Figure 2. The
figure makes clear that the NEA wins more often than they lose elections. This could indicate
either that NEA campaign contributions facilitate election success for recipients or that the NEA
successfully identifies and strategically targets their contributions to candidates that are likely to
win. Perhaps not surprisingly, their election success rates also track the proportion of A/B
members in the House and Senate over time as was shown in Figure 1. The NEA experienced
less success in the 2010 and 2014 mid-term elections in the Senate (and therefore I observe lower
proportions of A/B members in 2012 and 2016) and increasing success in the 2012 and 2014
midterms in the House (and therefore I observe higher proportions of A/B members in 2014 and
2016). Importantly, not only is the NEA contributing more money and contributing more
frequently to Republican A/B candidates as shown in Table 2 and Figure 2, but they are also
experiencing more success in these campaigns over time as shown by the red line in Figure 3.
What Federal Legislation (e.g. School Choice) has the NEA Officially Supported and
Opposed from 2009-2016 and What are the Outcomes of this Respective Legislation?
Table 3 documents the frequency of the votes opposed and supported in the NEA
Legislative Report Card overall and by topic. Again, the NEA Legislative Report Card is derived
from congressional votes on the NEA’s stated legislative priorities. Consequently, for an issue to
appear in Table 3 it must be 1) a priority of the NEA and 2) voted on by the House and/or
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 26

Senate. While the overall total reflects the sum frequency of all votes supported or opposed by
the NEA in a given year, only topics receiving three or more votes by lawmakers are shown in
more detail in the table and are discussed below.
During this eight-year span, the NEA took an official position on a combined total of 148
House and Senate roll call votes. Their success rate on these votes overall is the highest in the
Senate (79 percent) where the NEA maintained an ally majority during the observation period (as
was shown in Figure 1). However, the success rate is the lowest in 2016, after the Democrats lost
majority control of the Senate in the 2014 election. The success rate precipitously declines in the
House after 2010, when the Republican gained majority control and the proportion of NEA allies
dropped below 50 percent. However, the NEA House success rate increased again in 2016 (from
41 to 63 percent) when the NEA regained a majority of allies.
The NEA’s most successful and voted on issues across both chambers include higher
education (100 percent in the Senate and 67 percent in the House) (e.g. student loan policy),
social issues (100 percent in the House and the Senate) (e.g. equal pay measures), union rights
(67 percent in the House and the Senate) (e.g. union certification procedures), health care (86
percent in the Senate and 56 percent in the House) (e.g. the Affordable Care Act), benefits (100
percent in the Senate and the House) (e.g. unemployment insurance), facilities (100 percent in
the Senate and House) (e.g. school modernization), and budget and funding (68 percent in the
House and 57 percent in the Senate) (e.g. budget cuts).
On some topics, Senate votes consistently aligned with NEA positions but House votes
did not. For example, the Senate voted with the NEA 100 percent of the time on school choice
issues, whereas the House only supported the NEA position 38 percent of the time. In the case of
school choice, Republicans used their majority in the House to pass several bill amendments
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 27

related to the reauthorization of the school voucher program in the District of Columbia that
were staunchly opposed by the NEA in the 2011-2012, 2013-2014, and 2015-2016 legislative
sessions. Notwithstanding the support of many Republicans, the Senate failed to give final
approval for reauthorization of the program, consistent with the NEA position. Similarly, in
relation to tax cuts, the Senate voted consistent with NEA positions 100 percent of the time but
the House only aligned with the NEA 33 percent of the time. The NEA supported votes on the
Middle Class Tax Cuts Act, which received approval in the Senate but not in the House.
Figure 3 maps the overall NEA Senate and House policy success rates (black line) with
the NEA election success rates (yellow, blue, light blue, and red lines). The election and policy
success rates are placed on different x axes so that the election success rates are now
contemporaneous with next year’s congressional session (or in other words the 2010 election is
contemporaneous with the 2011-2012 congressional session in the figure because lawmakers
elected in 2010 did not take office until 2011). The figure shows that the NEA’s election success
tracks their policy success rate very closely in both the House and the Senate, suggesting that
NEA election success in a given election year is fairly indicative of their policy success in the
subsequent legislative session.
Conclusion
This study extends the small body of literature on teachers’ unions’ in federal politics by
exploring the nuances of the National Education Association’s efforts in Congress in a period of
rapid political change. With a Republican majority in the House and Senate and a presidential
administration friendly to charter and private schooling options, teachers’ unions face new policy
options that, in the words of the NEA president, are “reckless and wrong for students and
working families” (NEA, 2017b). In this paper, I explore the extent to which the shifting status
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 28

quo in partisan politics in Washington towards a Republican majority has changed the influence
of the NEA on Capitol Hill, including the distribution of their allied lawmakers in the House and
Senate, the degree to which their favored candidates win elections, and extent to which their
favored policy positions are supported by Congress.
A few important findings emerge from this analysis. First, I find that the onset of a
Republican majority in Congress certainly spurred a decrease in the number of Democrat allies
on Capitol Hill—the NEA lost 61 allied Democrat members in the House and 12 allied
Democrats in the Senate since the 2010 election. While this could surely spell doom for a labor
union that is traditionally dependent on Democrat support, the loss of ally members was partially
offset by an increase in Republicans friendly to NEA positions. The NEA added 4 A/B
Republican senators and 16 A/B Republican representatives, most in firmly Democrat states,
where taking union friendly positions is ostensibly more acceptable among the electorate. The
increase in Republican A/B candidates was substantial enough to help the NEA maintain their
allied majority in the Senate and regain their allied majority in the House after the 2014 mid-term
election. Further, I find that the NEA is contributing larger sums of money to allied Republican
campaigns over time which provides some suggestive evidence that partisanship matters less to
the union than how a member will vote on legislation.
Second, I find that NEA-supported candidates win more often than they lose in federal
elections. This does not necessarily connote that NEA contribution causes federal lawmakers to
succeed in their election bids, but instead could suggest that the NEA is aware of the electability
of candidates and targets their contributions to those campaigns with a high probability of
succeeding (or those with a high likelihood of generating a return on the election investment).
Importantly, I find a strong relationship between the election success rate of NEA-supported
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 29

candidates and the proportion of congressional votes that align with union interests. This means
that while the NEA policy success rate has declined in the House and Senate over time, the
election success of a new wave of NEA allies, particularly among Republicans, may help stay
the decline.
Finally, when looking at the relative financial influence of the NEA when compared to
their opponents, I observe that business and school choice groups rival, and in most cases,
exceed the NEA in donation amounts but they donate to a broader range of candidates,
contributing in some cases, to NEA allies in substantial amounts. Anti-labor groups donate
almost exclusively to NEA foes but their expenditures are nowhere near the amounts the NEA
contributes to NEA allies. This result diverges somewhat with the findings of Moe (2011) who
argues that teachers’ unions expend more AND are targeted in their political strategy when
compared to their opposition. Instead teachers’ unions are actually outspent by some groups and
they outspend others.
Overall, there is little to suggest in this paper that NEA influence in federal politics will
disappear or even substantially diminish with the shifting political climate in Washington. While
the results suggest some decline in the number of NEA allies in Congress and the subsequent
success of NEA policy positions, the presence of a contingent of Republican allies offers
renewed opportunity for the union to maintain a presence in federal politics.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 30

CHAPTER THREE- Preserving the Status Quo?: The Relationship between
Teachers’ Union Strength and State-Level Agenda- and Policy-Setting Influence
Introduction
Policy scholars have documented a marked shift in the jurisdictional authority of
education policy-setting in recent years (Conley, 2003; Fusarelli & Fusarelli, 2015; Gayley,
2015; Henig, 2013; Manna, 2006). Spurred by growing concern about the lagging achievement
of the United States versus other countries documented in reports like A Nation at Risk, the
modern education reform movement can be classified by increasing state and federal
involvement in education decision-making and the squeezing out of local control. Federal and
state education policies like No Child Left Behind (NCLB) (2001), Race to the Top (RTT)
(2009), Common Core State Standards (2009) (CCST), state recovery school districts,
performance evaluation, tenure, and compensation reform exemplify how state and federal
politicians are now involved in decision-making around the “core of the education enterprise”—
decisions about teaching, curriculum, performance, and compensation that were formerly the
prerogative of local school boards (Henig & Bulkley, 2010, p. 323).
The shifts in the education policy-setting arena to more state and federal control provide
new and different opportunities for political actors to influence the education policy agenda
(Cibulka, 2001; Gayley, 2015; Henig, 2013; McDonnell, 2013). In particular, researchers have
long noted that local education policy-setting favors traditional education interests (Goldstein,
2014; Henig, 2013; Moe, 2005; 2006; 2011). For example, historically low voter turnout in
school board elections affords teachers’ unions the opportunity to drastically shape the
composition of school boards and ultimately who ratifies their collectively bargained agreements
(Moe, 2005, 2006). Indeed, the lack of interest group pluralism in school districts means that
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 31

organized teachers’ groups have substantial power in shaping the content and implementation of
local education policy. However, as the axis of decision-making shifts to the state and federal-
level, traditional education interest groups, like teachers’ unions must compete (and succeed) in
these policy venues against interest groups that may not have a voice at the local-level, but are
primed to advocate for their policy positions in state and federal legislatures (Cooper & Sureau,
2008; Goertz, 2009; Henig, 2013).
Evidence also exists that the state and federal interest group space in education is
becoming more crowded (Bjork & Lindle, 2001; Mazzoni, 1995; Mahwinney & Lugg, 2001;
Opfer, 2001; Reckhow, 2013; Reckhow & Snyder, 2014). Teachers’ unions still maintain
impressive state- and federal-level organizations, with large staffs and significant resources
aimed at influencing legislative policies. They are consistently some of the largest spenders in
state and federal elections, contributing nearly 22 million dollars to candidates and committees in
the 2016 election cycle alone (National Institute on Money in State Politics, 2016). Nonetheless,
unions are now joined by new groups who aim to shape the education policy landscape (Opfer,
Young, & Fusarelli, 2008). For example, Reckhow and Snyder (2014) note the rise of
philanthropic foundations in education politics and their funding of “jurisdictional challengers”
to traditional education institutions like public school districts and teachers’ associations.
Additionally, others have noted a marked increase in the number of conservative policy think
tanks and school choice advocates who aim to expand schooling options and change the way
teachers are evaluated, compensated, and retained (Cibulka, 2001; Gayley, 2015; Sawchuk,
2012).
As the locus of education-decision making shifts and as the number of education interest
groups grows, teachers’ unions’ success in state and federal policymaking is perhaps more
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 32

critical to their survival than in times past but also more challenging (Moe, 2011). Theories of
interest group influence posit that as the institutional arrangements surrounding policy-making
change, so does interest group policy-setting power (e.g. Baumgartner & Jones, 1991; Jones &
Baumgartner, 2012; Sabatier and Jenkins- Smith, 1988, 1993; Sabatier and Weible, 2007). In the
case of teachers’ unions, it remains to be seen how the changing dynamics of state and federal
education decision-making, including the increase in interest group pluralism, have influenced
their ability to impact legislative outcomes. Existing research on teachers’ unions and
policymaking has primarily focused on their local policy-setting role in negotiating teacher
contracts with school districts (see Cowen & Strunk, 2015 for a review). A few scholars explore
the relationship between teachers’ union strength and state-level policy but do so at a single point
in time and fail to consider the role of competing interest groups as a check on teachers’ union
policy influence (e.g. Finger, 2017; Hartney & Flavin, 2011; Renzulli and Roscigno, 2005;
Shober, Manna, & Witte, 2006).
In this paper, I investigate the relationship between teachers’ union strength relative to
“opposing groups” and the proposal and adoption of state legislation across five state legislative
cycles from 2011 to 2015. I employ a self-collected nationwide dataset that chronicles all
proposed and enacted policies across 21 teacher policy topics and new measures of teachers’
union and opposition group strength that capture multiple dimensions of union power relative to
competing interests (i.e. membership and revenue power, election influence, and lobbying
influence). These measures are used in a series of within-state fixed effect models that predict
agenda-setting and policy-setting outcomes. The legislative database used in this study contains
nearly 4,000 state teacher laws that are coded based on their intent in limiting teacher and union
rights—I define “favorable” laws as those that preserve or enhance the rights of teachers or
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 33

teachers’ unions or the scope or impact of CBAs and “unfavorable” laws as those that constrain
the rights of teachers or teachers’ unions or the scope or impact of CBAs.
I find that teachers’ unions strength relative to competing interest groups is uniquely
associated with teachers’ agenda-setting success but not final policy outcomes. A higher
proportion of favorable teacher laws and a lower proportion of unfavorable teacher laws are
proposed in states where unions exert more strength in comparison to their opposition. Final
policy-setting outcomes, however, appear to be associated with the absolute strength of teachers’
unions in a state, and this is particularly true when considering teachers’ unions’ influence over
election outcomes. In short, consistent with theoretical predictions (e.g. Kingdon, 2003; Moe,
2011, 2015) minority education groups may have an easier time shaping the agenda-setting
process versus influencing final policy outcomes. Notwithstanding, the proliferation of new
interest groups in education politics, teachers’ union strength is the strongest predictor of
favorable teacher policy success and unfavorable teacher policy failure.
This paper is organized as follows: First, I provide theoretical background on how we
might expect teachers’ union power to change as the institutional environment surrounding
American education policymaking shifts. I then overview existing research on the relationship
between teachers’ union power and state policy outcomes. Next, I outline the data and methods
used in this study, including the measures of teachers’ union power and agenda- and policy-
setting outcomes. Finally, I discuss the results of the study and conclude with a discussion on
what the results entail for education policy-setting and teachers’ union influence.
Teachers’ Union Power in an Era of Institutional Change in Education Policymaking
Due to their large financial reserves, strong membership, and political involvement,
teachers’ unions are often recognized as the most influential interest group in public education.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 34

Terry Moe writes that “as the most powerful group in education, [teachers’ unions] use their
power to promote [their] special interests—in collective bargaining, in politics” (p. 23).
However, he further notes that teachers’ unions are powerful only to the extent that policymakers
and education bureaucrats are willing to entertain union interests. More formally, Peterson,
Henderson, and West (2012) classify the relationship between teachers’ unions, policy-makers
(i.e. state legislative committees), and government agencies (i.e. state departments of education)
using the classic political concept of the iron triangle, a difficult-to-break, mutually supportive
relationship between legislators, special interests, and bureaucrats that holds in to place the
education policy status quo. Teachers’ unions, as is common for donating special interests, trade
campaign contributions and their electoral prowess to lawmakers in exchange for favorable
policies that protect and promote their interests (Moe, 2011, 2015). Education bureaucrats,
valuing education-specific expertise and looking to support and build trust with teachers (or at
least not run afoul of union leadership), often support the union-favored policies handed down by
policymakers. The result, according to some, is an education system that reifies and reinforces
union power (Moe, 2011; Paige, 2006).
Theories of the policy process highlight how the dynamics of policy-setting and the
power of interest groups change with key institutional shifts. For example, Punctuated
Equilibrium Theory (PET) conceptualizes policymaking as a continual struggle between forces
that maintain stability and the status quo and those that drive policy systems into disequilibrium
and bring about rapid change (Baumgartner & Jones, 1991; 2002; 2009; Baumgartner, Jones, &
Mortensen, 2014; Jones & Baumgartner, 2012). PET argues that iron triangles, also termed
policy monopolies or policy subsystems, are not invulnerable forever, but can collapse as
institutional pressures for change overwhelm the forces holding the subsystem in place. The
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 35

Advocacy Coalition Framework (ACF) provides additional detail on the variety of forces both
within and outside of a policy subsystem that can spur policy change. ACF conceptualizes policy
subsystems as made up of coalitions, or groups of policy actors who hold different values and
conflicting policy goals (Sabatier, 1988; Sabatier and Jenkins-Smith, 1993; Sabatier & Weible,
2007; Weible, Sabatier & McQueen, 2009). Policy subsystems typically contain a dominant
majority coalition and several minority coalitions seeking to alter the prevailing status quo in
public policy, and ACF hypothesizes that the lineup of allies and opposition to the status quo are
relatively stable over time. Policy change occurs, however, as external shocks to the policy
subsystem (e.g. changes in socioeconomic conditions, public opinion, regime changes, or policy
outputs from other subsystems) redistribute the resources among disparate coalitions and as
minority coalitions seize opportunities to interject new policy narratives (Jenkins-Smith et al.,
2014; Sabatier and Jenkins-Smith, 1999; Sabatier & Weible, 2007). In the wake of external
shocks, minority coalitions may maximize their political resources (e.g. legal authority, public
opinion, information, membership base, financial means, and skillful leadership) to tip the
balance of power between coalitions and bring about policy change (Jenkins-Smith et al., 2014;
Weible, Sabatier & McQueen, 2009).
As argued above, current institutional shifts in education policy-making may challenge
the iron triangle of which unions are party to. Cibulka (2001) notes that the K-12 system “is
becoming more functionally centralized, which has changed the array of interest groups
operating in the K-12 arena, where they exercise their influence, whom they seek to persuade
and how, as well as the manner in which coalitions form and dissolve. The discontinuities
created between the old politics of education and the new politics are dramatic (p. 36)).” Jeffrey
Henig (2013) traces the movement of education-decision making away from single-purpose
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 36

governments (i.e. local school boards, state departments of education, U.S. Department of
Education) towards general-purpose governments (i.e. mayors and city councils, state
legislatures, Congress), a phenomenon which he calls “the end of American exceptionalism.” He
argues that education decision-makings has been “reabsorbed” into general purpose arenas,
where “the rules and regulations are less favorable,” for teachers’ unions and other education-
specific groups (Henig, 2013, p. 130). General purpose arenas are characterized by multiple,
intersecting policy domains (e.g. health, transportation) versus single-issue silos (e.g. education
policy), a larger array of competing interest groups, and less dependence on highly localized,
issue-specific expertise. Unlike in education-specific arenas, education groups, like teachers’
unions, are less able to cultivate relationships built on their close connection to the classroom and
to teachers, and instead must compete with other political actors, align their education proposals
with approaches in other domestic policy domains, and win allies “who might not have school-
age children and who might be quicker to ask, ‘how much will this cost’ and ‘how do I really
know this will work as claimed (Henig, 2013, p. 28).’”
The shifting locus of education decision-making is coupled with the rise of new interest
groups and coalitions in public education (Bjork & Lindle, 2001; Mazzoni, 1995; Mahwinney &
Lugg, 2001; Opfer, 2001; Reckhow, 2013; Reckhow & Snyder, 2014). In recent years, new
education advocacy groups have cropped up around the nation, including Stand for Children
(1996), K12 Inc. (2000), Democrats for Education Reform (2007), StudentsFirst (2010), 50Can
(2011), and Students Matter (2011), and these groups are increasingly lobbying lawmakers and
donating to election campaigns (Sawchuk, 2012). Superfine and Thompson (2016) note, for
example, how the Vergara v. California (2014) trial in California, a legal challenge to California
state statutes governing teacher tenure, dismissal, and layoffs brought on by Students Matter,
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 37

illustrates the shifting power of coalitions away from teachers’ unions and the use of new policy
venues by education advocates to influence education policy. Second, philanthropic foundations
are increasingly funding new research and reform efforts, thus bringing new ideas into education
policy-setting, some of which challenge traditional union positions (Reckhow & Snyder, 2014).
Finally, business groups are also becoming more involved in public education by financially
backing some of the advocacy groups mentioned above and by pushing for new accountability
policies and schooling options (Cibulka, 2001; Henig, 2013).
The struggle between teachers’ unions and these competing coalitions is of particular
importance for this study. These new interest groups could fundamentally alter the dynamics of
teachers’ union power and ultimately the mix of education policies proposed and enacted in
legislatures. For example, lawmakers could be less responsive to traditionally dominant special
interest groups, like teachers’ unions, as the influence of competing groups increases and as new
coalitions vie for power (Hansen & Gray, 2016; Mahoney, 2007; Mahwinney & Lugg, 2001;
Rosenthal, 1993). Considering the growing work on teachers’ unions’ role in local education
policy (e.g. Cowen & Strunk, 2015; Marianno & Strunk, 2018a; Moe, 2009; Strunk, 2011, 2012),
the extant literature on teachers’ unions’ influence on state and federal policy is relatively small
and has largely failed to consider the role of competing interests in shaping policy change.
Prior Research on Teachers’ Union and State and Federal Policy-Setting
Only a few empirical studies specifically focus on teachers’ unions’ influence on policy-
setting (Finger, 2017; Giersch, 2014; Hartney & Flavin, 2011; Renzulli & Roscigno, 2005;
Shober, Manna, & Witte, 2006; Stoddard & Corcoran, 2007). The bulk of this research
concentrates on the spread and content of charter school policies (Giersch, 2014; Renzulli &
Roscigno, 2005; Shober, Manna, & Witte, 2006; Stoddard & Corcoran, 2007). These studies
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 38

collectively demonstrate that teachers’ union power (as measured by union membership rates), is
associated with both the enactment and content of charter school legislation. Stoddard and
Corcoran (2007) establish that states with larger shares of unionized teachers are significantly
less likely to enact a charter school law, and when they did enact such legislation, the laws were
much weaker. This latter result replicates the findings of both Renzulli and Roscigno (2005) and
Shober, Manna, and Witte (2006).
Two studies move beyond a focus simply on charter school legislation to look at the
relationship between teachers’ union power and other education reform policies. Hartney and
Flavin (2011) operationalize teachers’ union influence as the proportion of campaign
contributions from teachers’ unions to candidates for public office relative to the total amount of
donations in a state. They find that states where teachers’ unions donate a higher proportion of
contributions to candidates are less likely to have reform-oriented school choice, teacher
evaluation, and teacher pay laws. In perhaps the most comprehensive study on the subject,
Finger (2017) uses principal component analysis to generate a scale score of union influence
based on the presence of public-sector bargaining and right-to-work laws and public-sector union
membership rates. Using a dataset of charter, voucher, and performance pay policies in an event
history framework, she finds that teachers’ union strength is associated with a decrease in the
likelihood of performance pay and charter school law adoption but is not significantly associated
with the passage of voucher legislation. She concludes that teachers’ unions “play a crucial role
in the diffusion of policy change” especially regarding policies that threaten them
organizationally (Finger, 2017, p. 16).

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 39

The Current Study
While prior work provides some evidence that teachers’ unions might matter for
education policy-setting, shortcomings in data, measurement, and identification make it difficult
to draw any definitive conclusions from this analysis. In particular, prior work fails to properly
consider the role of competing interest groups in explaining patterns in teachers’ union policy
success and failure. For example, Finger (2017) notes that while she captures whether an
opposition group is present in a state, she is unable to measure the intensity with which the group
operates and how that influences policy outcomes. As argued above, new interest groups might
represent an important constraint on teachers’ union policy power as the landscape of education
policy-setting changes, but to document their effects, we need to capture their political activity
relative to that of teachers’ unions.
Second, past research mainly conceptualizes union strength as a function of their
membership size. While this measure captures an element of interest group power, scholars have
called for measures that better encapsulate what interest groups do to influence policy (Moe,
2005; Pierson, 2015). With limited measures, we may actually underestimate the influence of
unions on policy. Similarly, most studies measure union power and legislative outcomes at a
single point in time. Thus, we are unable to tell, more generally, how teachers’ union power is
changing over time and how these changes are associated with legislative outcomes.
Finally, prior work mainly looks at the enactment of legislation while paying little
attention to agenda-setting, or the types of policies proposed. Kingdon (2003) notes, however,
that agenda- and policy-setting are governed by different processes. He defines the agenda as the
“list of subjects or problems to which governmental officials…are paying some serious attention
at any given time (p. 3).” Proposed policies in the legislature more closely align with Kingdon’s
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 40

classification of the “decision agenda,” or the policy subjects within the governmental agenda
that are actively being considered for new legislation. He argues that interest groups and policy
advocates succeed at getting a policy onto the decision agenda by linking their policy initiative to
a policy problem that lawmakers are pressed to solve. This does not ensure, however, that a
policy will be enacted. Moe (2011) notes, that the policy process is setup to make it very difficult
for minority interest groups to see that their policy ideas move beyond the decision agenda to
final adoption. New laws challenging the majority interest group and policy status quo must pass
through multiple committees that could sink the proposal with a single vote—and majority
interest groups will work hard to achieve that outcome. Because it is impossible to distinguish
between the role that interest group power plays in the agenda-setting versus policy enactment
process when looking simply at enacted legislation, existing studies focused on policy adoption
may, in fact, miss the efforts of minority interest groups to challenge union power. There is a
need to look at proposed policies as well.
I build on prior work by using a unique dataset of all proposed and enacted policies
related to 21 different teacher policy topics across five legislative sessions (2011 to 2015) in all
50 states. I propose a new way to capture state teachers’ union power that conceptualizes union
influence as a function of teachers’ union influence on three components relative to other interest
groups: their membership and revenue raising power, their election influence, and their lobbying
influence. I then estimate a series of within-state fixed effects models that account for fixed
unobserved heterogeneity in the state policy environment and estimate teachers’ union influence
on both agenda- and policy-setting outcomes.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 41

Data
State Teacher Policy Database
Scholars have made several efforts to collect and classify public policies over time,
though most of the existing datasets contain national level policies. For example, the
Comparative Agendas Project tracks new laws across 23 different policies areas (e.g. civil rights,
health, agriculture, labor, education) and 20 countries. Some non-profit and advocacy
organizations have made smaller attempts to track state policies in specific topic areas including
charter school laws (Center for Education Reform, 2017), energy policy (American Council for
an Energy Efficient Economy, 2018), college and career readiness and health policies (National
Association of State Boards of Education, 2018), collective bargaining laws (National
Conference on State Legislatures, 2018) and higher education policies (Western Interstate
Commission for Higher Education, 2018), to name a few.
Education Commission of the States (ECS) engages in the most comprehensive effort to
track education policy, following all enacted bills in 31 different education policy areas across all
50 states. The database used in this study has several advantages over existing education policy
databases. First, drawing on the setup of the ECS database, I collect all the same information that
ECS does on 21 different teacher policy topics (i.e. bill number, title, status, last action date, link
to bill text, sponsor, and bill summary). While ECS collects bills on more education policy
topics, my database includes more topics specific to teachers and their unions—for example,
collective bargaining, right-to-work/membership dues, teacher strikes, and teacher retirement
(the 21 topics are listed in Appendix Table 1)—and does so across five legislative cycles (2011
to 2015).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 42

ECS also only provides information on enacted policies versus those that were proposed
but defeated during the legislative process. Tracking proposed policies requires a more detailed
search of state legislative archives. I use the National Conference of State Legislatures (NCSL)
collective bargaining law database and the Lexis Nexis State Capital archives to find both
proposed and enacted teacher policies related to the 21 different topic areas. I used broad search
filters in these databases to ensure that I did not miss relevant laws. An initial search in the
NCSL database for collective bargaining laws and Lexis Nexis State Capital for bills related to
“education agencies and personnel” revealed a total of 51,304 law summaries. I read the
combined total of law summaries and categorized each relevant law based on based on their
topic (e.g. collective bargaining). Because of the broad search terms used in locating legislation,
a large portion of the law records were not relevant for the database, either because they did not
pertain to teachers’ unions or teachers or were duplicative of other laws proposed in a given
legislative session.
6
This lead to a final total of 3,944 relevant proposed and enacted bills
between 2011 and 2015. Specifically, to ensure that no enacted laws were missed, I compared
my database of enacted bills to the ECS database in the coinciding policy areas (e.g. teacher
evaluation) to ensure 100 percent overlap.
Finally, of upmost importance to this study is how proposed laws would change or how
enacted laws did change the rights of teachers and their unions. Consequently, I coded legislation
not only based on a broad topic but also on their “intent.” While other studies have made efforts
to code the “topic” (e.g. Worsham, 2006) and “flexibility” of public policies (e.g. Renzulli &
Roscigno, 2005; Shober, Manna, and Witte, 2006; Stoddard & Corcoran, 2007) or developed

6
Duplicate laws were mainly derived from some legislatures that employ the practice of introducing identical bills
in two chambers in the same year or by assigning revised legislation a new bill number. In order to avoid double-
counting laws, bills that were identical or nearly identical in wording in the same year were removed from the
database for the final analysis. In total, 511 duplicate bills were removed from the database.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 43

typologies to categorize laws based on whether they are “redistributive” or “developmental”
(Hwang and Gray, 1991; McFarlane & Meier, 2001), or restrictive of competition for public
office (Mulligan, Gil, & Sala-i-Martin, 2004), classifications based on the intent of a law towards
a dominant interest group, and the policy positions they espouse, do not exist in the current
extant literature.
I created a new typology that classifies laws based on whether proposed and enacted
teacher policies where “unfavorable”, “favorable”, or “neutral” towards teachers and their
unions. Laws were determined to be unfavorable if they constrained the rights of teachers’
unions or the scope, impact, or coverage of bargaining agreements, or if they were antithetical to
traditional union policy positions. Laws were categorized as neutral if they pertained to the rights
of teachers’ unions or the scope of collective bargaining agreements but made no substantive
changes. Laws were considered favorable if they recognized, preserved, or enhanced the rights of
teachers’ unions or the scope, impact, or coverage of collective bargaining agreements, or if they
were in alignment with traditional union policy positions. I determined traditional teachers’
union policy positions based on the stated stances on legislation of the NEA and its state
affiliates, made available in the “Issues and Action” section of their websites.
Table 4 provides some examples on how I classified legislation. For example, S.B. 441
(2013) in California, if enacted, would require permanent (tenured) teachers to be evaluated on a
three-year cycle, would remove the evaluation from the scope of collective bargaining, and
would require teachers to be evaluated on the basis of student test scores. CTA opposed the bill,
arguing that it would have “silenced teachers' voices on…evaluation” (California Teachers’
Association, 2013). Given the union’s opposition to the evaluation bill, I coded this legislation as
“unfavorable.” In contrast, S.B. 1458 (2013) in Texas increased the state contribution to
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 44

teachers’ retirement, provided a cost-of-living adjustment (COLA), and grandfathered in active
teachers into the state health plan if they had at least five years in the system. The Texas State
Teachers’ Association supported this bill, arguing that it “should help secure the defined benefit
plan for Teacher Retirement System members (Texas State Teachers’ Association, 2013).”
Consequently, I coded this bill as “favorable.” Because not all state union affiliates provide their
official stances on their websites, in cases where official union positions were not available, I
relied on NEA state affiliate positions on similar legislation in other states.
Policy Outcome Measures
First, I generate measures that are intended to capture the various policy ideas under
consideration by state legislatures or the extent to which the decision agenda, as defined by
Kingdon (2003), actively aligns with teachers’ union interests. I operationalize these measures as
the relative “favorableness” or “unfavorableness” of proposed and enacted policies in a given
legislative session as shown in equations (1) and (2). I use proportional rather than absolute
number measures because the latter would confound other state characteristics that may be
related to the sheer number of bills under consideration, like the size of the state, the length of
the legislative session, or the size of the legislative chambers. Furthermore, the proportional
measures provide a better indication of the relative intent of the policy agenda in a state towards
teachers and unions.

Prop. Favorable Proposed
st
=
# Favorable laws proposed
st
# total laws proposed
st

Prop. Unfavorable Proposed
st
=
# Unfavorable laws proposed
st
# total laws proposed
st

(1)
(2)
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 45

Table 5 provides summary statistics for the agenda-setting measures. Column (1)
indicates that the height of unfavorable legislative proposals occurred in 2011 when the
proportion of unfavorable legislative proposals in the average state was at 55 percent. The
proportion of unfavorable proposals decreased over time, reaching as low as 31 percent in 2014,
before climbing again to 44 percent in 2014.
7
Not surprising, the proportion of enabling laws
proposed follows an opposite trend, starting as low as 38 percent in 2011 and reaching as high as
55 percent in 2015.
To measure policy adoption, I follow studies from the legislative productivity literature
(e.g. Edwards, Barrett, & Peake, 1997; Krutz, 2000) to generate a series of policy success and
failure rates. In terms of policy outcomes, teachers’ unions are concerned with ensuring that their
favored policies are enacted and that damaging policies are defeated. Consequently, the
unfavorable bill failure rate is simply defined as the number of unfavorable policies enacted
divided by the number of unfavorable bills proposed (equation (3)). Similarly, the favorable bill
success rate is defined as the number of favorable policies enacted divided by the number of
favorable bills proposed (equation (4)).
Unfavorable Failure Rate
st
=
# Unfavorable laws failed
st
# Total unfavorable laws proposed
st

Favorable Success Rate
st
=
# Favorable laws enacted
st
# Total Favorable laws proposed
st

7
Adding the proportion of unfavorable laws and the proportion of favorable laws within a given year in Table 5
does not yield 100 percent because of the present of neutral laws, which are not analyzed in this chapter and which
do not appear in Table 5. Laws were categorized as “neutral” if they pertained to the rights of teachers’ unions or the
scope of collective bargaining agreements but made no substantive changes.
(3)
(4)
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 46

Table 5 shows the state policy success rates over time. In general, the failure rate is much
higher for unfavorable legislation than the success rate is for favorable legislation, consistent
with theoretical observations that the policy system is built in a way to ensure that blocking
legislation is much easier than taking actions to see that a bill is enacted (Moe, 2011, 2015).
8
The
success rate of unfavorable legislation reached its highest in 2013 at 79 percent whereas the
success rate of favorable legislation reached its highest in 2014 at 27 percent.
Relative Union Influence Measures: Membership/Resources, Election, and Lobbying
Power
Most existing studies employ unidimensional measures of teachers’ union power, like
membership rates and campaign contributions, without capturing the full breadth of how
teachers’ unions engage in policymaking and without considering the relative power of unions to
other groups in the policy subsystem (de Figueiredo & Richter, 2014; Lowery, 2013). Prior
research suggests that interest groups influence policy through three main avenues—elections
(Ansolabehere, de Figueiredo, & Snyder, 2003; Grossman & Helpman, 1994; 1996), lobbying
(de Figueiredo & Richter, 2014), and membership organizing (Leighley, 1996; Leighley and
Nagler, 2007). In this chapter, I argue that teachers’ unions will exert more pressure on policy
where they exercise more election, lobbying, and membership organizing influence relative to
opposing interest groups. Put another way, teachers’ unions will be less impactful in states where
their strength is countered with the strength of competing interests (Gray & Lowery, 1999;
Hansen & Gray, 2016; Mahoney, 2007; Mahwinney & Lugg, 2001; Rosenthal, 1993).

8
For example, Moe (2015) suggests that in order to see that a bill is enacted, policy reformers must ensure that their
bills pass multiple phases of the policy process, which includes several committee meetings and votes in both
legislative chambers. In contrast, a single committee or floor vote can stifle a bill—interest groups looking to block
unfavorable legislation must simply sway one vote.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 47

With the surge in education-specific groups in recent years, a number of organizations
may now challenge teachers’ unions’ policy-setting efforts. In this study, I define opposition
interests as school choice advocacy organizations and business groups. First, school choice
advocates include 116 organizations identified by the National Institute on Money in State
Politics as groups supportive of charter school and voucher expansion. This list includes some of
the most active education advocacy organizations involved in politics, including Democrats for
Education Reform, 50Can, Stand for Children, StudentsFirst, Education Reform Now, and
Families for Excellent Schools (Sawchuk, 2012). I hand-checked this list for completeness to
ensure that the largest school choice groups in each state were in the data. With the increasing
involvement of business groups and leaders in schools (Cibulka, 2001, Henig, 2013), I also
include as opposition groups over 40,000 business organizations across all 50 states. In what
follows, I describe the different components of teachers’ union relative strength and the
corresponding data sources (summarized in Table 6). I then describe how I construct the power
index for each component using principal components analysis (PCA).
Election Influence. Policy and political science theorist have long hypothesized that
special interest groups have influence on policy through electoral politics. Special interest groups
provide campaign contributions with the objective of maximizing election chances of their
favored candidates and the possibility of receiving a return on that investment in roll call votes
on legislation related to their interests (Ansolabehere, de Figueiredo, & Snyder, 2003; Grossman
& Helpman, 1994; 1996; Snyder, 1990, 1991). Contributions also serve to grant interest groups
access to lawmakers (Hall & Wayman, 1990; Kalla & Broockman, 2016), after which groups can
subsidize the informational resources of policymakers (Austen-Smith, 1995; Esterling, 2007;
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 48

Hall & Deardorff, 2006; Lohmann, 1995) and persuade them that their “electoral self-interest lies
in taking group-friendly positions (Hall & Deardorff, 2006, p. 71).”
I generate an overall election influence score from three measures of interest group
political involvement. First, borrowing from Hartney and Flavin (2011), I create a measure of
what they call “political activism” by dividing the number of campaign contributions from a
specific group in a state by the total amount of campaign contributions given to candidates for
state legislative offices. This measure has the attractive feature of reflecting the donating strength
of an organized interest relative to other interest groups in the state (e.g. a higher proportion of
contributions from teachers’ unions should reflect more influence from unions relative to other
actors). However, this measure fails to capture how much money interest groups donate to
candidates for public office (in absolute terms) and how often interest group- supported
candidates are elected. Consequently, I generate two additional measures of election influence—
the amount of contributions donated by teachers’ unions or opposition groups per candidate for
legislative office (logged for use in the measure) and the proportion of teachers’ union-supported
or opposition group-supported candidates elected. All data on contributions and election
outcomes were derived from the National Institute on Money in State Politics’ database on
political spending in state politics and only include contributions and outcomes for state house
and senate races. Because my outcome measures (agenda- and policy-setting) are measured
yearly and the election influence variables (i.e. win rates, campaign contributions) are only
available in years in which state elections are held (which, in most states, is every other year),
values for the election influence variables are copied for subsequent years until the next state
election is held.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 49

The summary statistics for the election influence variables shown in Table 7 provide
some indication for how teachers’ union election power might be changing over time relative to
other interest groups. In 2010, 45 percent of open state legislature seats were won by teachers’
union allies but this number dropped to 37 percent by 2014. Comparatively, the proportion of
seats won by business allies has remained relatively steady over time and the proportion of seats
won by school choice advocates has increased over the same time period (from 7 percent to 14
percent). Notwithstanding losing ground in the proportion of allies in state legislatures, teachers’
unions are donating more money to election campaigns—from 1,423 dollars per candidate in
2010 to 1,885 dollars per candidate in 2014. However, business groups and school choice
advocates also increased their contributions on a per candidate basis from 2010 to 2014.
Lobbying Influence. Separate from the resources donated to election campaigns, special
interest groups also devote a considerable amount of money towards direct lobbying efforts (de
Figueiredo & Richter, 2014). In this case, monetary contributions are not transferred to
lawmakers, but are instead employed in paying individuals to connect with and provide
information on a given issue, with the ultimate objective of swaying policy proposals and votes
(de Figueiredo, 2002). Consequently, employed lobbyists frequently target lawmakers in key
positions of power during the policymaking process, including politicians sponsoring or co-
sponsoring bills (Hojnacki & Kimball, 1998), sitting on key committees (Drope & Hansen, 2004;
Hojnacki & Kimball 2001), or serving as minority and majority party leaders (Evans, 1996).
Lobbyists often strategically target lawmakers allied with their interests as a mechanism for
subsidizing the amount of resources devoted to their favored causes (Hall & Deardorff, 2006);
however, there is some debate as to whether lobbyists also target rivals in order to prevent or
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 50

block unfavorable legislative pushes (Austen-Smith & Wright, 1994, 1996; Baumgartner &
Leech, 1996).
Despite the importance of lobbying to public policy outcomes, specific measures of
teachers’ union or opposition group lobbying efforts are largely absent from the extant literature
on teachers’ union political influence. New data from the National Institute on Money in State
Politics allows researchers to track the number of teachers’ unions and opposition groups
employing lobbyists and the specific number of lobbyists employed per teachers’
union/opposition group in a state. I generate indicators of teachers’ union and opposition group
lobbying influence by building on Gray and Lowery’s (1993, 1995) concept of interest group
density—I generate a measure of interest group lobbying density by dividing the total number of
teachers’ unions or opposition groups actively registered to lobby in a state by the total number
of organizations registered to lobby. Because a single organization may maintain multiple
lobbyist within their organization, I generate a second indicator of lobbying influence defined as
the number of teachers’ union or opposition group lobbyist employed per lawmaker in a state.
Summary statistics for all lobbying influence variables are shown in Table 7. While
teachers’ unions are employing a similar ratio of lobbyist to lawmakers over time, the proportion
of lobbying organizations that are registered to teachers’ unions has decreased slightly (from
0.64 percent to 0.510 percent). In contrast, school choice and business groups have increased
their lobbying presence from 2010 to 2014.
Membership Organizing. Finally, special interest groups, particularly labor unions,
often derive policy influence through the sheer size of their membership. Large membership
bases afford interest groups the opportunity to pursue influence by mobilizing their membership
in grass roots efforts to shape election campaigns and to influence legislative proposals
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 51

(Leighley, 1996; Leighley and Nagler, 2007; Verba, Scholzman, & Brady, 1995). Prior research
suggest that union membership is strongly, and positively associated with voter turnout in local,
state, and national elections (Leighley and Nagler, 2007; Moe, 2005; Moe, 2006b; Moe, 2011;
Radcliff, 2001; Radcliff and Davis, 2000). While union voter turnout might be more influential
in local elections, where there are fewer competing interests (as is the case with teachers’ unions
and school board elections (Moe, 2005; Moe, 2006b; Moe, 2011), unions still exert considerable
effort in identifying, endorsing, and rallying for their favored candidates in state and federal
elections. These endorsements serve as informational cues for both members and nonmembers
regarding the preferences and political orientation of the candidate (McDermott, 2006). Beyond
grass roots campaigns, interest groups with larger memberships are able to derive more resources
from the collection of membership dues which they can expend on building a stronger
organization.
Because school choice and business groups are not membership organizations, the
membership organizing measure is only available for teachers’ unions. To measure the size of
the teachers’ union membership, I draw upon data compiled from National Education
Association (NEA) reports by Mike Antonucci (Antonucci, 2017). Teachers’ union membership
rates are defined as the number of NEA members divided by the number of full time teachers
employed in a state. To define collective bargaining coverage, I employ data compiled by Hirsch
and Macpherson (2015) who use information from the Current Population Survey (CPS) to track
the percentage of public employees covered by a collective bargaining agreement in each state
from 1973 to 2017.
Teachers’ unions with a larger membership are also better able to raise revenue for their
organizational activities, including collective bargaining, legal and grievance proceedings,
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 52

employee salaries, health and retirement benefits, and political activities. The primary source of
this revenue is membership dues. Consequently, following Lott and Kenny (2013), I generate a
measure of the amount of dues revenue the state teachers’ union raises per teacher and the
amount of money they expend per student (both logged for use in the measure). The IRS requires
non-profit organizations to report their membership dues revenue and their total expenditures on
their 990 tax forms which are publicly available.
Summary statistics for the membership/resources variables (shown in Table 7) reveal
that while teachers’ unions’ dues revenue and expenditures have remained relatively stable over
time, teachers’ unions are losing membership. The proportion of teachers maintaining
membership in the union dropped over the duration of the panel. This is consistent with national
trends in teachers’ union membership over time (Marianno & Strunk, 2018b).
Component Creation. Following (Finger, 2017), I first use principal component analysis
to generate measures of the different components of teachers’ union and opposition group
influence. PCA is typically employed as a data reduction technique when faced with multiple
variables that are highly correlated (Julnes, 1999). The objective of PCA is to reduce the number
of variables to a smaller number of components that maximize the variation in the data.
9

Appendix Table 2A and 2B document the correlations between union influence and opposition
group influence variables within each component and over time.
10

Table 8, Panel A shows the results from the PCA run on all nine variables of union
strength in 2011. The results suggest that the first component captures the most variation in the
data at 43 percent, although the first three components all have eigenvalues over one.

9
PCA is separate and distinct from exploratory factor analysis, which is employed to discover the number of latent
constructs from an underlying dataset
10
Only the correlations in 2010 and 2011 are shown. A full correlation table with all years is available upon request.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 53

Components two and three explain an additional 21.90 percent and 14.0 percent of the variation
in the nine variables, and all three components collectively explain 78.90 percent of the
variation.
The four teachers’ union membership/resource variables all load highly on component
one with correlations of 0.835 or higher. The three election influence variables load the strongest
on component two with correlations at 0.592 or higher and the lobbying influence variables load
the highest on component three with correlations of 0.702 or higher. When looking at all the
variables on component one, teachers’ union lobbying density loads the lowest at 0.140, while all
other variables load at 0.388 or higher.
11

To generate a measure of overall union strength, I predict the score on component one for
each state in each year. PCA is done on each year separately to handle the non-independence of
the observations in the panel data structure. Consequently, as the measure of overall union
strength is standardized within year, such that the mean in each year is zero with a standard
deviation of one. The measure of overall teachers’ union strength is correlated with the measure
created by Winkler, Scull, and Zeehandelaar (2012) at 0.857.
I then re-run PCA on the variables for each component separately. In short, the same
process is followed for the four membership/resource variables (Table 8, Panel B), the three
election influence variables (Panel C), and the two lobbying influence variables (Panel D) such
that I predict a unique score for each component within each year. As with the overall teachers’
union strength variable, these variables are also standardized to have a mean of zero and a
standard deviation of one. Table 9 demonstrates that the membership/resources component is
correlated with the overall teachers’ union strength measure at 0.952. The election and lobbying

11
Table 8 presents the unrotated solutions. Given the strong patterns demonstrated in the unrotated solutions, the
rotated solutions did not add additional clarity.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 54

influence measures are not as highly correlated with the overall teachers’ union strength measure
(at 0.494 and 0.299, respectively). This is largely a result of the higher correlations of the
membership/resource variables with component one in the initial PCA analysis when compared
to the election and lobbying influence variables.
Like the creation of the teachers’ union influence measures, I employ principal
component analysis to generate three different measures of opposition group influence—overall
opposition group strength, opposition group election influence, and opposition group lobbying
influence. Results from the PCA analysis are shown in Appendix Table 3. I first create a measure
of overall opposition group strength by predicting the score on component one using all
opposition strength variables. I subsequently create measures of election and lobbying influence
by predicting the score on component one using only the election-specific and lobbying-specific
variables. Like the teachers’ union strength measures, these variables are standardized within
year to have a mean of zero and a standard deviation of one.
Measure of Relative Union Influence. To generate a specific measure of teachers’
unions’ influence relative to the influence of opposition groups, I take the difference between the
overall union influence measure and the overall opposition group influence measure within each
year. Because both measures are standardized within year, they are on comparable scales. This
measure has the attractive property of indicating the relative balance of power between teachers’
unions and their opposition in a state. I create similar measures for teachers’ union relative
election and lobbying influence by subtracting the opposition group election and lobbying
scores. These measures are all standardized within year to have a mean of zero and a standard
deviation of one.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 55

Table 10 shows the states ranked by the relative strength of their teachers’ union to the
strength of the opposition groups in the state. My relative group strength rankings (column 1) are
correlated with the Winkler, Scull, and Zeehandelaar (2012) state teachers’ union rankings
(column 4) at 0.697 and my state teachers’ union strength rankings (column 5) at 0.665. If strong
teachers’ union states also had weak opposition groups then we would expect a near perfect
negative correlation between the relative strength rankings and the relative opposition group
rankings (since lower numbers indicate stronger interest groups and higher numbers indicate
weaker interest groups). The moderate positive correlation suggests some variation in the
strength of state teachers’ unions and opposition groups. The rankings make clear that in some
states teachers’ unions are much stronger, in some states they are equally matched, and in some
states they are much weaker than their opposition. For example, the top ranked state is
Minnesota, which, as revealed by its overall teachers’ union strength rank, does not have the
strongest state teachers’ union in the nation (8
th
) but its opposition groups are ranked 45
th
,
suggesting a large power imbalance. In contrast, New York (ranked 26
st
) has the second
strongest teachers’ union in the nation but the sixth strongest opposition groups. Georgia has the
weakest state teachers’ union in the nation (50
th
), but its opposition groups are ranked as the 10
th

strongest. I would expect teachers’ unions in Minnesota to be much more successful in obtaining
their state-level policy priorities than teachers’ unions in Georgia, and quite possibly, even New
York.
Other Independent Variables
I control for several other variables that may make the proposal, passage, and failure of
state teacher legislation more likely. Using data from the National Conference of State
Legislatures, I account for the partisan control of the legislature. Indicators for a Republican-
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 56

controlled legislature and a split legislature are placed in the models with Democrat controlled
legislatures as the reference category.
Governments in financial trouble may also exert more pressure to alter teachers’ union
collective bargaining rights, teacher salaries, pension plans, and other policies that tie up state
dollars. Consequently, using data from the state financial census, I generate a measure of the
debt-to-service ratio (the amount of state debt relative to amount of state revenue), which
provides some indication of the ability of a state to cover their outstanding debts. I also control
for the economic health of a state by including a measure of the state unemployment rate.
State legislatures striving to improve education performance may also be more likely to
tackle teachers’ union rights and implement new education reforms. I control for the average
National Assessment of Educational Progress (NAEP) math score for each state in the models.
Because the exam is only administered every two years, I replace the missing year scores with
the midpoint score between the year prior and after. Using data from the National Council for
Education Statistics, I also control for the size of the K-12 student population in the state (natural
logged). This measure captures the size of the K-12 student constituency in a state.
Finally, I include a few controls to account for the unique dynamics of certain state
legislatures. First, I control for differences in state legislative session lengths as longer sessions
allow for more time to propose and enact legislation. I also include a control for if the state
proposed any relevant law in a given year. Because the agenda- and policy-setting variables are
proportional to the number of laws proposed, states that do not propose any legislation in a year
would be missing on all outcome variables. To include these states in the models, I recode these
missing states to equal zero and include a dichotomous variable in the model that equals one if
the state did not propose any relevant legislation. As Table 5 indicates, this ranges between 0
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 57

percent to 12 percent (6 states) of states in any given year. Finally, the extant literature has long
documented policy diffusion or spillover effects across states (e.g. Berry & Berry, 1990;
Nicholson-Crotty & Carley, 2016). To account for policy spillover, I include controls for the
agenda- and policy-setting variables in neighboring states.
Analytic Strategy
To identify the relationship between agenda- and policy- setting and teachers’ union
relative strength, I employ a within-state fixed effect model. This model has the advantage of
controlling for time-invariant unobserved heterogeneity between states that may shape policy
proposals or union strength. The model is estimated as shown in equation (6).
Y
st
= β
0
+β
1
X
st-1
+ S
st-1
β
2
+δ
j
+τ
t
+ε
st
(6)
The outcome variable Y
st
reflects the proportion of teacher policies proposed (agenda-setting) in
a given year or the policy success/failure rate (policy-setting). β
1
is the estimate of the
relationship between teachers’ union relative strength, indicated by X
st-1
, in year t-1, on Y
st
,
conditional Sst-1, a vector of time-varying state characteristics in year t-1, δ
j
, a state fixed effect,
and 𝜏 𝑡 , a year fixed effect.
12
If conditional on Sst-1, δ
j
, and τ
t
, X
st-1
is independent from ε
st
, then
β
1
will identify the causal effect of interest—the effect of teachers’ union relative influence on
state agenda- and policy-setting. It’s important to note however, that this assumption may not
hold. There are likely important unobserved time-varying variables that are related to state policy
and teachers’ union relative influence that are not accounted for in this model. Consequently, I
do not draw causal conclusions from the results presented in the next section.

12
Variables are lagged on year under the assumption that policymakers in year t are likely responding to conditions
in t-1 when considering new legislation. Results using contemporaneous controls are substantively similar and are
available upon request.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 58

Results
Table 11 shows the results from equation (6) predicting the proportion of unfavorable
laws (columns (1), (3), and (5)) and favorable laws (columns (2), (4), and (6)) proposed. I first
present the results for the measure capturing teachers’ unions’ relative strength to opposing
interest groups (columns (1) and (2)) and then for teachers’ union relative election and lobbying
strength to opposing interest groups (columns 3 through 6). Because of the repeated state
observations over time, standard errors are clustered at the state-level.
Consistent with expectations, Table 11 demonstrates that teachers’ unions experience
more agenda-setting success in states where they exert more influence relative to opposing
groups. In particular, column (1) shows that a one standard deviation increase in teachers’
unions’ influence relative to opposing groups is associated with a 5 percent decrease in the
proportion of unfavorable laws proposed (p<.0.05) and a 6 percent increase in the proportion of
favorable laws proposed (p<.0.10). These effects are slightly larger when looking specifically at
election influence (columns (3) and (4). A one standard deviation in teachers’ unions’ relative
election influence is associated with a 6 percent decrease in the proportion of unfavorable laws
proposed (p<.0.10) and a 7 percent increase in the proportion of favorable laws proposed
(p<.0.10). The effects are smaller for lobbying influence and not statistically significant, but the
signs on the coefficients are in the same direction (columns (5) and (6)).
Table 12 shows the results from equation (6) predicting unfavorable law failure rates
(columns (1), (3), and (5)), and favorable law success rates (columns (2), (4), and (6)). Column
(1) shows that a one standard deviation increase in teachers’ unions’ influence relative to
opposing groups is associated with a 17 percent increase in the failure rate of unfavorable teacher
policies (p<.0.01), which is consistent with theoretical observation that majority interest groups
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 59

are much more influential in protecting the status quo (blocking unfavorable policies) versus
ensuring that favorable policies are enacted (Moe, 2011, 2015). The effect of the overall relative
influence of teachers’ unions on the success rate of favorable legislation is not statistically
significant but is positive in direction. I do find that relative election influence is significantly
and positively related to the failure rate of unfavorable legislation and the success rate of
favorable legislation (at 11 percent (p<.0.01) and 7 percent, respectively (p<.0.10)). I also find
that the effects are smaller for lobbying influence and not statistically significant, but the signs
on the coefficients are in the same direction as the overall and election influence effects
(columns (5) and (6)).
Interestingly, I find very few consistent predictors of the proposal or outcome of state
legislation among the other state control variables. In short, the only variable that seems to
consistently matter is the degree of policy-setting activity in neighboring states, but only with
regard to the failure rate of unfavorable legislation. It appears that as the failure rate of
legislation in bordering states increases, the less likely a state is to reject unfavorable legislation.
This finding runs contrary to the notion that states mimic the policy responses of their neighbors
(e.g. Berry & Berry, 1990; Nicholson-Crotty & Carley, 2016). Instead, it appears that
unfavorable legislation has a stronger push in states bordering other locales where unfavorable
policies have been less successful.
Absolute influence measures. To explore whether the results are unique to the relative
influence measures versus absolute measures of teachers’ union power in a state, I run the same
regressions using the overall, election, and lobbying teachers’ union influence measures. The
results from these models are shown in the top panel of Table 13. Across all measures of
absolute union strength, I find no significant relationship with the proportion of favorable and
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 60

unfavorable laws proposed (columns (1) and (2)). However, when looking at the results
predicting legislative failure and success rates (shown in columns (4) and (5)), I find very similar
results to those shown in Table 12 using the relative influence measures—a one standard
deviation increased in the overall teachers’ union strength is associated with an 18 percent
increase in the failure rate of unfavorable legislation (p<0.01). Furthermore, the effects of
election influence on the failure rate of unfavorable legislation and the success rate of favorable
legislation are both positive and statistically significant as they are in Table 12.
13
Given the
comparability in the results across the different measures predicting policy outcomes, I am
unable to conclude that there is something about teachers’ union relative influence that is
uniquely associated with final policy decisions. Instead, it appears that relative influence is only
distinctly associated with the proposal of policies. This is consistent with the idea that even
minority interest groups can engage in agenda-setting activities if they are able to garner
sufficient resources to take advantage of open policy windows (Kingdon, 2003). They must
simply persuade an allied lawmaker to propose new legislation. However, while opposing
coalitions to teachers’ unions are strong enough to influence the agenda-setting process, they
may find it much more difficult to muster enough coalition force to see that the policies are
enacted, especially where teachers’ unions are at their strongest (Moe, 2011, 2015).
Top loading variables. To see if the results are robust to using PCA versus single
indicators of the different dimensions of union absolute and relative strength, I run the analysis
again, this time using the top loading variable for each dimension of interest group power from
PCA. These results are shown in the second and third panels of Table 13. The results from this

13
I also find directionally similar results when I use the Winkler, Scull, & Zeehandelaar (2012) overall teachers’
union strength rankings though the coefficients are smaller in magnitude and not statistically significant. This is
likely due to the fact that I am only able to use a single cross-section of data when modeling the proposal and
enactment of legislation as a function of the Winkler et al. rankings.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 61

analysis are substantively similar with a few exceptions. First, I find that teachers’ union
membership dues per teacher, as an indicator of overall teachers’ union strength, is positively
associated with the proportion of unfavorable laws proposed and negatively associated with the
proportion of favorable laws proposed. While these effects are in the same direction as those
found with the overall influence measure in the top panel of Table 13, they are much larger in
magnitude and are statistically significant. It is entirely possible, then, if this result is taken at
face value, that opposition coalitions are targeting their legislative proposals in locations where
teachers’ unions are at their strongest, at least in terms of their ability to raise revenues from
membership dues. Second, while I find no significant effect of lobbying influence on the
proposal or enactment of teacher legislation, I do find that the number of teachers’ union lobbyist
employed per lawmaker is significantly and positively associated with the failure rate of
unfavorable legislation.
Conclusion
The political rhetoric surrounding teachers’ unions often suggests that they are one of the
most, if not the most powerful special interest groups in public education (e.g. Burnette, 2016b;
Layton, 2015; Walker, 2013); yet, recent shifts in the jurisdictional authority of education policy-
making coupled with the rise of new interest groups involved in education politics may
fundamentally change the way teachers’ union power shapes policymaking (Cibulka, 2001;
Gayley, 2015; Henig, 2013; McDonnell, 2013). Drawing on a unique self-collected database of
proposed and enacted teacher policies across multiple legislative cycles linked to novel measures
of teachers’ union membership/revenue, electoral, and lobbying influence, this study represents
the most expansive effort on teachers’ unions state policy-setting influence to date.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 62

The findings of this study fit with existing theories on the dynamics of agenda-setting and
policy change. First, this study demonstrates the importance of understanding how teachers’
unions and other groups exert influence at different stages in the policy process. I find that
opposition group strength—conceptualized in this chapter as election and lobbying influence
from business associations and school choice advocate groups—often works in the opposite
direction of teachers’ union strength in the agenda-setting process. In short, opposition groups
experience more agenda-setting success where teachers’ unions are at their weakest and vice
versa. Existing theory suggests that the agenda-setting process is much more amenable to
influence by both minority and majority interest groups than other stages of the policy process
(Kingdon, 2003). This is because a host of factors can shape the ideas under active consideration
in the legislature, including prevailing public opinion, academic research, media coverage, and
ideas from other policy subsystems (Sabatier, 1988; Sabatier and Jenkins-Smith, 1993; Sabatier
& Weible, 2007). Minority interest groups can seize opportunities to attach their ideas to agenda
items that have been made prominent through other policy subsystem inputs without having to
marshal significant resources (Kingdon, 2003). Consequently, even though I show descriptively
that most education-related interest groups are still weaker in absolute terms than teachers’
unions, opposing interest groups may still garner enough strength to propose new policies that
run contrary to the education status quo. However, it is important to also note that inserting new
policy ideas into state legislatures becomes even easier in areas where dominant interest groups
and coalitions are growing weaker. This is because substantial effort is made by dominant groups
to ensure that some ideas never make on to the decision agenda (Baumgartner & Jones, 1991;
2002; 2009; Kingdon, 2003).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 63

While I also find that teachers’ unions experience more favorable policy adoptions
where they are stronger than competing groups, the presented robustness checks suggest that this
is more a function of teachers’ union overall strength rather than their power in comparison to
their opposition. In short, I do not find consistent evidence that opposition groups are more
successful at seeing their policy priorities enacted where teachers’ unions are growing weaker.
This may be because the balance of power has not shifted enough to allow competing interests
and coalitions to gain influence over the policy-setting process. As noted by Moe (2011, 2015),
interest groups whose priorities already align with existing status quo policies have an advantage
in policy-setting. These groups must simply ensure new policy proposals that run contrary to
their interests ultimately fail. And given the setup of the American legislative system, with
multiple votes on proposed legislation, this is relatively easy. Only a single vote must go in their
favor. The higher success rates for teachers’ unions on unfavorable legislation versus favorable
legislation provides some credence to this idea.
Finally, the findings of this study also have implications for the ability of state lawmakers
to reform public-schooling, particularly in ways that run contrary to the traditional interests of
teachers’ unions. As both PET and ACT point out, policy change occurs as external shocks to the
policy subsystem alter the alignment of interest group coalitions (Baumgartner & Jones, 2009;
Sabatier, 1988; Sabatier and Jenkins-Smith, 1993). Although, I highlight how recent institutional
shifts in education policy-making may serve to strengthen the opportunities of competing interest
groups (e.g. Henig, 2013), the results of this study suggest that teachers’ union strength is still
one of the strongest predictors of teacher policy outcomes favorable to unions. In short, if large-
scale reforms are to occur—the types of reforms that punctuate existing policy monopolies
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 64

favored by teachers’ unions—then new advocacy groups must grow in strength and/or further
institutional shifts must occur to alter the balance of power.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 65

CHAPTER FOUR- Where Have All the Senior Teachers Gone?: Teacher Transfer
Provisions and Between-School Gaps in Teacher Experience and Student Achievement
Introduction
Researchers consistently demonstrate that urban schools serving a high proportion of
certain student subgroups (e.g. racial minorities, socioeconomically disadvantaged, low
performing) face enormous challenges in attracting and retaining experienced, high-quality,
effective teachers (e.g. Adamnson & Darling-Hammond, 2012; Jacob, 2007; Peske & Haycock,
2006). This research generally finds that the least credentialed, experienced, and effective
teachers are concentrated in poor, minority, and low performing schools (Clotfelter, Ladd, &
Vigdor, 2005; Feng, 2010; Glazerman & Max, 2011; Goldhaber, Lavery, Theobald, 2015;
Isenberg, et al., 2013; Lankford, Loeb, & Wyckoff, 2002; Sass et al., 2012).
14
Research further
suggests that these stark differences in the quality of teachers across subgroups are partly to
blame for persistent racial and income-based achievement gaps (Reardon, 2011).
One of the challenges in improving the distribution of quality teachers is that teaching
assignments are strongly correlated with teacher preferences. Research demonstrates that teacher
mobility is related to the characteristics of students in schools, including student income levels,
race, and test scores, (Hanushek, Kain, & Rivikin, 2004; Scafidi, Sjoquist, & Stinebrickner,
2007). In addition, teacher mobility is associated with teachers’ overall working conditions (Cha
& Cohen-Vogel, 2011; Loeb, Darling-Hammond, & Luczak, 2005; Weiss, 1999) and the
school’s geographic proximity and demographic similarity to a teacher’s hometown (Boyd et al.
2005; Reininger, 2012). Moreover, teacher transfer policies found in teacher collective
bargaining agreements (CBAs) often grant teachers direct selection of their teaching assignment

14
However, some new evidence suggests that in a sample of select school districts high- and low-income students
have similar chances of being taught by the most effective teachers and the least effective teachers (Isenberg, et al.
2016).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 66

based on their length of service in the school district (Ballou, 2000a, 2000b; Hess & Kelly, 2006;
Johnson & Donaldson, 2006). To the extent that teachers prefer teaching in higher performing,
more advantage schools, experienced teachers may utilize seniority-based transfer provisions to
self-select into these locations, leaving the lowest performing, most disadvantaged schools with
the least experienced teachers.
Despite the presence of seniority-based transfer policies in some of the earliest negotiated
teacher CBAs (McDonnell & Pascal, 1979) and their growth over time (Johnson, Nelson, &
Potter, 1985), they are only recently being investigated as a source of the inequitable distribution
of quality teachers to students of varying levels of disadvantage (Anzia & Moe, 2014a; Cohen-
Vogel, Feng, Osborne-Lampkin, 2013; Goldhaber, Lavery & Theobald, 2016; Koski & Horng,
2007). Studies performed in California, Washington, and Florida find mixed results on the role
of transfer provisions in explaining teacher experience gaps. Some studies find that seniority-
based transfer provisions disproportionately burden high-minority schools with inexperienced
teachers (Anzia & Moe, 2014a; Moe, 2006b), while others find that seniority-based transfer
provisions may widen gaps in experience but improve gaps in effectiveness between advantaged
and disadvantaged schools (Goldhaber, Lavery, and Theobald, 2016), and still others find no
effect of the presence of seniority-based transfer provisions on the distribution of teacher
experience (Cohen-Vogel, Feng, Osborne-Lampkin, 2013; Koski & Horng, 2007)
While providing new texture to this ongoing debate, a major shortcoming of the existing
research is that the authors only examine the relationship between transfer policies and teacher
distribution at a single point in time. This reliance on cross-sectional data eliminates the authors
abilities to rule out endogeneity concerns in their estimates that may occur due to omitted
variables related to both the presence or absence of seniority-based transfer policies and the
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 67

distribution of experienced (or high quality) teachers within a district. One potential mechanism
for isolating the relationship between transfer policies on the distribution of teachers is to exploit
within-district contract changes over time rather than solely relying on likely endogenous
between district differences in CBA policies.
Second, researchers have yet to investigate how overall school staffing requirements –
not just transfer and vacancy provisions, but all provisions governing how teachers are hired and
assigned to schools and classrooms –may affect the distribution of teachers to schools. Collective
bargaining agreements contain many provisions that govern the assignment and transfer of
teachers beyond seniority-based voluntary and involuntary transfer provisions (Cohen-Vogel &
Osborne-Lampkin, 2007). For example, CBAs often require administrators to assign within-
district candidates for new positions before they consider hiring teachers from outside of the
district or place teachers returning from leave before hiring other teachers for an open position.
This provision and others like it are often criticized for imposing rigid hiring and assignment
procedures on schools that make it difficult for administrators to staff open positions with the
most qualified teacher (Ballou, 2000a; Ballou, 2000b; Hess & West, 2006; Hill, 2006; Levin,
Mulhern, & Schunck, 2005; Moe, 2006a).
A final concern with the existing studies on the topic is the failure of many to assess the
relationship between collective bargaining provisions like seniority-based transfer provisions and
differences between the achievement of more and less privileged groups of students. This is an
important equity concern that has gone largely unaddressed, even in the two studies that examine
the relationship between transfer and vacancy provisions and average achievement. To the extent
that teacher effectiveness is correlated with teacher experience (Aaronson, Barrow, & Sander,
2007; Clotfelter, Ladd, & Vigdor, 2007; Goldhaber, 2007; Goldhaber & Brewer, 2000;
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 68

Hanushek, 1986; Harris & Sass, 2011; Rivkin et al., 2005; Papay & Kraft, 2015; Wiswall, 2013),
then seniority-based provisions may leave the most disadvantaged students with not only the
least experienced teachers, but also the least effective. Seniority-based transfer provisions could
have wide-ranging consequences for within-district student achievement gaps.
This paper takes up the issue of seniority-based teacher transfer policies, this time using
data from a unique longitudinal dataset of CBAs from school districts in California over four
successive bargaining negotiations (2005-06, 2008-09, 2011-12, 2014-15). I expand on previous
research in several ways. First, I employ a unique longitudinal dataset and utilize district, year,
and district-by-year fixed effects specifications that gain identification from the within-district
changes in seniority-based teacher transfer policies over time. Second, I not only replicate the
seniority-based transfer measures utilized in prior research but also introduce a new measure of
transfer and vacancy “restrictiveness” that captures the relative degree of regulation around how
teachers are hired and assigned in California school districts. Finally, I explore the implications
of seniority-based and restrictive teacher transfer provisions for within-district achievement gaps.
Overall, I find no evidence that seniority-based transfer provisions or the overall
restrictiveness of transfer and vacancy provisions exacerbate gaps in the distribution of teacher
experience or teacher education over time in California. However, I do find that the relationship
between student achievement and the percentage of minority students in a school grows more
negative as district transfer and vacancy provisions increase in restrictiveness; however, once
accounting for the overall restrictiveness of the CBAs, this effect is no longer statistically
significant, suggesting that the observed relationship between transfer and vacancy
restrictiveness and student achievement is being driven by the overall determinativeness of
contract language and not by the transfer and vacancy provisions alone. I interpret these results
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 69

as evidence that, while a gap in teacher experience exists between advantaged and disadvantaged
schools, transfer policies do not make the situation worse. Nevertheless, increases in the overall
restrictiveness of teacher CBAs, not just over teacher hiring and placements, may magnify
within-district achievement gaps. These effects are small but persistent across all specifications.
In what follows, I overview the debate regarding seniority-based transfer provisions in
teacher CBAs and discuss prior research on teacher transfer policies and the distribution of
teachers to schools. I then describe the data and present empirical models that estimate within-
district gaps in experience and student achievement. I conclude with a discussion of the
implications of the results for policy decisions regarding the equity concerns of transfer
provisions and CBA restrictiveness, more broadly.
The Debate Regarding Seniority-Based Teacher Transfer Provisions in CBAs

Teacher transfers are generally grouped into two categories: involuntary and voluntary.
Involuntary transfers are district-initiated transfers and are generally utilized in times of
declining school enrollment, school closures, or program changes (or for other reasons specified
in the CBA), when teaching positions are altered or eliminated. In CBAs where seniority is the
determining factor in involuntary transfers, the least senior teacher is transferred if a more senior
teacher does not volunteer. In contrast, voluntary transfers are teacher-initiated. CBAs generally
outline procedures by which teachers can voluntarily select open vacancies before the beginning
of the next school year (and generally before outside hiring). When seniority is the determining
factor in such transfers, the most senior teacher is given first selection regardless of the
preferences of the principal at the new site.
Teachers’ unions (and presumably teachers) favor seniority provisions under the belief
that they limit or eliminate the arbitrary transfer of teachers to new positions based on the
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 70

capricious decisions of heavy-handed administrators. Seniority-based transfer provisions remove
the decision-making power from administrators and instead provide it to senior teachers, who,
whether involuntary or voluntarily transferred, have their choice of available vacancies within
the district above other, less senior, transferring teachers notwithstanding administrative
preferences.
However, some evidence suggests that utilizing seniority as a criterion for personnel
decisions is inefficient. For example, teacher layoffs are frequently made in reverse-order of
seniority, with the least senior teacher, selected as the first to lose their position. Evidence from
Los Angeles Unified School District and Washington State suggests that laying off teachers
based on seniority can intensify the negative effect of layoffs for students by increasing the level
of teacher churn, especially in schools with traditionally disadvantaged populations (Goldhaber,
Strunk, Brown & Knight, 2016; Knight & Strunk, 2016). In addition, studies from LAUSD, New
York state, Washington, and North Carolina find that seniority-based layoffs lead to the removal
of more effective teachers when compared to layoffs implemented based on measures of teacher
effectiveness (e.g., value-added measures of teachers’ contributions to student achievement or
observation-based teacher evaluation scores) and actually harm the effectiveness of teachers who
are impacted by seniority-based layoff processes yet remain in teaching (Boyd, Loeb, Lankford,
& Wyckoff, 2011; Goldhaber & Theobald, 2013; Kraft, 2015; Strunk, Goldhaber, Knight &
Brown, 2018). Critics of seniority-based transfer provisions use a similar logic; they argue that
such provisions are an inefficient and ineffective way to determine teacher placements because
they assign teachers to schools with little regard to the effectiveness of the teacher or their fit
with the new position (Ballou, 2000a; Ballou, 2000b; Hill, 2006; Levin, Mulhern, & Schunck,
2005; Moe, 2006a). Critics of these provisions further argue that they unnecessarily cede
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 71

administrative control of teacher placements over to senior teachers, who, if given the option,
will disproportionately choose to work in higher performing schools where working conditions
are better (Hess & West, 2006). Consequently, seniority-based transfer provisions could both
worsen gaps in teacher experience and student performance between the most advantaged and
disadvantaged schools.
Nonetheless, even as lawmakers strive to limit the rule of seniority in determining teacher
transfers, there is little evidence to support the contention that seniority-based provisions have an
adverse effect on students or exacerbate gaps in teacher quality across student subgroups.
Reforms to seniority-based transfer policies may not have the intended effects on student
achievement or on the distribution of teachers to schools. In particular, if teachers view these
protections as a form of compensation (see Strunk, Barrett and Lincove (2017) and Rothstein
(2015) for a discussion of how this may be the case), then the removal of critical teacher
protections could indeed adversely impact the overall supply of teachers and particularly of
experienced teachers in already hard-to-staff districts.
15

Prior Studies
Prior studies on teacher transfer provisions and the distribution of teachers are as varied
in findings as they are in methods, sample, coding scheme, and identification strategy (see Koski
& Horng (2014) and Anzia & Moe’s (2014b) reply for some of this discussion). Table 14
documents the similarities and differences between existing studies and compares them with the
present study on the topic. Although the earliest studies on CBA teacher transfer provisions

15
One way to view job protections is as a type of non-monetary form of compensation. For example, if teachers
value job security, then tenure and due process protections lower the opportunity of costs of teaching thereby
making the profession more attractive. Changes to these job protections, like seniority-based transfer provisions,
may then have important and perhaps unfortunate consequences for the distribution of teachers (Strunk, Barret, &
Lincove, 2017).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 72

(Koski and Horng, 2007; Moe, 2006c) examine CBAs from the same state (California), they
employ different data sets of CBAs from different years, different coding schemes for the CBA
transfer language, and different identification strategies, and come to different conclusions. As
shown in Column A of Table 14, the study by Terry Moe (2006c) uses a sample of 1,588
California schools nested within 115 school districts, to estimate a series of district fixed effects
models that predict the proportion of experienced teachers and credentialed teachers in a school
as a function of district transfer and vacancy policies found in those districts CBAs in the 1998-9
school year. He finds that strong transfer provisions are indeed associated with a higher
proportion of inexperienced and uncredentialled teachers in disadvantaged schools, though the
generalizability of his conclusions are limited by his restricted sample of school districts.
16

Koski and Horng (2007) (shown in Table 14, Column B) follow up on Moe’s study with
their own in California that employs a larger sample of districts (488 with four or more schools)
nearly a decade later (2005-6) and a more detailed transfer and vacancy scale measure that
captures the “restrictiveness” of a district’s transfer policy. The authors’ sum together the
collective strength of six different involuntary and voluntary transfer provisions garnered from
their sample of teacher CBAs. The authors also diverge from Moe’s fixed effect approach,
arguing that an HLM strategy that considers the nested structure of the data (schools within
districts) is more appropriate. Nonetheless, this approach ignores concerns about endogeneity in
teacher transfers that might be at least partially addressed using a district fixed effect. And while
the authors employ nearly identical outcomes and control measures to those used in Moe

16
He limits his sample to districts with four or more elementary schools and to districts where the median school has
between 15 to 85 percent minority students and at least 5 percent inexperienced teachers. This is based on the
concern that teacher transfers will matter less or not at all in areas with fewer schools and with more uniformity in
the demographic make-up of the student body. Unfortunately, this effectively excludes a large share of school
districts in California, including some of the most disadvantaged in the state.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 73

(2006b),
17
they do not limit their sample to only elementary schools or to schools in the middle
of the distribution of school disadvantage.
18
Koski and Horng’s (2007) findings diverge from
Moe’s (2006b). They conclude that “there is a teacher quality gap between schools and between
districts…but CBAs do little to make the situation worse (p. 298).”
Two studies later replicate both the original Koski & Horng and Moe studies, one in
Florida (Cohen-Vogel, Feng, Osborne- Lampkin, 2013) (shown in Table 14, column C) and one
in California (Anzia & Moe, 2014a) (shown in Table 14, column D). Cohen-Vogel, Feng, &
Osborne-Lampkin (2013) take up the issue of teacher transfer policies in Florida using a sample
of 66 out of the 67 unionized districts in Florida during the 2002-2003 school year. One notable
strength of this study is that they expand their measures of “teacher quality” to include not only
credentials and experience but also the percentage of teachers with National Board Certification,
and the average teacher SAT scores in a school. Replicating the same transfer and vacancy
measures and identification strategies of the two prior studies, they find little to suggest that
transfer and seniority provisions in CBAs worsen the teacher experience and quality gap between
more and less disadvantaged schools in Florida. Anzia and Moe (2014a) employ data that allows
them to perform an exact replication of Koski & Horng (2007) and of Moe (2006c), and find that
the insignificant finding on the interaction between teacher transfer policy and the percentage of
disadvantaged students is due to Koski and Horng using a two-tailed t-test rather than a one
tailed t-test, which Anzia and Moe argue is more appropriate given the negative hypothesized

17
Moe (2006c) operationalizes his experience measure as the proportion of teachers with less than three years of
experience whereas Koski and Horng (2007) employ a measure that is the proportion of teachers with more than two
years of experience. They do this because tenure is granted at two years in California. Koski and Horng add a few
additional controls including the number of schools in the district, school location (urban, rural, suburban), per pupil
expenditures, the number of school service days for teachers, and the percentage of college-educated adults in the
district boundaries.
18
Koski & Horng also include LAUSD in their sample whereas Moe (2006c) did not.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 74

direction of the transfer policy effect.
19
Nevertheless, given the limited evidence on the direction
of the effect of teacher transfer policies on the distribution of teachers, a two-tailed test is more
conservative. Using their one-tailed test, Anzia and Moe conclude from their analysis that
seniority-based transfer provisions do indeed burden disadvantaged schools with higher
proportions of inexperienced teachers.
The most recent study (shown in Table 14, Column E) and the only to employ teacher-
level data, draws upon longitudinal data on teacher assignments from Washington state with
nearly 200,000 teacher/year observations (Goldhaber, Lavery, & Theobald, 2016). The
advantage of teacher-level data is that it allows the authors to more fully explore and account for
other mechanisms that might cause an inequitable distribution of teachers across schools (e.g.
layoffs, attrition, hiring) above and beyond CBA transfer provisions. The availability of both
teacher-level and student-level data also provides the opportunity to create a more direct measure
of teacher effectiveness using value-added methods. Overall, Goldhaber et al., find that
involuntary transfer procedures (but not voluntary transfer procedures) do, indeed, intensify gaps
in teacher experience, as the likelihood of veteran teachers’ leaving disadvantaged schools and
novice teachers’ staying in disadvantaged schools is higher in districts that utilize seniority in
involuntary transfer decisions. However, they also find that more effective teachers are less
likely to exit disadvantaged schools when seniority is utilized in involuntary transfers, which
suggests that these provisions might improve the equitable distribution of quality teachers.
Given the differences in the studies (e.g. dataset, sample, coding scheme, and modeling
strategy), it is difficult to draw any definitive conclusions as to the impact of seniority-based
transfer provisions on the distribution of teacher experience. Not only is more research needed,

19
They also successfully replicate the Moe (2006c) results by excluding LAUSD from Koski and Horng’s data.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 75

but more research of a certain kind. The current study (shown in Table 14, Column F) builds on
existing research in at least three ways. First, to date, existing studies rely on CBA data from a
single period and thus are unable to assess the implications of shifts in transfer policies on the
distribution of teachers. If restrictive CBA transfer policies are driving the inequitable
distribution of teachers to schools, then we should expect to see the distribution of experienced
teachers grow more inequitable in districts where transfer policies grow more restrictive. This
study is the first to employ data from multiple bargaining cycles to explore the implication of
CBA transfer policy changes on the distribution of teacher experience. Second, this study
employs a more comprehensive measure of transfer and vacancy provisions than is used in prior
work. Transfer and vacancy sections of teacher CBAs contain many more provisions than those
simply governing the use of seniority in involuntary and voluntary transfer decisions that may
effectively restrict administrative control over teacher placements and thereby influence the
distribution of teachers. Generated using the Partial Independence Item Response (PIIR)
approach (Strunk & Reardon, 2010), the measure employed in this study relies on information on
the restrictiveness of 34 different transfer and vacancy policies in California CBAs. Finally, if
disadvantaged schools are bearing the brunt of restrictive transfer policies, then we might also
expect teacher transfer policies to explain school achievement gaps. This is the first study to
explore the implications of teacher transfer policies on within-district gaps in student
achievement.
Data
To assess the relationship between seniority-based transfer provisions and within district
gaps in teacher experience and student achievement, I draw upon longitudinal school- and
district-level data on teachers, students, and CBAs in California. Data on district CBA transfer
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 76

policies are drawn from a self-collected database of contracts from California school districts
with four or more schools. School-level data on teachers and students are drawn from the
California Department of Education’s (CDE) and National Center for Education Statistics’
(NCES) publicly-available data archives. These data sources and measures are described in more
detail below.
Transfer and Vacancy Measures
Data on transfer policies are generated from a dataset of CBAs that govern school district
operations over four successive bargaining cycles (in place in the 2005-6, 2008-9, 2011-12, and
2014-15 school years) in California school districts with at least four or more schools. The
sample was limited to districts with four or more schools because teacher transfer policies
operate to a much lesser extent in smaller districts where the within-district job choices of
teachers are more limited (Anzia & Moe, 2014a; Koski & Horng, 2007; Strunk, 2011, 2012;
Strunk & McEachin, 2011; Strunk & Reardon, 2010).
20
CBAs were collected every three years
because California law requires that CBAs are negotiated at least every three years, although
districts and teachers’ unions may negotiate more frequently if desired (California Teachers
Association, 2017). In total, the dataset of contracts contains 466 (80 percent) of California
school district contracts from districts with four or more schools in the 2005-2006 school year,
501 (86 percent) from the 2008-2009 school year, 490 (84 percent) from the 2011-2012 school
year, and 495 (86 percent) from the 2014-2015 school year. Contracts, including all
supplementary material (e.g. salary schedules, memorandums, tentative agreements), were
collected by first searching district websites. In the few cases where the documents were not

20
I further explored limiting the sample like Moe (2006c) to districts with four or more elementary schools, to
schools with more than 15 percent and less than 85 percent minority students (and also excluding LAUSD) and the
results were substantively similar (see Appendix Tables 6 and 7).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 77

readily available on the websites, official public records request were first emailed and then
mailed to school district human resources personnel and the superintendent.
I generate four separate measures of transfers and vacancies, following the previous
literature and then expanding to my own, I argue, more complete measure of the overall
restrictiveness of transfer and vacancy regulations in CBAs.
Anzia & Moe Measure. I first replicate the seniority-based transfer measure created by
Anzia and Moe (2014a) and Moe (2006c), who demarcate contracts with a 1 if the CBA requires
that seniority is the determining factor in deciding which teacher receives a transfer and 0
otherwise. Scored separately for involuntary versus voluntary transfers, the authors then sum the
two measures to create a final transfer score that ranges from 0 (seniority is not the determining
factor in voluntary or involuntary transfer decisions) to 2 (seniority is the determining factor in
both involuntary and voluntary transfer decisions). One weakness of this measure is that it lumps
voluntary and involuntary transfer provisions together, when in fact, these types of transfers
serve different purposes, are initiated for different reasons, and may interact with school
characteristics in different ways to influence the distribution of teachers (Goldhaver, Lavery, &
Theobald, 2016). Another weakness of this measure is that it only captures the degree to which
seniority determines who receives a transfer or not. Many transfer and vacancy provisions in
CBAs are, in fact, more nuanced in their treatment of seniority than the Anzia and Moe measure
captures.
Goldhaber et al. (2016) measures. To address the shortcomings in the Anzia and Moe
measure, Goldhaber, Lavery, & Theobald, (2016) utilize two categorical measures in
Washington state that provide more detail on the rules governing involuntary and voluntary
transfer processes. The authors determine whether seniority for a given district’s transfer policies
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 78

is prohibited from use, not addressed, addressed as one of several factors, is the tiebreaker when
two or more equally qualified teachers apply, or is the only factor used. An identical measure is
created for involuntary transfer procedures. I recreate these measures in this study, however, I
drop the “seniority is prohibited” category from my measures because California CBAs do not
expressly prohibit the use of seniority in transfer decisions. These measures clearly distinguish
between voluntary and involuntary transfers and allow us to distinguish to what degree seniority
factors into transfer decisions.
Koski & Horng (2007) Measure. The Anzia and Moe and Goldhaber et al. measures
focus explicitly on the role of seniority in voluntary and involuntary transfer decisions and there
is good reason for this. Seniority-based policies are the most controversial because they cede
administrative control over teacher assignments to the most senior teachers, who can choose their
teacher positions, generally free of other requirements on their qualifications or skills. This is
particularly important when thinking about the distribution of experienced teachers to low
performing and high-minority schools.
Nevertheless, the transfer and vacancy sections of teacher CBAs in California are lengthy
and contain many more provisions than those simply dictating the use of seniority in involuntary
and voluntary transfer decisions. These provisions also determine administrative control over
other aspects of transfers in ways that may shape the characteristics of teachers at their schools.
For example, some CBAs place restrictions on the reasons for which teachers may be
involuntarily transferred (e.g. for enrollment changes, professional growth, or because of the
elimination or change of a program), limit the number of times a given teacher can be
transferred, or determine teachers’ rights to return to former positions if transferred. While these
provisions are not explicitly focused on seniority, they, like the seniority-based provisions, limit
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 79

managerial discretion over the movement of personnel and consequently give teachers more
control over their teaching placements. Thus, not only should we be concerned about the role of
seniority-based transfer policies in the strategic sorting of teachers to more advantaged schools
but also the overall restrictiveness of the transfer and vacancy provisions in the CBA, which limit
the overall administrative options for addressing within-district inequities in teacher experience
and quality (Grissom, Loeb, & Nakashima, 2014).
Koski and Horng (2007) generate an initial measure of the “restrictiveness” or “strength”
of a given CBA’s transfer provisions by summing together their coding of six items (See
Appendix 4 for the items), which results in a final transfer and vacancy scale score that ranges
from 1 to 10. I replicate this measure as closely as possible, although I am missing two of the six
items from their original measure in the database. This results in a transfer and vacancy scale
score that ranges from 1 to 7.
Transfer and Vacancy Restrictiveness Measure. While the Koski and Horng measure
considers more items than the previous two measures, absent from the measure are several
additional items that further restrict managerial control over teacher transfers and vacancy
assignments. Furthermore, the items in their measure are given equal weight in terms of their
contribution to latent transfer and vacancy restrictiveness. One could imagine that some items
hold more significance in practically restricting the latitude of management versus others.
In lieu of these shortcomings, I create an additional measure of transfer and vacancy
restrictiveness that considers a broader set of items governing transfer and vacancies (34 items,
see Appendix 4 lists the provisions included in the measure). This measure includes the
traditional items governing the role of seniority in involuntary and voluntary transfers but also
considers other provisions, including restrictions on the number of times a teacher can be
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 80

involuntarily transferred, the order in which new employees are considered for vacancies, and
the procedures by which a teacher may return to a school in which they were involuntarily
transferred from in subsequent school years.
I generate the transfer and vacancy restrictiveness measure using a process pioneered by
Strunk and Reardon (2010) and validated in other research on overall CBA restrictiveness (e.g.
Goldhaber, Lavery, Theobald, 2014; Goldhaber, Lavery, Theobald, D'Entremont, & Fang, 2013;
Strunk, 2011; 2012; Strunk & Grissom, 2010; Strunk & McEachin, 2011). The benefit of this
approach versus others is that it considers the conditional structure of CBAs and allows for each
CBA items to be differentially correlated with the restrictiveness trait. In this process, every
CBA is read and coded for the presence of 34 contract provisions governing teacher transfers.
Contractual provisions that impose a restriction are demarcated with a 1 in the formal coding
sheet, and if a subsequent, more restrictive, but related provision is also present in the CBA, that
item is also given a 1, such that each contract item is assessed along a continuum of
restrictiveness based on the degree to which higher-order, more restrictive provisions are found
in the CBA.
Following the extant literature (e.g., Goldhaber, Lavery, Theobald, 2014; Goldhaber,
Lavery, Theobald, D'Entremont, & Fang, 2013; Marianno, Killbride, Theobald, Stunk, Cowen,
& Goldhaber, 2018; Strunk, 2011; 2012; Strunk & Grissom, 2010; Strunk & McEachin, 2011;
Strunk & Reardon, 2010), the coded information is then used in an adaptation of the Partial
Independence Item Response (PIIR) model developed by Reardon and Raudenbush (2006). This
model, first applied to measuring CBA strength by Strunk and Reardon (2010), is a generalized
hybrid of a discrete time hazard model and a Rasch model that adjusts for the conditional
structure of “response” patterns in a CBA. The PIIR model measures the underlying latent
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 81

restrictiveness of the transfer process, using information from the individual regulations garnered
from the contract, where each item is weighted based on the conditional probability that the item
appears in CA CBAs and its unique contribution to the latent restrictiveness trait. The PIIR
model is formally estimated as a multilevel random effects logistic regression, where contract
items are nested within contracts, within years. The model predicts the likelihood that a given
transfer provision is found within the CBA conditional on the inclusion of earlier, less restrictive
transfer item and as a function of some latent level of transfer item restrictiveness over time.
As in earlier work, the intercept of the multilevel random effects logistic regression is the
parameter of interest, interpreted as the contract-specific latent transfer restrictiveness which is
allowed to vary across contracts and districts. Let Ykig equal the outcome (0,1) of each item k in
contract-year i, in district g, and hkig represent the presence of the gate item for provision k in
contract i in district g so that 𝜑 𝑘𝑖𝑔
= 𝑃𝑟 (𝑌 𝑘𝑖𝑔
= 1|ℎ
𝑘𝑖𝑔
= 1). The gate item represents the
conditional structure of CBAs, where the presence of a given item in a CBA (in each year) is
dependent on a higher order item being represented. For example, a contract can only specify
that seniority is the determining factor in the transfer process when it first stipulates that seniority
is considered in transfer decisions. Thus, 𝜑 𝑘𝑖𝑔
is the conditional probability of a positive
response to item k for contract-year i in district g, conditional on passing through the gate item
hkig, where hkig is equal to 1 when the gate item is represented in the CBA in a given year (Strunk
& Reardon, 2010, p. 645).
The structural model then takes the following form:
𝑙𝑜𝑔 [
𝜑 𝑘𝑖𝑔 1−𝜑 𝑘𝑖𝑔 ] = 𝜃 𝑖𝑔
+ ∑ 𝛾 𝑗 𝐷 𝑗𝑖𝑔 + 𝝉 𝒊
𝐾 𝑗 =1
(1)
Where the conditional probability of provision k appearing in contract-year i in district g is a
function of 𝜃 𝑖𝑔
, or the latent restrictiveness of the transfer process in CBA-year i in district g, and
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 82

ϒj which is the coefficient on a vector of dummy variables for each contract item (Djig) and
represents the conditional restrictiveness of each item. In sum, the model is estimating the log
likelihood that a given transfer provision is included in the CBA in each year, conditional on the
gate contract provision being included and as a function of latent transfer restrictiveness.
In previous cross-sectional work, 𝜽 𝒊 represented the latent level of CBA transfer and
vacancy subarea restrictiveness in a single year. Here 𝜽 𝒊𝒈
is the latent contract restrictiveness in
each contract–year within a given district, g, and the year-specific level of CBA
restrictiveness, 𝝉 𝒊 . To accurately capture the restrictiveness of a contract’s transfer and vacancy
provisions in each year, then, I now capture and add the contract/negotiation year random effect
back to my estimated latent restrictiveness (𝜽 𝒊𝒈
+ 𝝉 𝒊 ) to obtain the total transfer and vacancy
restrictiveness of a contract that district administrators experience in each year of data.
Outcome Measures of Teacher Experience
Following prior research (Anzia & Moe, 2014a; Cohen-Vogel, Feng, Osborne-Lampkin,
2013; Koski & Horng, 2007; Moe, 2006c), I use two measures of school-aggregated teacher
experience as outcome variables: (1) the percentage of teachers in a school with more than two
years of teaching experience (referred to as teacher experience) and (2) the percentage of
teachers with at least a Master’s degree (referred to as teacher education). These data are derived
from the CDE’s school-level Professional Assignment Information Form (PAIF) staff profile
data for the 2005-6 through the 2014-15 school years. In California, it makes sense to use the
two-year demarcation point because teachers move from probationary to permanent status after
two years of in-district experience.
Although these measures are often referred to in prior research as measures of teacher
quality (e.g. Cohen-Vogel, Feng, Osborne-Lampkin, 2013; Koski & Horng, 2007) this paper
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 83

makes a careful distinction between measures of teacher experience/education and measures of
teacher quality. Prior research suggests that individual teacher attributes are only weakly related
to student achievement (Aaronson, Barrow, & Sander, 2007; Clotfelter, Ladd, & Vigdor, 2007;
Goldhaber, 2007; Goldhaber & Brewer, 2000; Hanushek, 1986; Harris & Sass, 2011; Rivkin et
al., 2005) and thus many not be adequate proxies for teacher quality.
However, there are still good reasons to care about the inequitable distribution of teacher
experience and education within school districts. First, if teacher performance continues to
improve over the course of the teaching career (e.g. Harris & Sass, 2011; Papay & Kraft, 2015;
Wiswall, 2013), then there may be substantial differences in the quality of teaching in schools
with a significant number of novice teachers compared to those with teachers who have more
years of experience. Second, there are non-test student outcomes associated with greater teacher
experience or education that are indirectly related to student achievement, including absences,
reported disruptive classroom offenses, time spent completing homework, and time spent reading
for pleasure (Ladd & Sorensen, 2016). Finally, some research suggests that there are significant
peer effects associated with being in a school with a high proportion of experienced teachers,
which may lead to better student outcomes (Jackson & Bruegmann, 2009). Thus, even though
experience or education is not an adequate proxy for teaching quality, there is still strong
justification to pay attention to how more experienced/educated teachers are distributed to
schools.
Outcome Measures of Student Achievement
School-level aggregated raw test scores by grade are publicly available from the CDE’s
Standardized Testing and Reporting (STAR) and California Assessment of Student Performance
and Progress (CAASPP) systems. STAR testing data are available from 1997-1998 to the 2012-
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 84

2013 school year in English-language arts (ELA) and math. Due to the switch to Common Core
aligned exams, no test scores are available in California for the 2013-2014 year. The CAASPP
system now archives school-aggregated results in ELA and math on the new Smarter Balanced
exams beginning with the 2014-2015 school year through 2016-2017. For this analysis, I rely on
student achievement data from 2004-2005 to 2015-2016.
There are two challenges for this analysis in estimating achievement over time given the
change in exams with Common Core. The first is that the Smarter Balanced exams are vertically
aligned such that they are scored on a continuum from grade-to-grade for both math and ELA.
When using school-aggregated raw test-score averages, some schools may appear to have higher
mean test scores simply because they enroll more students in higher grades. Second, STAR and
Smarter Balanced exams are different tests that are aligned to different standards and are
therefore testing different content. They are also scored on different scales. On the STAR exam,
a student could score from 150 to the maximum score of 600 in each grade and for both ELA and
math, whereas the Smarter Balanced exam a student can score from 2114 to 2862 depending on
their grade and subject (California Department of Education, 2016).
To make test scores comparable across grades and over time, I follow prior research and
standardized the SBAC tests within grade, subject, and year for all students such that the within
grade, subject, year average test score has a mean of zero and a standard deviation of 1 (e.g.,
Backes, et al. 2016). To aggregate to the school-level, these grade-level scores are then weighted
by the number of test takers in each grade in each school, so that each grade receives its proper
weight in the final school-level average. A given school’s test score in, for example, ELA, can
then be interpreted as that school’s deviation from the school population mean in ELA. Because
the STAR exams are similarly scaled across grades, these exam scores are not standardized
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 85

across grades but are standardized within year, such that a given school’s STAR score can also
be interpreted as their deviation from the school population mean.
Other School and District Control Variables
The primary measure of school disadvantaged used in prior research is school minority
composition (the percentage of students that are African American, Hispanic, or Native
American), which I also employ in this study (Anzia & Moe, 2014a; Cohen-Vogel, Feng,
Osborne- Lampkin, 2013; Koski & Horng, 2007; Moe, 2006c).
21
I also control for other
important school-level factors that may influence the transfer rates out of schools or student
achievement, including the natural log of school enrollment, the school-wide average student-to-
teacher ratio, school location (urban, rural, suburban), the natural log of district enrollment, and
the percentage of free and reduced price lunch students in the district. These variables are drawn
from the National Center for Education’s (NCES) Common Core Data Files and the CDE’s
Annual Performance Index (API) data files.
Analytic Strategy
I first replicate prior research (Anzia & Moe, 2014a; Cohen-Vogel, Feng, & Osborne-
Lampkin, 2013; Moe, 2006c) by estimating a two-way interaction model where the relationship
between school-level teacher experience or school-level teacher education and the percentage of

21
I also estimated equations 2, 3, and 4 using an alternative measure of school disadvantage (the percentage of free-
and reduced-price lunch students in the school). Results are shown in Appendix Tables 8 and 9. The results from the
models predicting teacher experience and teacher education are substantively similar, however, I do find some
differences in the models predicting student achievement. While the coefficients on the key interactions between
district transfer and vacancy policy and school disadvantaged are still negative, as in the main specifications (shown
in Table 15), some coefficients change slightly in magnitude (and change in statistical significance). While I find
that the relationship between the percentage of minority students and math achievement is more negative for schools
in districts where transfer and vacancy provisions are growing more restrictive over time (Table 15, panel C,
columns 5 and 6, rows 6, and 9) these effects are not statistically significant when using school percent FRL as my
measure of disadvantage (Appendix Table 9, panel C, columns 5 and 6, rows 6, and 9). Furthermore, I find that
switching from not using seniority in involuntary transfer decisions to using seniority as the determining factor is
significantly associated with a decrease in ELA and math scores for every 1 percent increase in FRL (Appendix
Table 9, panel B). In my main results, I only find statistically significant changes in math.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 86

minority students is allowed to vary across districts with different CBA transfer provisions on a
cross-section of data from the 2005-2006 school year. The model is formally specified as
follows:
Y
ij
= β
0
+β
1
minority
ij
+β
2
minority
ij
x transfer
j
+ S
ij
β
3
+δ
j
+ε
ij

Where Y
ij
is either the percentage of experienced teachers or the percentage of teachers with
a master’s degree at school i in district j. Y
ij
is a function of the percentage of minority students
in school i in district j, and the main effect of interest—the interaction between the percentage of
minority students in a school and a district’s transfer policy. If seniority-based transfer and
vacancy provisions exacerbate the gap in teacher experience/education between schools of
differing levels of disadvantage, then β
2
should be negative. S
ij
is a vector of school
characteristics, including student enrollment and the student-to-teacher ratio. δ
j
is a district fixed
effect. Because this model is identified from the variation in district teacher transfer policy in a
single year, the main effect of teacher transfer policy on the percentage of experienced teachers
is perfectly collinear with the district fixed effect and is therefore not included in the model.
While these cross-sectional results are informative, they do not account for some sources
of endogeneity that may bias the observed relationships. Specifically, because CBAs are not
randomly assigned to school districts, important unobserved differences may exist between
districts with more or less determinative transfer policies that are also related to the distribution
of experienced and graduate-educated teachers in the district. For example, experienced teachers
with strong preferences to avoid disadvantaged schools may self-select into districts with transfer
policies that allow them to move easily between schools, which would bias the effect of
seniority-based transfer policies on the distribution of teacher experience upward (Anzia & Moe,
2014a; Goldhaber, Lavery, Theobald, 2015). However, because I have bargaining data across
(2)
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 87

multiple CBA cycles, I can more closely approximate the relationship between transfer and
vacancy policies, school characteristics, and the distribution of teacher experience/education by
exploiting within-district CBA changes. In this case, the identifying variation on the main effect
of interest is derived only from districts who change their transfer and vacancy policies across
the four observed contract cycles. This model is specified by adding a time dimension, t, to
equation (1) and by maintaining the district fixed effect:
Y
ijt
= β
0
+β
1
transfer
jt
+ β
2
minority
ijt
+β
3
minority
ijt
x transfer
jt
+D
jt
β
4
+ S
ijt
β
5
+δ
j
+ τ
t
+ε
ijt

Now the outcome variable, Y
ijt
, is a function of the transfer policy in district j in year t, the
percentage of minority students in school i in district j in year t, and the main effect of interest,
the two-way interaction between the percentage of minority students and the district transfer
policy. β
3
can be interpreted as the differential impact of a district’s transfer and vacancy policy
growing more restrictive (or when using dichotomous indicators, switching from not factoring in
seniority to utilizing seniority) on the relationship between the percentage of minority students
and school-level teacher experience/teacher education and the percentage of minority students,
holding all else equal. In this model, every district serves as their own control, and the effects are
interpreted as deviations from the within-district average over time. It’s important to note that β
3

represents the causal effect of teacher transfer policies only if, after conditioning on D
jt
and S
ijt
,
the vectors of time-varying district and school controls, and after including δ
j
, the district effect,
and τ
t
, the year fixed effect, ε
ijt
, is independent. However, there may be unaccounted for district-
specific trends related to both changes in teacher transfer policies and the distribution of teacher
experience/education in schools that make this assumption unrealistic. For example, changes in
teacher transfer policy and the distribution of teacher experience or teacher education in certain
(3)
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 88

districts may be strongly influenced by reform-oriented superintendents who implement new
reforms geared towards improving teacher quality.
In equation (3), I control for district-specific time trends by substituting the district and
year fixed effect in equation (2) for a district-by-year fixed effect, δ
j
*τ
t
:
Y
ijt
= β
0
+β
1
transfer
jt
+ β
2
minority
ijt
+β
3
minority
ijt
x transfer
jt
+D
jt
β
4
+ S
ijt
β
5
+δ
j
*τ
t
+ε
ijt

β
3
represents the causal effect of teacher transfer policies only if, after conditioning on D
jt
,
S
ijt
, and δ
j
*τ
t
(the district-specific time trends), ε
ijt
, is independent. This is unlikely given that
equation (4) no longer accounts for fixed district characteristics that might shape transfer policies
and the distribution of teacher experience/education. Furthermore, it is highly unlikely that
equation (3) or (4) adequately controls for all the school characteristics that might be related to
school-level teacher experience and the adoption of specific transfer policies. Comparing the
estimates from equations (3) and (4) likely yield more unbiased findings than presented in prior
work, but these results should not be read as implying causal relationships.
As mentioned above, seniority-based and restrictive teacher transfer policies may also
shape school achievement gaps. If seniority-based transfer provisions afford more
experienced/educated teachers greater opportunities to transfer to more advantaged locations
(thus, leaving disadvantaged schools with less experienced and effective teachers), then I expect
larger school achievement gaps in school districts where seniority is the determining factor in
transfer decisions. Furthermore, increased administrative restrictions on teacher hiring and
placement may affect student achievement independent of the influence of seniority provisions
on the distribution of teacher experience. Increased hiring and assignment restrictions may make
it difficult for administrators to match classroom teachers with assignments that best fit their
qualifications, particularly in disadvantaged locations where there is more teacher turnover.
(4)
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 89

To explore the relationship between seniority-based transfer policies, transfer and vacancy
restrictiveness, and achievement, I substitute the outcome variables in equations (2) through (4)
for school-level achievement measures (Math and ELA scores). The model specifications remain
otherwise unchanged.
Results
How Much Do Teacher Transfer Provisions Change Over Time?
Table 15, Panel A provides summary statistics for the different contract restrictiveness
measures in each year (columns as well as the average change in transfer provisions across
negotiation cycles). Overall, only 3 percent of school districts in the sample use seniority as a
determining factor in voluntary transfers in 2014-2015 (compared to 5.8 percent in 2005-2006). I
also observe a decrease in the percentage of school districts that do not consider seniority at all in
voluntary transfer decisions, down from 29.3 percent in 2005-2006 to 21.5 percent in 2014-2015.
This is met by a corresponding increase in the percentage of districts that use seniority as one of
many factors in making voluntary transfer decisions (from 23.9 percent in 2005-2006 to 30.5
percent in 2014-2015). There is also a similar decrease in the percentage of districts that do not
consider seniority when making involuntary transfers (from 26.5 percent in 2005-2006 to 17.1
percent in 2014-2015). This corresponds with an increase in the percentage of districts using
seniority as a determining factor in involuntary transfers (from 14.7 percent to 22 percent from
2005-2006 to 2014-2015).
Of course, these sample averages mask some of the individual changes districts have
made across contract years. To explore these in more depth, I reduce the sample to a balanced
panel (districts for which I observe a CBA in all contract-years, n=363) and explore the types of
within-district provision changes over time. Thirty-six percent of school districts made at least
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 90

one change to their voluntary transfer provisions and 30 percent made at least one change to their
involuntary transfer provisions across the four bargaining cycles. Most of these changes are
moderate. For example, between the first two bargaining cycles in the panel (2005-06 and 2008-
09), of the 103 school districts that made changes to their voluntary transfer provisions, 60
percent were made between adjacent levels (e.g. from seniority as the determining factor to all
else equal, or from no seniority to seniority is considered, etc.). Only two districts switched from
not considering seniority to using seniority as the determining factor and no districts made
changes in the other direction (i.e. from using seniority as the determining factor to not
considering seniority). The percentages are similar for involuntary transfers.
Table 15 also shows the average change in the overall restrictiveness of district transfer
and vacancy provisions. The average year-to-year change on the Koski and Horng measure is
0.147 or 18 percent of a standard deviation. On the transfer and vacancy restrictiveness measure,
I find a year-to-year change of essentially zero, though some districts change quite considerably
around that mean, as indicated by the standard deviation of 0.284. To explore how much
variation is between and within districts in these measures, I partition the variation between-
districts and within-districts. Approximately, 68 percent of the variation in the Koski and Horng
measure and 56 percent of the variation on the transfer and vacancy restrictiveness measure is
between districts. Thus, between approximately one-third to one-half of the variation on these
continuous measures can be attributed to change over time.
Are Changes in Transfer Provisions Associated with the Way Experienced Teachers Are
Distributed Across Schools?

Table 16 presents the estimated coefficients from the teacher experience and teacher
education models. Columns 1 and 4 show estimates from the cross-sectional models (equation
2), columns 2 and 5 present the results derived from the district fixed effect models (equation 3),
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 91

and columns 3 and 6 show the results from the district-by-year fixed effect models (equation 4).
Each of these models is run separately for the voluntary and involuntary transfer seniority
dummies (panels A and B) and each of the combined provision measures (panel C), including the
Anzia and Moe, Koski and Horng, and new transfer and vacancy restrictiveness measure.
Table 16 shows that, all else equal, schools with higher concentrations of minority
students have fewer experienced teachers and teachers with graduate degrees. These results
support the well-documented finding that higher-minority schools have difficulty attracting and
retaining experienced and highly-qualified teachers (Clotfelter, Ladd, & Vigdor, 2005; Feng,
2010; Glazerman & Max, 2011; Goldhaber, Lavery, Theobald, 2015; Isenberg, et al., 2013;
Lankford, Loeb, & Wyckoff, 2002; Sass et al., 2012).
To explore whether the within-district gap in teacher experience and teacher education
vary significantly across districts with different seniority-based transfer provisions, I examine the
coefficients on the interactions between the percentage of minority students and the seniority
transfer variables. Unlike Anzia and Moe (2014) and to some extent, Goldhaber, Lavery, &
Theobald, (2016), but similar to Cohen-Vogel, Feng, Osborne-Lampkin, (2013) and Koski &
Horng, (2007), I do not find consistent evidence across any of these measures or any of the
model specifications that gaps in teacher experience between schools with different
concentrations of minority students are wider in school districts with seniority-based voluntary
and involuntary transfer policies or more restrictive transfer and vacancy provisions.
Are Changes in Transfer Provisions Associated with Between-School Gaps in Student
Achievement?

Table 17 presents the estimated coefficients from the student achievement models. All
regressions shown in Table 17 provide evidence that all else equal, schools with higher
concentrations of minority students have lower student achievement. Specifically, a one percent
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 92

increase in the percentage of minority students is associated with approximately a 2 percent SD
decrease in math and ELA scores.
The key question explored here is whether seniority and the overall restrictiveness of
transfer and vacancy provisions are associated with increases in the observed gap in student
achievement between advantaged and disadvantaged schools.
Table 17, Panels A and B reveal that the relationship between percentage minority and
student achievement is more negative as school districts using seniority as a determining factor
in voluntary and involuntary teacher transfer decisions versus those not using seniority in making
transfer decisions (albeit not significantly). This is true across all model specifications. Similar
negative but null results are found in Panel C for the Anzia and Moe measure. I do find, however
that the relationship between the percentage of minority students and student achievement does
become more negative for schools in districts where transfer and vacancy provisions are growing
more restrictive over time (when using the Koski & Horng measure and the transfer and vacancy
restrictiveness measure). I observe these effects more consistently in math than in ELA and in
the district and district-by-year fixed effect versus the cross-sectional specifications. Although
the effects are small (less than 1 percent of a SD), these results suggest that there could be
something, about overall restrictions on staffing that disproportionately harm schools that are
more disadvantaged (e.g. poor teacher-position fit, as mentioned earlier).
Given that prior research provides some suggestive evidence that overall contract
restrictions are associated with student achievement gaps (Strunk & McEachin, 2011), I explore
whether the transfer and vacancy restrictiveness measure is picking up something related to the
overall determinativeness of the contract that is not wholly specific to hiring and assignment
provisions. As a robustness check, I generate a measure of overall contract restrictiveness based
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 93

on 34 items culled from a set of over 600 provisions across all areas of the contract (see Strunk
& Reardon, 2010 for a review of how this measure was created). The measure was created in
similar way to the transfer and vacancy measure. Appendix 10 addresses in more detail how the
34 items were selected to represent the overall “latent restrictiveness” of California CBAs and
how the measure was constructed. Appendix Table 11 lists the 34 items contained in the measure
(which notably contains four items included in the transfer and vacancy restrictiveness measure).
Table 18 presents the results from equations 2, 3, and 4 with the measure of overall
contract restrictiveness in lieu of the transfer and vacancy restrictiveness measure. I find that the
minority ELA and math achievement gap is intensified in schools in districts where contracts are
becoming more restrictive (as evidenced by the significant negative effects on the interaction
between school minority composition and overall contract restrictiveness). This result is robust
across all specifications. To estimate the effect of overall CBA restrictiveness on achievement,
holding the restrictiveness of the transfer and vacancy section constant, I enter both measures
into the equations. Appendix Table 5 reveals that the transfer and vacancy restrictiveness
measure and the overall contract restrictiveness measure are correlated at 0.311, which I interpret
as evidence that these measures are related but are capturing the restrictiveness of different parts
of the contract Table 18, panel B, columns 1 through 6, row 5 demonstrate that this result holds
even when holding constant transfer and vacancy restrictiveness.
22
The coefficient on the
interaction between transfer and vacancy restrictiveness and school minority composition
essentially goes to zero (especially in the over time models, shown in columns 2, 3, 5, and 6),

22
There may concerns about multicollinearity when introducing both overall restrictiveness and transfer and
vacancy restrictiveness into the same model. I do not observe meaningful differences in the standard errors on the
overall contract restrictiveness estimates between the results presented in Tables 18 and 19, which suggests that
multicollinearity is not a problem. We should interpret the estimates in Table 19 as the effect of overall contract
restrictiveness when holding constant the restrictiveness of the transfer and vacancy section of CBAs.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 94

when controlling for overall contract restrictiveness. Yet, the significant negative effects on the
interaction between school minority composition and overall contract restrictiveness remain.
Consequently, the main specifications shown in Table 17, panel C may be picking up something
about the overall restrictiveness of CBAs that matter for within-district achievement gaps versus
something inherent to or solely related to determinative hiring and transfer provisions.
Conclusion
Given the well-documented gap in the experience-, education-, and effectiveness-levels of
the teaching workforce in advantaged and disadvantaged schools (Clotfelter, Ladd, & Vigdor,
2005; Feng, 2010; Glazerman & Max, 2011; Goldhaber, Lavery, Theobald, 2015; Isenberg, et al.,
2013; Lankford, Loeb, & Wyckoff, 2002; Sass et al., 2012), policymakers and education reformers
are searching for solutions to improve the distribution of quality teachers to hard-to-staff locations.
Obama-era programs (e.g. the Teacher Incentive Fund (TIF), Race to the Top (RTT), provided
grant-based support to education agencies willing to experiment with new evaluation systems and
compensation incentives to identify, recruit, and retain high quality teachers to teach in the hardest-
to-staff locations. By 2016, 44 states had passed new legislation overhauling traditional teacher
evaluation systems to include test-based measure of teacher performance, additional performance
rating categories, and/or require evaluation ratings to factor into promotion, dismissal, and
compensation decisions (NCTQ, 2016). Outside of the legislature, education advocacy groups
have challenged teacher policies in court, filings lawsuits in California (Vergara v. California),
New York (Wright v. New York), and Minnesota (Forslund v. Minnesota). Plaintiffs in these cases
argue that state layoff, tenure and dismissal protections for teachers unduly burden students in
disadvantaged schools with the least effective teachers.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 95

Despite these and other efforts in states and districts across the nation, there is little to
suggest that gaps in teacher experience and quality across disadvantaged student subgroups have
attenuated. This paper takes up the issue of seniority-based teacher transfer policies and their role
in explaining within-district gaps in teacher experience/education and student achievement and is
the first to estimate these relationships over multiple bargaining cycles.
Findings from this study support the well-documented result that higher-minority schools
have difficulty attracting and retaining experienced and highly-qualified teachers; however, I
find that seniority-based transfer provisions and the overall restrictiveness of CBA transfer and
vacancy provisions do not make the situation worse. In short, it appears that the gap in teacher
experience and education credentials between schools with higher and lower proportions of
minority students does not appear to widen in districts that switch to using seniority in transfer
decisions or where provisions governing teacher hiring and placement are becoming more
restrictive towards management. This finding, now demonstrated across multiple bargaining
cycles, is consistent with the results of earlier cross-sectional studies (Cohen-Vogel, Feng, &
Osborne-Lampkin, 2013; Koski & Horng, 2007).
This study is also the first to explore the association between seniority-based and
restrictive teacher hiring and assignment provisions and student achievement. While I fail to
detect widening achievement gaps in school districts where transfer policies give more weight to
seniority, I do find larger achievement gaps between schools where district CBAs are growing
more restrictive overall, not just in teacher hiring and assignment policies as anticipated. This is
not the first study to highlight the relationship between CBA restrictiveness and within-district
achievement gaps (Strunk & McEachin, 2011) but is the first to do so over time.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 96

This result has important implications for research on teacher collective bargaining
agreements in public education. Recent evidence suggests that increases in the restrictiveness of
teacher CBAs, on average, are not significantly related to changes in district-aggregated student
achievement, even for high-poverty, Black, Hispanic, and English-Language Learner students
(Marianno & Strunk, 2018a). Nevertheless, the present study demonstrates the need to explore
disproportionate effects of restrictive CBAs between schools, within districts. While somewhat
surprising, there are important reasons for why restrictive CBA content might be unrelated to
changes in average district achievement over time but related to within-district gaps in
achievement. Some of the earliest work on teachers’ unions and collective bargaining in
education hypothesized that collective bargaining provisions have disproportionate effects on
student learning. Eberts and Stone (1987) observed that students at the tails of the achievement
distribution performed worse in unionized districts when compared to students at the tails of the
distribution in non-unionized districts. They argued that unionization and collective bargaining
impose standardized workplace procedures that make it more difficult for administrators to
implement personalized learning practices or specialized programs. Instead, standardized work
rules allow schools and teachers to address the needs of the average student, at the expense of the
lowest performing students, who might need additional and more tailored assistance.
The findings of this study suggest that policymakers might consider the ways in which
collective bargaining provisions bump against efforts to improve the performance of the lowest
performing schools. For example, researchers have documented how specific regulations on
class size, school schedule, performance evaluation, compensation, and teacher transfers make it
incredibly difficult to implement new multiple-measure evaluation systems and professional
development supports for low performing teachers (Strunk, 2014). Others have shown how
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 97

current step-and-lane salary schedules make it difficult to retain novice teachers and attract high
quality teachers to the lowest performing schools (Hess & West, 2006; Strunk & Grissom, 2010).
These studies exemplify the ways in which restrictive CBAs might hamstring administrators in
their efforts to tailor education policies to address the needs and improve the performance of
their neediest students.
Nonetheless, readers should exercise caution in interpreting these results as evidence that
removing teacher bargaining rights will improve achievement gaps. It is difficult to say what the
general equilibrium effects of removing collective bargaining rights for teachers might be. New
evidence on the impacts of tenure reform on teacher attrition in Louisiana, for example, reveal
that there was significant reform-induced attrition for teachers in schools that serve more
disadvantaged populations (Strunk, Barrett, and Lincove, 2017). To the extent that teachers view
tenure or analogous types of protections in CBAs as forms of compensation, then their removal
could have harmful effects on teacher recruitment, retention, and ultimately student learning,
thereby negating any positive benefit derived from enhanced flexibility. In lieu of reducing
teacher collective bargaining rights altogether, a more modest approach is likely necessary—one
that allows schools districts to partner with teachers’ unions and state education agencies to
apply for selective waivers to bargaining policies in order to better serve the needs of
disadvantaged students in the lowest performing schools.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 98

CHAPTER FIVE- The Future of Teachers’ Unions in Public Education: Cross-Cutting
Themes from Three Papers on Teachers’ Union Influence
Current legal challenges to public-sector union rights and teacher policy present
formidable obstacles for teachers’ unions in public education (Marianno, 2015; Marianno &
Strunk, 2018b). As noted in chapter three, theory suggests that the recent proliferation of
education interest groups in education and the centralization of education decision-making to
state and federal venues (where these groups can have more influence) should present new
opportunities for opposing groups to weaken public policy related to teachers’ unions’ interests
(Gayley, 2015; Henig, 2013). A look at recent proposed policies in state legislatures (e.g.
Freeman & Han, 2012; Marianno, 2015), and current state and Supreme Court cases suggest that
this might be the case. Since 2011, every state in the nation has proposed, and in some cases,
enacted reforms to unions’ rights to negotiate collective bargaining agreements (CBAs) with
local school districts (e.g. Wisconsin Budget Repair Bill; Freeman & Han, 2012; Marianno,
2015). State courts have considered reforms to tenure rights, seniority-based layoff provisions,
teacher discipline procedures, and evaluation policies (e.g. Doe v. Antioch, Vergara v. the State
of California; Wright v. New York; Forslund v. Minnesota). The Supreme Court case Janus v.
American Federation of State, County, and Municipal Employees, Council 31, funded, in part, by
advocacy organizations like the National Right to Work Legal Defense Foundation and the
Liberty Justice Center, will likely end the practice of collecting fair-share fees from non-union
members covered by existing collective bargaining agreements. As demonstrated by trends in
recent Right-to-Work states, an unfavorable ruling for unions on the Janus case could drastically
decrease their membership and revenue, thereby potentially weakening teachers’ unions political
influence (Marianno & Strunk, 2018b).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 99

This dissertation yields new insights on how we might expect the dynamics of teachers’
union power and influence in education to change considering new institutional and legal
challenges. First, chapter two, in exploring descriptive trends in teachers’ union campaign
contributions, election outcomes for supported lawmakers, and success rates on federal
legislation, demonstrates that drastic shifts in the political environment may not spell immediate
doom for teachers’ unions. Chapter two shows that in a period when Republicans gained control
of both chambers of Congress, teachers’ unions lost 61 allied Democrat members in the House
and 12 allied Democrats in the Senate since the 2010 elections. Certainly, this could harm the
political influence of a union that is traditionally reliant on Democrat support. Nonetheless, I find
that 79 percent of votes in the Senate and 64 percent of votes in the House still aligned with the
NEA’s policy positions during the last four legislative cycles. Their ability to maintain relatively
high success rates in Congress likely stems from their success in garnering support from a
growing contingent of Republican lawmakers, a broader trend in union political strategy that
others have noted (Rich, 2012; Winkler, Scull, & Zeehandelaar, 2012). I find that the NEA added
4 A/B Republican senators and 16 A/B Republican representatives in the 2010, 2012, and 2014
elections. In short, in the face of changing partisan politics, teachers’ unions adapted to find new
allies in unlikely places.
Chapter three investigates the relationship between teachers’ unions’ political
involvement and state agenda- and policy-setting, and demonstrates that state teachers’ union
strength, particularly their election strength, remains the strongest predictor of both the proposal
and outcome of policies aligned with union interests. I find some evidence that relative union
strength is associated with the mix of policies under consideration in state legislatures, but I do
not find that relative strength is uniquely associated with the enactment of policies. Current
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 100

trends, then, in the growth and strength of competing interest groups have yet to drastically
overtake or weaken teacher unions’ policy-setting influence. To the extent that Janus and state-
level Right-to-Work legislation reduces teachers’ unions abilities to raise revenue (as a result of
declining membership and a loss in fee-payers), particularly for political purposes, we might
expect a “leveling of the political playing field” for other interest groups in education. With less
ability to raise funds for operational expenses, teachers’ unions might change their spending
patterns and curb their political donations to election campaigns, thereby decreasing their
political influence relative to other groups.
However, we might also envision a different response from teachers’ unions to an
unfavorable Janus decision—one of adaption, a characteristic of teachers’ unions demonstrated
by the findings of chapter two. For example, there is some evidence that state union affiliates
facing losses in membership and membership dues as a result of Right-to-Work legislation are
reallocating their revenues to focus more on state-level political advocacy, not less (Marianno &
Strunk, 2018b). These groups are freeing up political resources by cutting expenditures in other
areas (e.g. staffing and benefit levels) and by engaging in political action fundraising campaigns.
If state teachers’ union affiliates in recent Right-to-Work states are any guide, we may see unions
respond to Janus by redistributing their resources, even as they become scarcer, to increase their
focus on political advocacy. This response to institutional and legal challenges also makes sense.
Increasing political action might allow teachers’ unions to protect other policy areas, like
collective bargaining, that remain central to their organization’s survival.
Even if it is likely that teachers’ unions survive current institutional and legal challenges,
they will continue to face opposition from critics. Most of this scrutiny is in fact, built on
divisive rhetoric and beliefs about how teachers’ unions affect public education. For example,
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 101

former secretary of education Rod Paige referred to the nation’s teachers’ unions as “terrorists”
in his 2006 book The War against Hope: How Teachers’ Unions Hurt Children, Hinder
Teachers, and Endanger Public Education. Former New Jersey governor Chris Christie
compared teachers’ unions to the “mafia” (Burnette, 2016a) and on another occasion said they
deserved a “punch in the face” because they are the “single most destructive force in public
education in America (Layton, 2015).” Scott Walker, governor of Wisconsin and author of
controversial reforms to union rights, stated he saved his state from the “stranglehold” of union
rules in order to “improve education for my kids, my nieces, and all of the other children across
Wisconsin (Walker, 2013, p. 3, 152).” It is important to note that advocates on the other side of
the spectrum also hold their own opposing beliefs about how unions affect schooling. For
example, proponents of unions believe that teachers’ unions protect teachers “against the
arbitrary exercise of power by heavy-handed administrators” (Ravitch, 2006), attend to “the
actual conditions of teachers’ work” (Bascia & Rottmann, 2011, p. 792; Casey, 2006), and help
improve student learning (Bascia & Rottmann, 2011).
Determining what should be made of all the rhetoric on teachers’ unions requires
empirical investigation into the effect of unions on public school outcomes (Cowen & Strunk,
2015; Strunk, Goldhaber, & Cowen, 2018). Chapter four of this dissertation helps in this regard.
In short, I demonstrate the importance of exploring the relationship between collective
bargaining policy and within-district, between-school student outcomes. The results of this study
hearken back to some of the earliest theories on the effects of collective bargaining agreements
on public schooling. Eberts and Stone (1987) argued that collective bargaining agreements
standardize schooling in ways that make schools less responsive to the needs of students at the
bottom of the achievement distribution. Indeed, I find that achievement gaps are becoming larger
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 102

in districts where collective bargaining agreements are growing more restrictive towards
management. This may be because the rules imposed by teacher collective bargaining
agreements make it difficult for lower performing, more disadvantaged schools to implement
tailored programs and reforms to address the needs of their students.
Moving beyond the rhetoric, I provide evidence, then, that there may be good reasons to
loosen some of the contractual protections in teacher bargaining agreements. Because of the
equity concerns associated with contract provisions, local and state policymakers should consider
ways to provide selective relief to school districts from bargaining provisions that obstruct the
provision of additional supports to disadvantaged students. In loosening contractual provisions,
we must also acknowledge that some contract items afford important protections for teachers, or
are, at the very least, viewed as such by teachers and their unions. Indeed, policymakers must
walk a careful balance when implementing reforms to teacher contracts. If the recent teacher
strikes in six states are any guide (Rossman, 2018), where teachers and their unions feel under
attack by reform efforts, they will rise up and use their voices to protest (Marianno & Strunk,
2018b). Consequently, efforts to reform teacher contracts may amount to a gradual refinement of
the provisions that are most detrimental for disadvantaged students. Continued efforts by
researchers to identify these provisions will help in both maintaining important and valuable
teacher contract provisions and removing or changing those that negatively affect schools’ most
vulnerable students.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 103

References
Aaronson, D., Barrow, L., & Sander, W. (2007). Teachers and student achievement in the
Chicago public schools. Journal of Labor Economics, 25(2), 95–135.
Adamnson, F., & Darling-Hammond, L. (2012). Funding disparities and the inequitable
distribution of teachers: Evaluating sources and solutions. Education Policy Analysis
Archives, 20(37), 1–46.
American Federation of Teachers. (2016). About us. Retrieved from http://www.aft.org/about.
Ansolabehere, S., de Figueiredo, J. M., & Snyder, J. M. (2003). Why is there so little money in
U.S. politics? Journal of Economic Perspectives, 17(1), 105–130.
Antonucci, M. (2017). NEA active membership actually behind where it was 20 years ago.
Retrieved from http://www.eiaonline.com/intercepts/2017/06/26/nea-active-membership-
actually-behind-where-it-was-20-years-ago/.
Anzia, S. F., & Moe, T. M. (2014a). Collective bargaining, transfer rights, and disadvantaged
schools. Educational Evaluation and Policy Analysis, 36(1), 83–111.
Anzia, S. F., & Moe, T. M. (2014b). Focusing on fundamentals: A reply to Koski and Horng.
Educational Evaluation and Policy Analysis, 36(1), 120-123.
Argys, L. M., & Rees, D. I. (1995). Unionization and school productivity: A reexaminiation.
Research in Labor Economics, 14, 49–68.
Austen-Smith, D. (1995). Campaign contributions and access. American Political Science
Review, 89(3), 566–581.
Austen-Smith, D., & Wright, J. (1994). Counteractive lobbying. American Journal of Political
Science, 38(1), 25–44.
Austen-smith, D., & Wright, J. R. (1996). Theory and evidence for counteractive lobbying.
American Journal of Political Science, 40(2), 543–564.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 104

Backes, B., Cowan, J., Goldhaber, D., Koedel, Cory, Miller, L., & Xu. Z. (2016) The Common
Core conundrum: To what extent should we worry that changes to assessments and
standards will affect test-based measures of teacher performance. (CALDER Working
Paper No. 152).
Ballou, D. (2000a). Teacher contracts in Massachusetts. Boston, MA: Pioneer Institute.
Ballou, D. (2000b). Contractual constraints on school management: Principals’ perspectives on
teacher contracts. In D. Ravitch & J. P. Viteritti (Eds.) City schools: Lessons from New York
(pp. 89–116). Baltimore, MD: John Hopkins University Press.
Bascia, N., & Rottmann, C. (2011). What’s so important about teachers’ working conditions?
The fatal flaw in North American educational reform. Journal of Education Policy, 26(6),
787–802.
Baumgartner, F., & Leech, B. (1996). The multiple ambiguities of “Counteractive Lobbying.”
American Journal of Political Science, 40(2), 521–542.
Baumgartner, F. R., & Jones, B. D. (1991). Agenda dynamics and policy subsystems. The
Journal of Politics, 53(4), 1044–1074.
Baumgartner, F. R., & Jones, B. D. (2002). Policy dynamics. Chicago, IL: University of Chicago
Press.
Baumgartner, F. R., & Jones, B. D. (2009). Agendas and instability in American politics.
Chicago, IL: University of Chicago Press.
Baumgartner, F. R., Jones, B. D., & Mortensen, P. B. (2014). Punctuated equilibrium theory:
explaining stability and change in public policymaking. In P.A. Sabatier & C.M. Weible
(Eds.), Theories of the policy process (59-103). Boulder, CO: Westview Press.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 105

Berry, F.S., & Berry, W. D. (1990). State lottery adoptions as policy innovations: An event
history analysis. The American Political Science Review, 84(2), 395–415.
Björk, L., & Lindle, J. C. (2001). Superintendents and interest groups. Educational Policy, 15(1),
76-91.
Boyd, D., Lankford, H., Loeb, S., & Wyckoff, J. (2005). The draw of home: How teachers’
preferences for proximity disadvantage urban schools. Journal of Policy Analysis and
Management, 24(1), 113–132.
Boyd, D., Lankford, H., Loeb, S., Rockoff, J., & Wyckoff, J. (2008). The narrowing gap in New
York City teacher qualifications and its implications for student achievement in high-
poverty schools. Journal of Policy Analysis and Management, 27(4), 793–818.
Boyd, D., Lankford, H., Loeb, S., Wyckoff, J., Finance, S. E., & Summer, N. (2011). Teacher
Layoffs: An empirical illustration of seniority versus measures of effectiveness.
Education Finance and Policy, 6(3), 439–454.
Burnette, D. (2016a, September, 6). N.J. Gov. Chris Christie calls teachers union 'mafia,' signs
education bills Education Week. Retrieved from:
http://blogs.edweek.org/edweek/state_edwatch/2016/09/gov_chris_christie_calls_teacher
s_union_mafia_signs_education_bills.html
Burnette, D. (2016b, November). Teachers’ unions spend big, reap little in elections. Education
Week. Retrieved from: http://www.edweek.org/ew/articles/2016/11/16/teachers-unions-
spend-big-reap-little-in.html
California Department of Education. (2016). Smarter Balanced scale score ranges. Retrieved
from http://www.cde.ca.gov/ta/tg/ca/sbscalerange.asp.
California Teachers Association. (2013). “Bad evaluation bill defeated, win for students and
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 106

teachers.” Retrieved from https://www.cta.org/en/Issues-and-Action/Legislation/Capitol-
News/Legislative-Wins-for-Students-and-Teachers.aspx.
California Teachers Association. (2017). Collective bargaining. Retrieved from
http://www.cta.org/en/Issues-and-Action/Collective-Bargaining.aspx
Casey, L. (2006). The educational value of democratic voice. In J. Hannaway & A. J. Rotherham
(Eds.), Collective bargaining in education: Negotiating change in today’s schools (pp.
181–202). Cambridge, MA: Harvard Education Press.
Cha, S.-H., & Cohen-Vogel, L. (2011). Why they quit: A focused look at teachers who leave for
other occupations. School Effectiveness and School Improvement, 22(4), 371–392.
Chetty, R., Friedman, J. N., & Rockoff, J. E. (2014). Measuring the impacts of teachers:
Evaluating bias in teacher value-added estimates. American Economic Review, 104(9):
2593-2632.
Cibulka, J. (2001). The changing role of interest groups in education: Nationalisation and the
new politics of education productivity. Education Policy, 15(1), 12–40.
Clotfelter, C. T., Ladd, H. F., & Vigdor, J. (2005). Who teaches whom? Race and the distribution
of novice teachers. Economics of Education Review, 24(4), 377–392.
Clotfelter, C. T., Ladd, H. F., & Vigdor. J. L. (2007). Teacher credentials and student
achievement: Longitudinal analysis with student fixed effects. Economics of Education
Review, 26 (6): 673-682.
Clotfelter, C. T., Ladd, H. F., & Vigdor. J. L. (2010). Teacher credentials and student
achievement in high school: A cross-subject analysis with student fixed effects. Journal
of Human Resources, 45(3): 655-681.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 107

Cohen-Vogel, L., & Osborne-Lampkin, L. (2007). Allocating quality: Collective bargaining
agreements and administrative discretion over teacher assignment. Educational
Administration Quarterly, 43(4), 433–461.
Cohen-Vogel, L., Feng, L., & Osborne-Lampkin, L. (2013). Seniority provisions in collective
bargaining agreements and the “teacher quality gap.” Educational Evaluation and Policy
Analysis, 35(3), 324–343.
Conley, D.T. (2003). Who governs our schools? New York, NY: Teachers College Press.
Cooper, B. S., & Sureau, J. (2008). Teacher unions and the politics of fear in labor relations.
Educational Policy, 22(1), 86–105.
Cowen, J. M., & Strunk, K. O. (2015). The impact of teachers’ unions on educational outcomes:
What we know and what we need to learn. Economics of Education Review, 48, 208–223.
de Figueiredo, J. M. (2002). Lobbying and information in politics. Business and Politics, 4(2),
125–129.
de Figueiredo, J. M., & Richter, B. K. (2014). Advancing the empirical research on lobbying.
Annual Review of Political Science, 17(1), 163–185
Drope, J. M., & Hansen, W. L. (2004). Purchasing Protection? The Effect of Political Spending
on U.S. Trade Policy. Political Research Quarterly, 57(1), 27–37.
Eberts, R., & Stone, J. (1987). Teacher unions and the productivity of public schools. Industrial
and Labor Relations Review, 40(3), 354–363.
Edwards, G.C., Barrett, A., & Peake, J. (1997). The legislative impact of divided government.
American Journal of Political Science, 41:545- 63.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 108

Esterling, K. M. (2007). Buying expertise: Campaign contributions and attention to policy
analysis in congressional committees. American Political Science Review, 101(01), 93–
109.
Evans, D. (1996). Before the roll call: Interest group lobbying and public policy outcomes in
house committees. Political Research Quarterly, 49(2), 287–304.
Feng, L. (2010). Hire today, gone tomorrow: New teacher classroom assignments and teacher
mobility. Education Finance and Policy, 5(3), 278–316.
Finger, L. K. (2017). Vested interests and the diffusion of education reform across the states.
Policy Studies Journal.
Fusarelli, B.C., & Fusarelli, L.D. (2015). “Federal education policy from Reagan to Obama:
Convergence, divergence, and ‘control’.” In B.S. Cooper, J.G. Cibulka, & L.D. Fusarelli
(Eds.), Handbook of Education Politics and Policy, (pp. 189–210), New York:
Routledge.
Foundation Center. (2016). IRS 990 finder. Retrieved from http://foundationcenter.org/find-
funding/990-finder
Freeman, R. B., & Han, E. (2012). The war against public sector collective bargaining in the US.
Journal of Industrial Relations, 54(3), 386–408.
Galey, S. (2015). Education politics and policy: Emerging institutions, interests, and ideas.
Policy Studies Journal, 43(S1), S12–S39.
Ghianni, T. (2011, June, 1). Tennessee limits collective bargaining rights for teachers. Reuters.
Retrieved from http://www.reuters.com/article/us-unions-states-tennessee-
idUSTRE75071I20110601.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 109

Glazerman, S., & Max, J. (2011). Do low income students have equal access to the highest-
performing teachers?, Retrieved from http://files.eric.ed.gov/fulltext/ED517966.pdf.
Goertz, M.E. (2009). Standards-based reform: Lessons from the past, directions for the future.”
In K.K. Wong, & R. Rothman, Clio at the Table: Using History to Inform and Improve
Education Policy, (pps. 201-219). New York: Peter Lang Publishing.
Goldhaber, D. (2007). Everyone’s doing it, but what does teacher testing tell us about teacher
effectiveness? Journal of Human Resources, 42(4): 765-794.
Goldhaber, D. & Brewer, D. (2000). Does teacher certification matter? High school teacher
certification status and student schievement. Educational Evaluation and Policy Analysis,
22(2): 129-145.
Goldhaber, D., & Theobald, R. (2013). Managing the teacher workforce in austere times: The
determinants and implications of teacher layoffs. Education Finance and Policy, 8(4),
494–527.
Goldhaber, D., Lavery, L., Theobald, R., D’Entremont, D., & Fang, Y. (2013). Teacher
collective bargaining: Assessing the internal validity of partial independence item
response measures of contract restrictiveness. SAGE Open, 3(2).
Goldhaber, D., Lavery, L., & Theobald, R. (2014). My End of the bargain: Are there cross-
district effects in teacher contract provisions ? ILR Review, 67(4), 1274-1305.
Goldhaber, D., Lavery, L., & Theobald, R. (2015). Uneven playing field? Assessing the teacher
quality gap between advantaged and disadvantaged students. Educational Researcher,
44(5), 293–307.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 110

Goldhaber, D., Lavery, L., & Theobald, R. (2016). Inconvenient truth? Do collective bargaining
agreements help explain the mobility of teachers within school districts? Journal of
Policy Analysis and Management.
Goldhaber, D., Strunk, K. O., Brown, N., & Knight, D. S. (2016). Lessons learned from the
Great Recession: layoffs and the RIF-induced teacher shuffle. Educational Evaluation
and Policy Analysis, 20(3), 517-548.
Goldschmidt, S., & Stuart, L. (1986). The extent and impact of educational policy bargaining.
Industrial and Labor Relations Review, 39(3), 350–360.
Goldstein, D. (2014). The teacher wars: A history of America’s most embattled profession. New
York: Doubleday.
Giersch, J. (2014). Aiming for giants: Charter school legislation and the power of teacher unions.
Education and Urban Society, 46(6), 653–671.
Gray, V., & Lowery, D. (1993). Stability and change in state interest group systems, 1975-1990.
State & Local Government Review, 25(2), 87–96.
Gray, V., & Lowery, D. (1995). Interest representation and democratic gridlock. Legislative
Studies Quarterly, 20(4), 531–552.
Grissom, J. A., Loeb, S., & Nakashima, N. A. (2014). Strategic involuntary teacher transfers and
teacher performance: Examining equity and efficiency. Journal of Policy Analysis and
Management, 33(4), 112–140.
Grossman, G. M., & Helpman, E. (1994). Protection for sale. The American Economic Review,
84(4), 833–850.
Grossman, G. M., & Helpman, E. (1996). Electoral competition and special interest politics. The
Review of Economic Studies, 63(2), 265.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 111

Grunwald, M. (2015, October). Obama vs. teachers unions: It’s still on. http://www.politico.com
/agenda/story/2015/10/obama-vs-teachers-unions-its-stilll-on-000264
Hall, R. L., & Deardorff, A. V. (2006). Lobbying as legislative subsidy. American Political
Science Review, 100(1), 69–84.
Hall, R. L., & Wayman, F. W. (1990). Buying time : Moneyed interests and the mobilization of
bias in congressional committees. Political Science, 84(3), 797–820.
Hansen, E. R., & Gray, V. (2016). Interest group density and policy change in the states.
Retrieved from
http://hansen.web.unc.edu/files/2014/12/Hansen_Gray_Density_Policy_Change.pdf.
Hanushek, E. A. (1986). The economics of schooling: Production and efficiency and public
schools. Journal of Economic Literature, 24(3), 1141–1177.
Hanushek, E. A., Kain, J. F., & Rivkin, S. G. (2004). Why public schools lose teachers. The
Journal of Human Resources, 39(2), 326–354.
Harris, D. N., & Sass, T. 2011. Teacher training, teacher quality and student achievement.
Journal of Public Economics, 95, 798–812.
Hartney, M., & Flavin, P. (2011). From the schoolhouse to the statehouse: Teacher union
political activism and U.S. state education reform policy. State Politics & Policy
Quarterly, 11(3), 251–268.
Hess, F. M., & Kelly, A. P. (2006). Scapegoat, albatross, or what? The status quo in teacher
collective bargaining. In J. Hannaway & A. Rotherham (Eds.), Collective bargaining in
education: Negotiating change in today’s schools (pp. 53–88). Cambridge, MA: Harvard
Education Press.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 112

Hess, F., & Loup, C. (2008). The leadership limbo: Teacher labor agreements in America’s fifty
largest school districts. Retrieved from http://eric.ed.gov/?id=ED500316
Hess, F., & West, M. (2006). A better bargain: Overhauling teacher collective bargaining for
the 21st Century. Retrieved from http://eric.ed.gov/?id=ED498038
Hill, P. T. (2006). The cost of collecting bargaining agreements and related district policies. In J.
Hannaway & A. J. Rotherham (Eds.), Collective bargaining in education: Negotiating
change in today’s schools (pp. 89–110). Cambridge, MA.
Hirsch, B. & Macpherson, D. (2015). Union Membership and Coverage Database from the
CPS [Data file]. Retrieved from http://www.unionstats.com/
Hojnacki, M., & Kimball, D. (1998). Organized interests and the decision of whom to lobby in
Congress. American Political Science Review, 92(4), 775–790.
Hojnacki, M., & Kimball, D. C. (2001). Lobbying contacts in congressional committees.
Political Research Quarterly, 54(1), 161–180.
Hwang, S. D., & Gray, V. (1991). External limits and internal determinants of state public
policy. Western Political Quarterly, 44(2), 277-298.
Isenberg, E., Max, J., Gleason, P., Potamites, L., Santillano, R., Hock, H., & Hansen, M. (2013).
Access to effective teaching for disadvantaged students. Retrieved from
https://ies.ed.gov/ncee/pubs/20144001/pdf/20144001.pdf
Jackson, K., & Bruegmann, E. (2009). Teaching students and teaching each Other: The
importance of peer learning for teachers. American Economic Journal: Applied
Economics, 1(4), 85–108.
Jacob, B. A. (2007). The challenge of staffing urban schools with effective teachers. Future of
Children, 17(1), 129–153.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 113

Jenkins-Smith, H. C., Nohrstedt, D., Weible, C. M., & Sabatier, P. A. (2014). The advocacy
coalition framework: Foundations, evolution, and ongoing research. In P.A. Sabatier &
C.M. Weible (Eds.), Theories of the Policy Process (pp. 183–223). Boulder, CO:
Westview Press
Johnson, M. (2017). Organizing for North Carolina: Social movement unionism in a southern
state. Peabody Journal of Education, 92(1), 127–140.
Johnson, S. M., & Donaldson, M. L. (2006). The effects of collective bargaining on teacher
quality. In J. Hannaway & A. J. Rotherham (Eds.), Collective bargaining in education:
Negotiating change in today’s schools (pp. 111–140). Cambridge, MA: Harvard
Education Press.
Johnson, S. M., Nelson, N. C. W., & Potter, J. (1985). Teachers Unions, School Staffing, and
Reform.
Jones, B. D., & Baumgartner, F. R. (2012). From there to here: Punctuated equilibrium to the
general punctuation thesis to a theory of government information processing. Policy
Studies Journal, 40(1), 1–20.
Julnes, G. (1999). Principal component analysis, factor analysis, and cluster analysis. In G. J.
Miller & M. L. Whicker (Eds.), Handbook of research methods in public administration
(pp. 549-598). New York, NY: Marcel Dekker.
Kalla, J. L., & Broockman, D. E. (2016). Campaign contributions facilitate access to
congressional officials: A randomized field experiment. American Journal of Political
Science, 60(3), 545–558.
Kelley, C., & Mead, J. (2017). Revolution and counter-revolution: Network mobilization to
preserve public education in Wisconsin. Peabody Journal of Education, 92(1), 103–114.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 114

Kleiner, M. M., & Petree, D. L. (1988). Unionism and licensing of public school teachers:
Impact on wages and educational output. In R. B. Freeman, & C. Ichniowski (Eds.),
When public sector workers organize. Chicago, IL: University of Chicago Press.
Kingdon, J.W. (2003). Agendas, alternatives, and public policies (2nd ed.). New York, NY:
Longman Press.
Knight, D. S., & Strunk, K. O. (2016). Who bears the costs of district funding cuts? Reducing
inequality in the distribution of teacher layoffs. Educational Researcher, 45(7), 395–406.
Koski, W. S., & Horng, E. L. (2007). Facilitating the teacher quality gap? Collective bargaining
agreements, teacher hiring and transfer rules, and teacher assignment among schools in
California. Education Finance and Policy, 2(3), 262–300.
Koski, W. S., & Horng, E. L. (2014). Keeping sight of the forest through the trees: Response to
“Collective Bargaining, Transfer Rights, and Disadvantaged Schools”. Educational
Evaluation and Policy Analysis, 36(1), 112-119.
Kraft, M. A. (2015). Teacher layoffs, teacher quality, and student achievement: Evidence from a
discretionary layoff policy. Education Finance and Policy, 10(4), 467–507.
Krutz, G. (2000). Getting around gridlock: The effect of omnibus utilization on legislative
productivity. Legislative Studies Quarterly, 25(4), 533–549.
Kurth, M. M. (1987). Teachers’ unions and excellence in education: An analysis of the decline in
SAT scores. Journal of Labor Research, 8(4), 351– 367.
Ladd, H. F., & Sorensen, L. C. (2016). Returns to teacher experience: Student achievement and
motivation in middle school. Education Finance and Policy.
Lankford, H., Loeb, S., & Wyckoff, J. (2002). Teacher sorting and the plight of urban schools: A
descriptive analysis. Educational Evaluation and Policy Analysis, 24(1), 37–62.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 115

Layton, L. (2015, August, 3). Chris Christie to teachers union: You deserve a punch in the face.
The Washington Post. Retrieved from:
https://www.washingtonpost.com/local/education/chris-christie-to-teachers-union-you-
deserve-a-punch-in-the-face/2015/08/03/86358c2c-39de-11e5-8e98-
115a3cf7d7ae_story.html
Leighley, J. (1996). Group membership and the mobilization of political participation. The
Journal of Politics, 58(2), 447–463.
Leighley, J. E., & Nagler, J. (2007). Voter and class bias in the turnout, unions, electorate, 1964-
2004. Journal of Politics, 69(2), 430–441.
Loewus, L.H. (2014, July). NEA calls for Secretary Duncan’s resignation. Education Week.
Retrieved from
http://blogs.edweek.org/edweek/teacherbeat/2014/07/nea_calls_for_sec_duncans_resi.ht
ml
Lohmann, S. (1995). Information, access, and contributions: A signaling model of lobbying.
Public Choice, 85(3), 267–284.
Lott, J., & Kenny, L. W. (2013). State teacher union strength and student achievement.
Economics of Education Review, 35, 93–103.
http://doi.org/10.1016/j.econedurev.2013.03.006
Lowery, D. (2013). Lobbying influence: Meaning, measurement and missing. Interest Groups &
Advocacy, 2(1), 1–26.
Mahoney, C. (2007). Lobbying Success in the United States and the European Union. Journal of
Public Policy, 27(1), 35–56.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 116

Mawhinney, H. B., & Lugg, C. A. (2001). Introduction: Interest groups in United States
education. Educational Policy, 15(1), 3–11.
Manna, P. (2006). Teachers unions and No Child Left Behind. In J. Hannaway, & A. Rotherham,
Collective Bargaining in Education (pp. 159-180). Cambridge, MA: Harvard Education
Press.
Marianno, B. D. (2015). Teachers’ unions on the defensive?: How recent collective bargaining
laws reformed the rights of teachers. Journal of School Choice, 9(4), 551–577.
Marianno, B. D. & Strunk, K. O. (2018a). The Bad End of the Bargain?: Revisiting the
Relationship between Collective Bargaining Agreements and Student Achievement.
Economics of Education Review.
Marianno, B. D. & Strunk, K. O. (2018b). After Janus: A new era of teachers’ union activism.
Education Next.
Marianno, B.D., Kilbride, T., Theobald, R., Strunk, K.O., Cowen, J., & Goldhaber, D. (2018).
Cut from the same cloth? Comparing urban district CBAs within states and across the
U.S. Educational Policy, 32(2), 334-359.
Maxwell, L.A. (2017, March). School choice a big winner in President Trump’s budget.
Education Week. Retrieved from
http://blogs.edweek.org/edweek/charterschoice/2017/03/school_choice_a_big_winner_in
_president_trumps_budget.html
Mazzoni, T. L. (1995). State policy making and school reform: Influences and influentials. In J.
D. Scribner & D.H. Layton (Eds.), The study of educational politics (pp. 53-73).
Washington, DC: Falmer Press.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 117

McDermott, M. L. (2006). Not for members only: Group endorsements as electoral information
cues. Political Research Quarterly, 59(2), 249–257.
McDonnell, L. M. (2013). Educational accountability and policy feedback. Educational Policy,
27(2), 170–189.
McFarlane, D.R. & Meier, K.J. (2001). The politics of fertility control: Family planning and
abortion policies in the American states. New York: Seven Bridges Press.
Moe, T. (2005). Teacher unions and school board elections. In W. G. Howell (Ed.), Besieged:
School boards and the future of education (pp. 254-287). Washington, D.C.: Brookings.
Moe, T.M. (2006a). Union power and the education of children. In J.Hannway, & A. Rotherham
(Eds.), Collective bargaining in education: Negotiating change in today’s schools (pp.
229–256). Cambridge, MA: Harvard Education Press.
Moe, T. M. (2006b). The union label on the ballot box. Education Next, 6, 59–66.
Moe, T.M. (2006c). Bottom-up structure: Collective bargaining, transfer rights, and the plight of
disadvantaged schools. Stanford University working paper. Available at:
http://eric.ed.gov/?id=ED508944.
Moe, T.M. (2009). Collective bargaining and the performance of public schools. American
Journal of Political Science, 53(1), 156–174.
Moe, T. M. (2011). Special interest: Teachers unions and America’s public schools. Washington
D.C.: Brookings Institution Press.
Moe, T. M. (2015). Vested interests and political institutions. Political Science Quarterly,
130(2), 277–318.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 118

Mulligan, C. B., Gil, R., & Sala-i-Martin, X. (2004). Do democracies have different public
policies than nondemocracies?. Journal of Economic Perspectives, 18(1), 51-74.
Mulvihill, (2017, May). School choice-billionaires at odds. Associated Press. Retrieved from
https://www.usnews.com/news/best-states/washington/articles/2017-05-16/correction-
school-choice-billionaires-at-odds-story
National Conference on State Legislatures. (2018). Collective bargaining and labor union
legislation database [Date File] Retrieved from: http://www.ncsl.org/research/labor-and-
employment/collective-bargaining-legislation-database.aspx
National Council on Teacher Quality. (2016). District policy/state influence fact sheet. Retrieved
from http://www.nctq.org/districtPolicy/stateInfluence.do
National Institute on Money in State Politics. (2016). Follow the Money Database. Retrieved
from https://www.followthemoney.org/.
National Education Association. (2015). Congressional Report Card. Retrieved from
http://www.nea.org/home/65601.htm
National Education Association. (2017a). NEA President: Trump-DeVos budget is a wrecking
ball aimed at public schools. Retrieved from http://www.nea.org/home/70780.htm
National Education Association. (2017b). NEA President comments on Trump budget’s
proposed public education slashes. Retrieved from http://www.nea.org/home/70747.htm.
Nelson, H. F., & Rosen, M. (1996). Are teachers unions hurting American education? A state-
by-state analysis of the impact of collective bargaining among teachers on student
performance. Retrieved from
http://files.eric.ed.gov.libproxy1.usc.edu/fulltext/ED404746.pdf
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 119

Nicholson-Crotty, S., & Carley, S. (2016). Effectiveness, implementation, and policy diffusion:
Or “Can we make that work for us?” State Politics and Policy Quarterly, 16(1), 78–97.
Opfer, V. D. (2001). Beyond self-interest: Educational interest groups and congressional
influence. Educational Policy, 15(1), 135–152.
Opfer, D., Young, T., & Fusarelli, L. (2008). Interest groups and institutional effects on school
choice. Educational Policy, 193–216.
Papay, J. P., & Kraft, M. A. (2015). Productivity returns to experience in the teacher labor
market : Methodological challenges and new evidence on long-term career improvement.
Journal of Public Economics, 130, 105–119.
Paige, R. (2006). The war against hope: How teachers’ unions hurt children, hinder teachers,
and endanger public education. Nashville, TN: Thomas Nelson, Inc.
Peske, H. G., & Haycock, K. (2006). Teaching inequality: How poor and minority students are
shortchanged on teacher quality: A report and recommendations by the education trust.
Retrieved from http://eric.ed.gov/?id=ED494820
Peterson, P. E., Henderson, M., & West, M. R. (2014). Teachers versus the Public: What
Americans Think about Schools and How to Fix Them. Washington D.C. : Brookings
Institution Press.
Pierson, P. (2015). Power and path dependence. In J. Mahoney & K. Thelen (Eds.), Advances in
Comparative-Historical Analysis, (pp. 123–146). Cambridge, MA: Cambridge University
Press.
Radcliff, B. (2001). Organized labor and electoral participation in American national elections.
Journal of Labor Research, 22(2), 405–414.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 120

Radcliff, B., & Davis, P. (2000). Labor organization and electoral participation in industrial
democracies. American Journal of Political Science, 44(1), 132–141.
Ravitch, D. (2006). Why teacher unions are good for teachers—and the public. American
Federation of Teachers. Retrieved from: http://www.aft.org/periodical/american-
educator/winter-2006-2007/why-teacher-unions-are-good-teachers-and
Reardon, S. F. (2011). The widening academic achievement gap between the rich and the poor:
New evidence and possible explanations. In G.J. Duncan & R.J. Murnane (Eds.), Whither
opportunity? Rising inequality and the uncertain life chances of low-income children (pp.
91–116). New York: Russell Sage Foundation.
Reardon, S. F., & Raudenbush, S. W. (2006). A partial independence item response model for
surveys with filter questions. Sociological Methodology, 36, 257-300.
Reckhow, Sarah. (2013). Follow the money: How foundation dollars change public school
politics. Oxford, UK: Oxford University Press.
Reckhow, S., & Snyder, J. W. (2014). The expanding role of philanthropy in education politics.
Educational Researcher, 43(4), 186–195.
Reininger, M. (2012). Hometown disadvantage? It depends on where you’re from: Teachers’
location preferences and the implications for staffing schools. Educational Evaluation
and Policy Analysis, 34(2), 127-145.
Renzulli, L. A., & Roscigno, V. J. (2005). Charter school policy, implementation, and diffusion
across the United States. Sociology of Education, 78, 344–366.
Rivkin, S. G., Hanushek, E. A., & Kain, J. F. (2005). Teachers, schools, and academic
achievement. Econometrica, 73, 417- 458.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 121

Rockoff, J.E., 2004. The impact of individual teachers on student achievement: evidence from
panel data. American Economic Revie, 94 (2), 247–252.
Rosenthal, A. 1993. The third house: Lobbyists and lobbying. Washington, DC: Congressional
Quarterly.
Rossman, S. (2018, March). Teachers are striking all over. What is going on? USA Today.
Retrieved from https://www.usatoday.com/story/news/nation-now/2018/03/30/teachers-
striking-oklahoma-west-virginia-arizona-kentucky/472742002/.
Rothstein, J. (2015). Teacher quality policy when supply matters. American Economic Review,
105(1), 100-130.
Sabatier, P.A. (1988). An advocacy coalition model of policy change and the role of policy?
Oriented learning therein. Policy Sciences, 21: 129–168
Sabatier, P. A., & Jenkins-Smith H. (1988). An Advocacy Coalition Model of policy change and
the role of policy orientated learning therein. Policy Sciences, 21: 129–68.
Sabatier, P. A., & Jenkins-Smith H. (1993). Policy change and learning: An advocacy coalition
approach. Boulder, CO: Westview Press.
Sabatier, P. A., & Jenkins-Smith H. (1999). The Advocacy Coalition Framework: An assessment.
In P.A. Sabatier (Eds.) Theories of the Policy Process. Boulder, CO: Westview Press.
Sabatier, P.A., & Weible, C.M. (2007). The Advocacy Coalition Framework: Innovations and
Clarifications. In P.A. Sabatier (Eds.) Theories of the Policy Process (2
nd
Ed.), (pp. 189–
222). Boulder, CO: Westview Press.
Sass, T. R., Hannaway, J., Xu, Z., Figlio, D. N., & Feng, L. (2012). Value added of teachers in
high-poverty schools and lower poverty schools. Journal of Urban Economics, 72(2–3),
104–122.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 122

Sawchuk, S. (2012, May). New advocacy groups shaking up education field. Retrieved from
https://www.edweek.org/ew/articles/2012/05/16/31adv-overview_ep.h31.html
Sawchuk, S. (2016, July). NEA's membership numbers: Correcting the record. Retrieved from
http://blogs.edweek.org/edweek/teacherbeat/2016/07/neas_membership_numbers_correc.
html?qs=NEA
Scafidi, B., Sjoquist, D. L., & Stinebrickner, T. R. (2007). Race, poverty, and teacher mobility.
Economics of Education Review, 26(2), 145–159.
Shober, A. F., Manna, P., & Witte, J. F. (2006). Flexibility meets accountability: State charter
school laws and their influence on the formation of charter schools in the United States.
Policy Studies Journal, 34(4), 563–587.
Smith, L. (2009, December). D.C. Schools chief Michelle Rhee fights union over teacher pay.
US News. Retrieved from https://www.usnews.com/news/national/articles/2009/12/21/dc-
schools-chief-michelle-rhee-fights-union-over-teacher-pay
Snyder, J.M., 1990. Campaign contributions as investments: The U.S. House of Representatives,
1980–1986. Journal of Political Economy, 98 (6), 1195–1227.
Snyder, J.M., 1991. On buying legislatures. Economics and Politics, 3 (2), 93–109.
Steelman, L. C., Powell, B., & Carini, R. M. (2000). Do teacher unions hinder educational
performance?: Lessons learned from state SAT and ACT scores. Harvard Educational
Review. 70(4): 437-466.
Stoddard, C., & Corcoran, S. P. (2007). The political economy of school choice: Support for
charter schools across states and school districts. Journal of Urban Economics, 62(1),
27–54.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 123

Strauss, V. (2017, July). Teachers union leader: We won’t work with Trump and DeVos because
‘I do not trust their motives’. Washington Post. Retrieved from
https://www.washingtonpost.com/news/answer-sheet/wp/2017/07/03/teachers-union-
leader-we-wont-work-with-trump-and-devos-because-i-do-not-trust-their-
motives/?utm_term=.49e3e9ba0ed9
Strunk, K. O. (2011). Are teachers’ unions really to blame? Collective bargaining agreements
and their relationships with district resource allocation and student performance in
California. Education Finance and Policy, 6(3), 354–398.
Strunk, K. O. (2012). Policy poison or promise: Exploring the dual nature of California school
district collective bargaining agreements. Educational Administration Quarterly, 48(3),
506–547.
Strunk, K. O. (2014). The role of collective bargaining agreements in the implementation of
education reforms: Perils and possibilities. In F.M. Hess & M.Q. McShane (Eds.)
Teacher quality 2.0: Toward a new era in education reform (pp. 155–180). Cambridge,
MA: Harvard Education Press.
Strunk, K. O., Barrett, N., & Lincove, J. A. (2017). When tenure ends: Teacher turnover in
response to policy changes in Louisiana. Retrieved from
http://educationresearchalliancenola.org/files/publications/022217-Strunk-Barrett-
Arnold-Lincove-When-Tenure-Ends-Teacher-Turnover-in-Response-to-Policy-Changes-
in-Louisiana.pdf
Strunk, K. O., Goldhaber, D., Knight, D.S. & Brown, N. (2018). Are there hidden costs
associated with conducting layoffs? The impact of RIFS and layoffs on teacher
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 124

effectiveness. Retrieved from
https://caldercenter.org/sites/default/files/WP%20140_0.pdf
Strunk, K. O., & Grissom, J. A. (2010). Do strong unions shape district policies?: Collective
bargaining, teacher contract restrictiveness, and the political power of teachers’ unions.
Educational Evaluation and Policy Analysis, 32(3), 389–406.
Strunk, K. O., & McEachin, A. (2011). Accountability under constraint: The relationship
between collective bargaining agreements and California schools’ and districts'
performance under No Child Left Behind. American Educational Research Journal,
48(4), 871–903.
Strunk, K. O., & Reardon, S. F. (2010). Measuring the strength of teachers’ unions: An empirical
application of the partial independence item response approach. Journal of Educational
and Behavioral Statistics, 35(6), 629–670.
Superfine, B. M., & Thompson, A. R. (2016). Interest groups, the courts, and educational
equality: A Policy regimes approach to Vergara v. California. American Educational
Research Journal, 53(3), 573–604.
Texas State Teachers Association (2013). Summary of the 83
rd
Legislature. Retrieved from
http://tsta.org/sites/default/files/83rdLegislativeSessionReport_0.pdf.
Verba, S., Scholzman, K. H., & Brady, H. (1995). Voice and Equality. Cambridge: Harvard
University Press.
Walker, S. (2013). Unintimidated: A governor’s story and a nation’s challenge. New York, NY:
Sentinel
Weible, C.M., Sabatier, P.A. & McQueen, K. (2009). Themes and variations: Taking stock of the
Advocacy Coalition Framework.” Policy Studies Journal, 37(1): 121–40.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 125

Weiss, E.M. (1999). Perceived workplace conditions and first-year teachers’ morale, career
choice commitment, and planned retention: A secondary analysis. Teaching and Teacher
Education, 15, 861–879.
Winkler, A. M., Scull, J., & Zeehandelaar, D. (2012). How strong are U.S. teacher unions: A
state-by-state comparison. Washington, DC: Fordham Institute
Wiswall, M. (2013). The dynamics of teacher quality. Journal of Public Economics, 100, 61–78.
Young, T. V. (2011). Teachers unions in turbulent times: Maintaining their niche. Peabody
Journal of Education, 86(1), 338–351.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 126

Figures and Tables

Notes: A/B and D/F lawmakers are defined as those that received a A/B or a D/F on the NEA’s
Legislative Report Card. The year on the horizontal axis denotes the last year of the congressional
session (e.g. 2010 represents the 2009-2010 congressional session). The proportion of Democrats
and proportion of Republicans shown correspond with the partisan makeup of that congressional
session.
Figure 1. The Proportion of NEA A/B and D/F Legislators in Congress Over Time
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 127

Figure 2. NEA Election Contribution Rates Over Time
Notes: A/B and D/F lawmakers are defined as those that received a A/B or a D/F on the NEA’s
Legislative Report Card. The year on the horizontal axis denotes the election year. The vertical
axis represents the contribution rate or the proportion of candidates the NEA contributed to in a
given election year. Senate A/B Dem Cont rate represents the contribution rate to A/B Democrat
incumbents in the Senate. Senate D/F Dem Chall Cont rate represents the contribution rate to
Senate Democrat challengers to D/F incumbents.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 128

Figure 3. NEA Election Success Rates Over Time
Notes: A/B and D/F lawmakers are defined as those that received a A/B or a D/F on the NEA’s
Legislative Report Card. The year on the upper horizontal axis denotes the election year. The year
on the lower horizontal axis denotes the last year of the congressional session (e.g. 2010
represents the 2009-2010 congressional session. The vertical axis represents the NEA’s legislative
or election success rate. Election A/B Dem Cont rate represents the election success rate of NEA-
supported A/B Democrat incumbents. Election D/F Dem Chall Cont rate represents the election
success rate of NEA-supported Democrat challengers to D/F incumbents.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 129

Table 1. Change in the Distribution of A/B, C, and D/F Seats in the House and Senate by Lawmaker Party and State Partisanship

Overall
(1)
Firmly Democrat State
(2)
Firmly Republican State
(3)
Swing State
(4)
A/B C D/F A/B C D/F A/B C D/F A/B C D/F
# ∆ # ∆ # ∆ # ∆ # ∆ # ∆ # ∆ # ∆ # ∆ # ∆ # ∆ # ∆
A. Senate
Overall
Overall -8 8 2 -1 0 2 -5 6 0 -2 2 0
Dem -12 0 0 -1 0 0 -8 0 0 -3 0 0
Rep 4 8 2 0 0 2 3 6 0 1 2 0
∆ 2010 to
2012
Overall -5 1 5 -3 0 3 -1 1 1 -1 0 1
Dem -5 0 0 -3 0 0 -1 0 0 -1 0 0
Rep 0 1 5 0 0 3 0 1 1 0 0 1
∆ 2012 to
2014
Overall 2 8 -10 2 0 -1 -2 7 -6 2 1 -3
Dem 1 0 0 2 0 0 -2 0 0 1 0 0
Rep 1 8 -10 0 0 -1 0 7 -6 1 1 -3
∆ 2014 to
2016
Overall -5 -1 7 0 0 0 -2 -2 5 -3 1 2
Dem -8 0 0 0 0 0 -5 0 0 -3 0 0
Rep 3 -1 7 0 0 0 3 -2 5 0 1 2
B. House
Overall
Overall -35 12 28 1 7 -11 -18 2 23 -18 3 16
Dem -61 -1 -2 -17 0 0 -21 -1 -2 -23 0 0
Rep 26 13 30 18 7 -11 3 3 25 5 3 16
∆ 2010 to
2012
Overall -45 8 41 -11 3 10 -17 3 15 -17 2 16
Dem -57 -1 -2 -19 0 0 -18 -1 -2 -20 0 0
Rep 12 9 43 8 3 10 1 4 17 3 2 16
∆ 2012 to
2014
Overall 4 1 -7 4 3 -13 0 0 6 0 -2 0
Dem 8 0 0 6 0 0 2 0 0 0 0 0
Rep -4 1 -7 -2 3 -13 -2 0 6 0 -2 0
∆ 2014 to
2016
Overall 6 3 -6 8 1 -8 -1 -1 2 -1 3 0
Dem -12 0 0 -4 0 0 -5 0 0 -3 0 0
Rep 18 3 -6 12 1 -8 4 -1 2 2 3 0
Notes: The overall row represents the change in seats from the 2009-2010 congressional session to the 2015-2016 congressional session. Changes between
adjoining sessions are shown below the overall rows (e.g. ∆ 2010 to 2012 represents the change in seats from the 2009-2010 congressional session to the 2011-
2012 congressional changes). Changes between the observed sessions occurred as a result of the 2010, 2012, and 2014 elections. The changes are shown by
candidate party (also on the rows) and state partisanship (on the columns). The partisan political affiliation of the state was determined by the partisan vote share
received by Democrat and Republican presidential candidates since the 2000 election (e.g. Republican states are defined as those that consistently voted majority
Republican in presidential elections from 2000 to 2012).
The Effects of Teachers' Unions on Congress, Statehouses, and Schools

130

Table 2. NEA, School Choice Advocate, Anti-Labor, and Business Campaign Contributions Per Candidate for
Senate and House “A/B” and “D/F” Races (2010, 2012, 2014 elections)
Senate House
Race Cont.
Overall
(1)
2010
(2)
2012
(3)
2014
(4)
Overall
(5)
2010
(6)
2012
(7)
2014
(8)
Dem
A/B
NEA 8,726 8,267 8,925 8,889 5,788 6,200 5,150 5,725
SC 13,319 16,333 11,409 12,928 2,571 2,670 2,748 2,248
AL 0 0 0 0 0 0 0 0
BU 334,933 408,806 293,362 319,563 70,482 69,570 71,983 70,537
Rep Chall.
NEA 0 0 0 0 0 0 0 0
SC 5,301 4,612 4,432 6,729 247 195 354 230
AL 510 676 243 643 23 41 1 14
BU 154,920 184,987 109,331 176,168 19,215 24,965 14,009 14,932
Rep
A/B
NEA 1,000 3,000 0 0 3,500 2,125 2,714 5,000
SC 17,134 25,800 9,903 15,700 3,939 2,500 3,821 4,611
AL 0 0 0 0 0 0 0 0
BU 298,372 326,953 259,696 308,466 126,508 62,963 139,914 132,552
Dem
Chall
NEA 3,333 0 10,000 0 517 0 1,071 0
SC 6,567 0 14,500 5,200 428 500 0 945
AL 0 0 0 0 0 0 0 0
BU 168,165 25,649 473,693 5,155 17,658 20,002 10,500 25,916
Rep
D/F
NEA 267 250 0 455 112 337 6 38
SC 16,394 15,200 23,219 13,355 2,656 2,837 2,858 2,282
AL 1,170 1,567 7,57 1,000 71 32 68 106
BU 415,411 393,876 459,651 41,0751 105,041 85,128 117,839 107,900
Dem
Chall
NEA 2,641 2,875 3,333 1,818 966 480 1,606 678
SC 3,117 4,267 2,762 2,155 26 8 48 18
AL 0 0 0 0 0 0 0 0
BU 71,604 114,303 34,104 55,706 6,137 3,668 9,802 4,218
Notes: A/B and D/F incumbents are defined as those that received a A/B or a D/F on the NEA’s legislative
report card for the congressional session preceding the election. A/B and D/F challengers are those running in
the general election against a A/B or a D/F candidate. Contributions include all election-related contributions
from individuals and non-individuals and their PACs to candidates for federal office and their PACs.
Contributions amounts are shown per candidate. Races with Democrat “D/F” candidates are not reported
because no Democrat “D/F” candidates exist in the Senate.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 131

Table 3. Topic and NEA Position on Votes Scored in NEA Legislative Report Card (2009-10, 2011-12, 2013-14, 2015-16)
Senate House
Overall 2010 2012 2014 2016 Overall 2010 2012 2014 2016
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
#S #O
SR
%
Total 50 18 79 11 3 93 8 4 75 15 3 94 16 8 63 42 38 63 20 0 100 4 15 47 9 13 41 9 10 63
Budget/Funding 11 8 68 1 1 50 1 2 67 4 2 83 5 3 63 8 13 57 2 0 100 1 4 40 3 6 44 2 3 80
Higher Education 6 1 100 3 0 100 1 1 100 2 0 100 0 0 - 6 3 67 4 0 100 1 1 50 0 1 0 1 1 50
Health Care 5 2 86 3 0 100 0 1 100 0 0 - 2 1 67 5 4 56 4 0 100 0 2 0 0 0 - 1 2 33
Social Issues 7 0 100 2 0 100 2 0 100 2 0 100 1 0 100 4 1 80 2 0 100 0 1 0 1 0 100 1 0 100
School Choice 0 4 100 0 2 100 0 0 - 0 0 - 0 2 100 3 5 38 0 0 - 0 3 67 3 0 33 0 2 0
Union Rights/Dues 0 3 67 0 0 - 0 1 100 0 1 100 0 1 0 2 4 67 1 0 100 0 2 50 1 0 100 0 2 50
Tax Cuts 2 0 100 0 0 - 2 0 100 0 0 - 0 0 - 2 1 33 0 0 - 2 0 50 0 1 0 0 0 -
Compensation/Benefits 2 0 100 1 0 100 0 0 - 1 0 100 0 0 - 2 0 100 1 0 100 1 0 100 0 0 - 0 0 -
Wages 1 0 0 0 0 - 1 0 0 0 0 - 0 0 - 0 3 100 0 0 - 0 2 100 0 1 100 0 0 -
Education Job Creation 2 0 50 1 0 100 1 0 0 0 0 - 0 0 - 2 0 100 2 0 100 0 0 - 0 0 - 0 0 -
School Modernization 1 0 100 1 0 100 0 0 - 0 0 - 0 0 - 3 0 100 3 0 100 0 0 - 0 0 - 0 0 -
Gun Violence
Prevention
3 0 33 0 0 - 0 0 - 1 0 100 2 0 0 1 0 100 0 0 - 0 0 - 1 0 100 0 0 -
Immigration Reform 1 1 50 0 0 - 0 0 - 1 0 100 0 1 0 0 2 0 0 0 - 0 0 - 0 1 0 0 1 0
Notes: The NEA Legislative Report Card tracks legislative votes on bill amendments and final bills. #S represents the number of votes supported by the NEA. #O represents the number of votes opposed by the
NEA. The success rate (SR) is defined as the number of vote outcomes (pass/fail) that supported the NEA official position divided by the total number of votes taken. The first row reflects the total number of
votes taken in the Senate and House supported and opposed by the NEA over time. The subsequent rows reflect the totals for topics with at least three combined votes in the House and Senate. A small
percentage of votes cover more than one topic. For these reasons, totaling across topics in each column will not always reflect the total shown in the first row, which covers all votes taken and does not double
count votes covering more than one topic.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 132

Table 4. Examples of Coded Legislation
Bill Number State Year Topic
Final
Action
Summary Stated NEA Position Law Intent
S.B. 441 CA 2013 Evaluation Failed
Relates to evaluation of certificated education
employees assigned as classroom teachers who have
been employed for a specified time. Provides for rating
levels. Requires the governing board to avail itself of
the advice of parents of pupils. Change permanent
employees who have been employed in the district for
ten years to a three-year evaluation cycle instead of a
five-year evaluation cycle.
Oppose
(“Bad Evaluation
Bill Defeated, Win
for Students and
Teachers”-
California Teachers’
Association, 2013)
Unfavorable
S.B. 866 VA 2015
Salary &
Benefits
Failed
Relates to health insurance for local school board
employees; allows local school boards to elect to have
all their employees and retirees, as well as the
dependents of employees and retirees, eligible to
participate in the state employee health insurance plan
in lieu of the current state- administered local health
insurance plan; provides that the local school board
shall be responsible for whatever portion of the cost of
such insurance is not paid by the employee.
Oppose
(Virginia Education
Association
Legislative Report
Card, 2015)
Unfavorable
S.B. 1587 NJ 2014 Retirement Failed
Requires spousal consent to election of certain pension
payout options under Teachers' Pension and Annuity
Fund, Judicial Retirement System and Public
Employees Retirement System.
No Official Position Neutral
H.B. 30 UT 2015
Salary &
Benefits
Enacted
Relates to math teacher training; expands a grant
program for teacher training in math by allowing a grant
to be used to provide a stipend, professional
development, and leadership opportunities to an
experienced mathematics teacher to assist the teacher in
becoming a teacher leader.
Support
(Utah Education
Association
Legislative Tracking
Sheet, 2015)
Favorable
S.B. 1458 TX 2013 Retirement Enacted
Relates to the administration of and benefits payable by
the Teacher Retirement System of Texas; provides an
optional group health plan for retirees.
Support
(Texas State
Teachers’
Association
Summary of the 83
rd

Legislature, 2013
Favorable
The Effects of Teachers' Unions on Congress, Statehouses, and Schools

133

Table 5. Summary Statistics for Dependent and Independent Variables
Dependent Variables
2011
Mean
(SD)
2012
Mean
(SD)
2013
Mean
(SD)
2014
Mean
(SD)
2015
Mean
(SD)
All Years
Mean
(SD)
Agenda-Setting
Proportion Unfavorable Laws Proposed
55.012
(21.372)
48.217
(28.680)
47.278
(23.131)
30.596
(25.196)
44.356
(21.104)
45.092
(25.191)
Proportion Favorable Laws Proposed
37.588
(18.894)
37.496
(24.412)
50.142
(22.933)
52.653
(29.953)
55.080
(20.643)
46.592
(24.672)
Legislative Success
Unfavorable Law Success Rate
71.472
(26.654)
65.999
(38.820)
78.654
(28.124)
61.962
(43.016)
77.955
(31.075)
71.208
(34.483)
Favorable Law Success Rate
19.622
(25.756)
14.440
(24.240)
23.971
(23.762)
27.703
(33.448)
21.868
(29.711)
21.521
(27.764)
Independent Variables
Partisan Control
Republican Control 0.520 0.560 0.540 0.560 0.620 0.560
Democrat Control 0.320 0.300 0.380 0.380 0.220 0.320
Split Control 0.160 0.140 0.080 0.060 0.160 0.120
Legislative Cycle
Session Calendar Days (ln)
4.917
(0.493)
4.369
(1.442)
4.893
(0.585)
4.221
(1.436)
4.932
(0.599)
4.666
(1.047)
No Laws Proposed 0.020 0.014 0.000 0.120 0.000 0.056
Fiscal and Economic Health
2010
Mean
(SD)
2011
Mean
(SD)
2012
Mean
(SD)
2013
Mean
(SD)
2014
Mean
(SD)
All Years
Mean
(SD)
Debt-to-Service Ratio
54.051
(25.304)
49.362
(24.090)
58.018
(28.846)
49.886
(25.358)
47.148
(24.151)
51.040
(25.598)
Unemployment Rate
8.946
(8.946)
8.468
(2.064)
7.538
(1.867)
7.050
(1.646)
6.104
(1.425)
6.891
(1.987)
Education System Performance/Size
NAEP Math Score
240.3867
(5.462)
240.850
(5.256)
241.504
(5.157)
242.158
(5.222)
241.285
(4.878)
241.242
(5.051)
Student Enrollment (ln)
13.301
(1.039)
13.305
(1.035)
13.305
(1.036)
13.311
(1.035)
13.316
(1.035)
13.312
(1.027)

The Effects of Teachers' Unions on Congress, Statehouses, and Schools

134

Table 6. Sources for Measures of Teachers’ Union and Opposition Group Influence
Measure Years Data Source
Selection of Studies Using
Measure
Election Influence
Proportion of Open Seats
Won by Interest Group Allies
Election
Years
(2010-2015)
National Institute on
Money in State
Politics, Follow the
Money Database

Contributions Per Candidate
Election
Years
(2010-2015)
National Institute on
Money in State
Politics, Follow the
Money Database

Proportion of Contributions
from Interest Group
Election
Years
(2010-2015)
National Institute on
Money in State
Politics, Follow the
Money Database
Hartney & Flavin (2011); Winkler,
Scull, & Zeehandelaar (2012)
Lobbing Influence
Lobby Organization Density 2010-2015
National Institute on
Money in State
Politics, Follow the
Money Database
Gray & Lowery (1993, 1995)
Lobbyist Per Lawmaker 2010-2015
National Institute on
Money in State
Politics, Follow the
Money Database

Membership Influence (Teachers’
Unions Only)

Teachers’ Union Membership
Dues Per Teacher
2010-2015
IRS 990 tax forms,
Foundation Center
Lott & Kenny (2013)
Teachers’ Union Spending
Per Student
2010-2015
IRS 990 tax forms,
Foundation Center
Lott & Kenny (2013); Winkler,
Scull, & Zeehandelaar (2012)
NEA Membership Rate 2010-2015 Antonucci (2017)
Finger (2017); Kleiner & Petree
(1988); Renzulli and Roscigno,
(2005); Shober, Manna, and Witte,
(2006); Stoddard & Corcoran,
(2007); Winkler, Scull, &
Zeehandelaar (2012)
Proportion Covered by
Collective Bargaining
Agreement
2010-2015
Hirsch and
Macpherson (2015)
Kleiner & Petree (1988); Kurth
(1987); Nelson & Rosen (1996);
Steelman, Powell, & Carini (2000)
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 135

Table 7. Summary Statistics for All Interest Group Strength Variables

2010
Mean
(SD)
2011
Mean
(SD)
2012
Mean
(SD)
2013
Mean
(SD)
2014
Mean
(SD)
All Years
Mean
(SD)
Teachers’ Union Strength
Overall Teachers’ Union Strength
Election Influence
Proportion of Open Seats Won by Teachers’ Union Allies
44.641
(22.177)
43.129
(21.428)
43.320
(21.236)
43.663
(21.746)
36.841
(21.440)
42.300
(21.609)
Teachers’ Unions Contributions Per Candidate
1422.864
(2176.431)
1428.621
(2144.503)
1783.199
(2978.382)
1801.962
(2987.728)
1885.345
(3764.452)
1666.346
(2862.645)
Proportion of Contributions from Teachers’ Union
1.446
(1.179)
1.397
(1.171)
1.463
(1.474)
1.488
(1.478)
1.394
(1.667)
1.438
(1.398)
Lobby Strength
Teachers’ Union Lobby Organization Density
0.647
(0.590)
0.612
(0.421)
0.536
(0.383)
0.527
(0.351)
0.510
(0.332)
0.566
(0.426)
# Teachers’ Union Lobbyist Per Lawmaker
0.090
(0.076)
0.085
(0.075)
0.084
(0.080)
0.081
(0.080)
0.089
(0.082)
0.086
(0.078)
Revenue, Dues, Coverage, and Membership Size
Teachers’ Union Membership Dues Per Teacher
319.708
(222.787)
317.119
(231.033)
314.444
(233.441)
316.908
(239.163)
319.445
(244.401)
317.525
(232.399)
Teachers’ Union Spending Per Student
25.530
(17.115)
24.558
(17.885)
23.946
(17.622)
24.441
(18.151)
24.857
(19.159)
24.667
(17.862)
NEA Membership Rate
91.341
(43.317)
88.039
(44.227)
85.782
(44.058)
84.086
(44.600)
83.582
(45.205)
86.566
(44.021)
Proportion Covered by Collective Bargaining Agreement
37.740
(18.337)
36.466
(18.018)
35.796
(18.271)
36.176
(18.862)
36.198
(17.200)
36.475
(18.012)
Opposition Groups
Overall Opposition Group Strength Score
Opposition Group Election Influence
Proportion of Open Seats Won by Business Group Allies
86.278
(23.278)
86.489
(26.853)
83.788
(26.175)
83.850
(26.200)
83.142
(26.615)
84.697
(25.704)
Business Group Contributions Per Candidate
5772.491
(7146.041)
6802.446
(9526.518)
8168.804
(10929.970)
8011.200
(10845.270)
7297.187
(9859.028)
7222.022
(9735.971)
Proportion of Contributions from Business Groups
5.710
(2.948)
5.769
(2.907)
6.064
(2.990)
6.072
(2.978)
5.603
(3.206)
5.845
(2.990)
Proportion of Open Seats Won by School Choice Group Allies
6.891
(11.506)
6.719
(11.305)
8.801
(13.493)
8.960
(13.423)
13.641
(19.798)
9.019
(14.368)
School Choice Group Contributions Per Candidate
199.737
(657.426)
211.297
(652.844)
291.968
(782.010)
293.757
(781.412)
395.095
(775.060)
279.005
(730.321)
Proportion of Contributions from School Choice Groups
0.164
(0.588)
0.162
(0.576)
0.195
(0.539)
0.196
(0.538)
0.247
(0.540)
0.193
(0.553)
Opposition Group Lobby Strength
Business Group Lobby Organization Density
13.506
2.073
13.754
(2.512)
13.965
(2.662)
14.107
(2.319)
13.867
(2.319)
13.840
(2.375)
# Business Group Lobbyist Per Lawmaker
1.197
1.646
1.380
(1.845)
1.202
(1.449)
1.523
(1.908)
1.562
(2.266)
1.373
(1.834)
School Choice Group Lobby Organization Density
0.196
0.263
0.234
(0.256)
0.339
(0.374)
0.349
(0.342)
0.339
(0.375)
0.291
(0.330)
# School Choice Group Lobbyist Per Lawmaker
0.020
0.035
0.033
(0.056)
0.033
(0.040)
0.050
(0.060)
0.045
(0.058)
0.036
(0.051)
Teachers’ Union/Opposition Group Difference Score
Teachers’ Union Strength Relative to Opposition Group Strength
Notes: All rows in grey represent component variables generated from principal component analysis of the underlying variables. Only summary statistics for the
underlying variables are shown because all component variables are standardized within-year to have a mean of zero and a standard deviation of one. The teachers’
union strength relative to opposition group strength variable is created by subtracting a given state’s component score on the overall strength component from their
component score on the overall opposition group strength component.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 136

Table 8. Eigenvalues and Component Loading on Components from Principal Component Analysis on Teachers’ Union Influence
Measures (Year=2011)
Panel A. Overall Teachers’ Union Influence Measure

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 3.865 1.897 0.430 0.430
Factor 2 1.969 0.706 0.219 0.648
Factor 3 1.262 0.367 0.140 0.789
Factor 1 Factor 2 Factor 3 Uniqueness
Teachers’ Union Membership Dues Per Teacher (ln) 0.931 -0.183 -0.114 0.087
Teachers’ Union Spending Per Student (ln) 0.894 -0.342 -0.167 0.055
NEA Membership Rate 0.894 -0.282 0.021 0.120
Proportion Covered by Collective Bargaining Agreement 0.835 -0.317 0.108 0.192
Proportion of Open Seats Won by Teachers’ Union Allies 0.559 0.637 0.006 0.281
Teachers’ Unions Contributions Per Candidate (ln) 0.316 0.869 -0.185 0.111
Proportion of Contributions from Teachers’ Union 0.388 0.592 -0.388 0.350
Teachers’ Union Lobby Organization Density 0.140 0.169 0.702 0.460
# Teachers’ Union Lobbyist Per Lawmaker 0.348 0.314 0.730 0.248
Panel B. Membership/Resource Strength Measure (Year=2011)

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 3.487 3.183 0.872 0.872
Factor 1 Uniqueness
Teachers’ Union Membership Dues Per Teacher (ln) 0.949 0.099
Teachers’ Union Spending Per Student (ln) 0.961 0.076
NEA Membership Rate 0.885 0.218
Proportion Covered by Collective Bargaining Agreement 0.938 0.120
Panel C. Election Influence Measure (Year=2011)

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 2.124 1.490 0.708 0.708
Factor 1 Uniqueness
Proportion of Open Seats Won by Teachers’ Union Allies 0.824 0.321
Teachers’ Unions Contributions Per Candidate (ln) 0.770 0.147
Proportion of Contributions from Teachers’ Union 0.923 0.408
Panel D. Lobbying Influence Measure (Year=2011)

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 1.324 0.647 0.662 0.662
Factor 1 Uniqueness
Teachers’ Union Lobby Organization Density 0.814 0.338
# Teachers’ Union Lobbyist Per Lawmaker 0.814 0.338
Notes: Results from PCA for each component in 2011 are shown in this table. All teachers’ union strength measures are created by
predicting the score for each state in each year on component one. Only factor with eigenvalues over 1.000 are shown in the table. I did
not employ any factor rotation. Given the strong loadings onto factor one for each component, the rotated solutions did not add clarity
on patterns in the data.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 137

Table 9. Bivariate Correlations Between Teachers’ Union and Opposition Group Strength Measures (2011)

Teachers’ Unions Opposition Groups
Overall
Influence
Rev Elect Lobb
Overall
Influence
Elect Lobb
Teachers’
Unions
Overall
Influence
1.000
Rev 0.952 1.000
Elect 0.494 0.228 1.000
Lobb 0.299 0.160 0.202 1.000
Opposition
Groups
Overall
Influence
0.191 0.045 0.515 0.111 1.000
Elect 0.137 -0.003 0.515 0.056 0.927 1.000
Lobby 0.218 0.121 0.306 0.155 0.727 0.417 1.000
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 138

Table 10. State Rankings on Interest Group Influence Variables (All Years)

TU
Relative
Influence
Rank
TU
Relative
Election
Rank
TU
Relative
Lobbying
Rank
Winkler et
al. (2012)
TU
Strength
Rank
Teachers’ Unions Opposition Groups
Overall
Rev,
Dues,
Mem,
Cov
Election Lobby Overall Election Lobby
MN 1 7 4 14 8 8 27 9 45 46 33
AK 2 15 6 15 6 3 26 8 36 37 34
ND 3 5 10 24 25 26 36 11 50 50 39
HI 4 9 16 1 12 11 12 29 38 35 47
NJ 5 8 27 7 1 2 8 6 21 30 7
ME 6 34 35 22 23 10 49 45 47 48 36
RI 7 4 9 5 15 17 14 10 40 40 32
MT 8 28 13 3 19 9 45 18 41 42 42
MA 9 16 7 21 4 4 28 1 25 38 9
NE 10 1 3 26 13 25 2 4 31 25 38
WY 11 3 30 29 30 29 31 48 48 49 43
CT 12 30 19 17 22 14 47 22 43 44 28
NH 13 37 26 30 26 18 48 44 46 45 46
DE 14 17 18 19 18 15 25 37 35 33 37
VT 15 36 11 11 34 27 50 20 49 47 49
AL 16 2 12 20 14 20 3 5 22 20 30
KY 17 10 8 28 28 30 19 17 42 41 48
OR 18 19 34 2 3 5 4 27 11 7 18
MD 19 14 32 23 21 24 18 39 29 27 29
WA 20 39 2 10 5 7 23 2 12 12 10
MI 21 32 33 16 9 6 16 19 9 10 16
WI 22 25 5 18 20 23 22 12 24 22 27
WV 23 6 15 13 32 37 13 31 39 39 50
SD 24 13 40 34 38 35 33 50 44 43 45
OH 25 22 31 12 16 16 15 23 17 17 12
NY 26 23 50 9 2 1 11 40 6 11 2
KS 27 24 20 32 33 33 30 33 33 26 41
PA 28 35 23 4 10 12 5 28 7 3 31
NV 29 27 17 25 17 22 7 7 8 4 17
ID 30 33 38 36 36 32 40 41 32 34 25
CA 31 21 49 6 11 13 6 38 4 9 3
UT 32 29 1 39 40 40 37 3 30 32 23
OK 33 20 37 43 37 38 24 36 28 29 21
IA 34 49 42 27 29 21 46 49 18 14 24
VA 35 38 14 47 39 39 32 21 27 21 40
CO 36 12 43 35 27 31 10 32 15 19 8
NM 37 18 22 37 41 42 20 15 26 31 15
IL 38 11 46 8 7 19 1 24 2 2 4
AR 39 44 24 48 44 44 44 46 34 28 44
MS 40 26 29 46 48 47 38 42 37 36 35
MO 41 43 28 38 42 41 35 26 20 16 22
AZ 42 40 47 51 43 43 42 25 14 24 5
NC 43 47 39 40 46 45 43 43 19 15 26
TN 44 45 45 41 35 36 21 35 5 6 6
SC 45 42 36 49 50 50 41 34 23 23 20
TX 46 31 21 44 45 49 17 13 13 13 14
FL 47 46 48 50 31 34 29 14 3 5 1
IN 48 50 25 31 24 28 9 16 1 1 19
LA 49 41 41 42 49 48 34 30 16 18 11
GA 50 48 44 45 47 46 39 47 10 8 13
Notes: The Winkler et al. (2012) state teachers’ union rankings (column 2) and my state teachers’ union rankings (column 3) are correlated at
0.852. The Winkler et al. (2012) state teachers’ union rankings (column 2) are correlated with my relative strength rankings (column 1) at 0.697.
My state teachers’ union strength rankings (column 3) are correlated with my relative strength rankings (column 1) at 0.665.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 139

Table 11. OLS Regressions of Teachers’ Union Relative Influence on State-Level Agenda-Setting (2011-2015)

Prop. Unfav.
Laws
(1)
Prop. Fav.
Laws
(2)
Prop. Unfav.
Laws
(3)
Prop. Fav.
Laws
(4)
Prop. Unfav.
Laws
(5)
Prop. Fav.
Laws
(6)
Teachers’ Union Influence Relative
to Other Groups
Overall Relative Influence
-5.262*
(2.569)
6.236+
(3.121)

Election Relative Influence
-5.980+
(3.285)
6.508+
(3.715)

Lobbying Relative Influence
-0.511
(1.990)
0.141
(2.189)
Partisan Control
Republican Control
(ref= Democrat Control)
0.282
(5.878)
-2.172
(5.056)
-0.865
(5.943)
-0.867
(5.267)
2.181
(5.098)
-4.737
(4.755)
Split Control
-0.423
(5.721)
-1.455
(6.038)
-0.084
(5.730)
-1.839
(5.970)
-1.163
(5.435)
-0.892
(5.716)
Fiscal and Economic Health
Debt-to-Service Ratio
-0.252
(0.304)
0.161
(0.341)
-0.277
(0.302)
0.188
(0.340)
-0.176
(0.300)
0.062
(0.334)
Unemployment Rate
-0.200
(1.808)
0.212
(1.801)
0.012
(1.739)
-0.009
(1.756)
-0.091
(1.810)
0.125
(1.772)
Education System Performance/Size
NAEP Math Score
-1.356
(1.030)
1.750
(1.264)
-1.088
(1.007)
1.435
(1.227)
-1.101
(1.037)
1.405
(1.273)
Student Enrollment (ln)
-130.360
(93.508)
168.809+
(97.144)
-181.264*
(87.526)
223.749*
(93.197)
-120.221
(95.555)
151.317
(99.104)
Legislative Cycle Length
Session Calendar Days (ln)
0.953
(2.142)
-0.692
(1.962)
0.513
(2.104)
-0.208
(2.008)
1.258
(2.159)
-1.053
(1.996)
Policy Spillover
Neighboring States’ Agenda-
Setting
0.065
(0.121)
-0.059
(0.104)
0.058
(0.120)
-0.065
(0.104)
0.054
(0.122)
-0.037
(0.107)
Control for No Laws Proposed X X X X X X
Year FE X X X X X X
State FE X X X X X X
R-squared 62.12 58.16 62.53 58.62 61.65 57.55
N 248 248 248 248 250 250
Notes: + p<0.10 *p<0.05 **p<0.01 ***p<0.001; Standard errors are shown parentheses and are clustered at the
state-level.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 140

Table 12. OLS Regressions of Teachers’ Union Relative Influence on State-Level Policy-Setting (2011-2015)

Unfav. FR
(1)
Fav. SR
(2)
Unfav. FR
(3)
Fav. SR
(4)
Unfav. FR
(5)
Fav. SR
(6)
Teachers’ Union Influence
Relative to Other Groups

Overall Relative Influence
17.084**
(6.001)
2.111
(4.710)

Election Relative Influence
10.562**
(3.466)
7.415+
(4.422)

Lobbying Relative Influence
2.749
(2.073)
0.923
(2.434)
Partisan Control
Republican Control
(ref= Democrat Control)
-11.013
(9.385)
13.860
(10.307)
-8.460
(9.179)
15.018
(10.636)
-9.208
(8.385)
8.344
(9.983)
Split Control
-1.139
(7.589)
-5.335
(8.831)
-2.029
(6.700)
-5.678
(8.913)
-3.952
(7.193)
-3.352
(9.020)
Fiscal and Economic Health
Debt-to-Service Ratio
0.814*
(0.368)
-0.116
(0.291)
0.847*
(0.368)
-0.078
(0.302)
0.922*
(0.385)
-0.287
(0.323)
Unemployment Rate
-2.185
(1.620)
-3.335
(2.009)
-2.479
(1.592)
-3.655+
(2.122)
-1.682
(1.882)
-3.712+
(2.071)
Education System
Performance/Size

NAEP Math Score
-0.331
(1.204)
1.280
(1.206)
-1.144
(1.185)
1.142
(1.199)
-1.002
(1.490)
1.095
(1.230)
Student Enrollment (ln)
37.409
(96.502)
14.322
(101.218)
116.499
(112.219)
86.369
(87.145)
38.014
(100.884)
-3.390
(104.099)
Legislative Cycle Length
Session Calendar Days (ln)
1.305
(2.154)
-0.318
(1.837)
1.963
(2.049)
0.405
(1.779)
0.965
(2.206)
-0.755
(1.875)
Policy Spillover
Neighboring States’ Policy-
Setting
-0.264*
(0.127)
-0.064
(0.109)
-0.278*
(0.126)
-0.085
(0.107)
-0.284*
(0.129)
-0.051
(0.110)
Control for No Laws Proposed X X X X X X
Year FE X X X X X X
State FE X X X X X X
R-squared 74.72 57.75 74.61 58.50 73.43 56.71
N 248 248 248 248 250 250
Notes: + p<0.10 *p<0.05 **p<0.01 ***p<0.001; Standard errors are shown parentheses and are clustered at
the state-level.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 141

Table 13. OLS Regressions of Teachers’ Unions Influence on State-Level Agenda- and Policy-
Setting (2011-2015)

Prop. Unfav.
Laws
(1)
Prop. Fav.
Laws
(2)
Unfav. FR
(4)
Fav. SR
(5)
Teachers’ Union Overall Influence
Overall Influence
9.457
(7.526)
-12.933
(9.310)
18.723**
(6.177)
2.939
(13.916)
Election Influence
0.677
(8.527)
-1.207
(9.098)
9.011*
(4.095)
16.789*
(7.733)
Lobbying Influence
-0.423
(1.945)
-0.114
(2.136)
1.372
(1.593)
0.224
(2.063)
Overall Influence Highest Loading Variables
Teachers’ Union Membership Dues
Per Teacher
31.986**
(10.939)
-33.044**
(11.855)
17.188
(17.119)
-8.342
(20.662)
Teachers’ Unions Contributions Per
Candidate (ln)
-0.953
(4.778)
1.085
(5.154)
5.727*
(2.223)
10.799*
(4.100)
# Teachers’ Union Lobbyist Per
Lawmaker
-12.291
(21.447)
-6.136
(23.935)
33.134*
(13.111)
-40.947
(28.607)
Relative Influence with Highest Loading Variables
Teachers’ Unions Relative
Contributions Per Candidate (ln)
-1.034
(0.673)
1.619+
(0.815)
1.896
(1.214)
1.533
(1.067)
# Teachers’ Union Relative
Lobbyist Per Lawmaker
1.232
(1.237)
-1.229
(1.558)
-1.467
(0.949)
0.930
(1.043)
Notes: + p<0.10 *p<0.05 **p<0.01 ***p<0.001; Standard errors are shown parentheses and are
clustered at the state-level.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 142

Table 14. Comparison of Studies on Transfer and Vacancy Provisions and the Distribution of Teachers
Study
Moe (2006c)
A.
Koski & Horng (2007)
B.
Cohen-Vogel, Feng, Osborne-
Lampkin (2013)
C.
Anzia & Moe (2014a)
D.
Goldhaber, Lavery, &
Theobald (2016)
E.
Current Study
F.
Level of Analysis
Schools nested within
districts
Schools nested within
districts
Schools nested within
districts
Schools nested within
districts
Teachers, nested within
schools, nested within
districts
Schools nested within
districts
Cross-
sectional/Longitudinal
Cross-sectional Cross-sectional Cross-sectional Cross-sectional
Longitudinal teacher data;
Cross-sectional CBA data for
most districts
Longitudinal school, district,
and CBA data
Sample Size
1,588 schools, 115 school
districts in California
5,199 schools, 420 school
districts in California
2,297 schools, 66 school
districts in Florida
Same as Koski & Horng
(2007); Moe (2006c)
190,469 teachers/year, 7,159
school/year, 994 district/year
observations
Similar to Koski & Horng
(2007); Anzia & Moe,
(2014a)
Sample Restrictions
Districts with four or more
elementary schools, and
schools with more than 15
percent and less than 85
percent minority students,
excludes LAUSD
School districts with four or
more schools, includes
LAUSD
Unionized districts with four
or more schools
Same as Koski & Horng
(2007); Moe (2006c)
None
School districts with four or
more schools, includes
LAUSD
Year(s) of CBA Data 1998-1999 2005-2006 2002-2003
Same as Koski & Horng
(2007); Moe (2006c)
2010-2011
2005-2006, 2008-2009, 2011-
2012, 2014-2015
Measure(s) of
Transfer and Vacancy
Provisions
Voluntary transfers and
involuntary
transfers are given a score of
1 if seniority is overriding
factor, and 0 otherwise. The
two scores are then added to
yield a measure that takes on
the values 0, 1, or 2
Transfer and Leave Score
(TLS) based on 6 provisions
from CBAs. Also use 1 item
measure of seniority in
voluntary transfers
Same as Koski & Horng
(2007)
Same as Koski & Horng
(2007); Moe (2006c)
Separate four-point scale for
voluntary and involuntary
transfers from prohibited
from use to only factor in
decisions.
Same as Koski & Horng
(2007); Moe (2006c); Cohen-
Vogel, Feng, Osborne-
Lampkin (2013); Anzia &
Moe, (2014a); Goldhaber,
Lavery, & Theobald (2016);
additional measure of TV
restrictiveness
Outcome Measure(s)
Proportion of teachers in the
school who are inexperienced
(teaching for less than three
years) or not fully
credentialed
The percent of teachers in the
district who are fully
credentialed and the percent
of teachers in the district with
more than two years of
teaching experience
The percentage of
professionally certified
teachers in a school; the
percentage of teachers in a
school with at least 3 years of
experience; school average
teaching experience; National
Board Certification; SAT
scores.
Same as Koski & Horng
(2007); Moe (2006c)
The probability that teacher i
in school j, district k, and year
t transfers to another school in
the district the following year
(relative to staying in the
same school)
Same as Koski & Horng
(2007); Moe (2006b); Cohen-
Vogel, Feng, Osborne-
Lampkin (2013); Anzia &
Moe, (2014a); Student
Achievement
Indicator of
Disadvantage
School’s percentage of
minority students (African-
Americans and Latinos)
Same as Moe (2006c)
School’s percentage of black
students; School’s percentage
of Hispanic students
Same as Koski & Horng
(2007); Moe (2006c)
percentage of
“disadvantaged” students in
school
Same as Koski & Horng
(2007); Moe (2006c)
Identification Strategy OLS w/ fixed effect HLM HLM/ OLS w/ fixed effect HLM/ OLS w/ fixed effect
Multinomial logit with
district-by-year fixed effects
OLS w/ district, year, and
district-by-year fixed effects
Findings
Strong transfer provisions are
associated with lower teacher
quality in disadvantaged
schools
Null results Null results
Strong transfer provisions are
associated with lower teacher
quality in disadvantaged
schools, especially in large
districts
Veteran teachers’ more likely
to leave in districts with
strong CBA involuntary
transfer protections; more
effective teachers (higher
VAM) are less likely to exit.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 143

Table 15. Summary Statistics for All Transfer and Vacancy Measures and Other District-Level Controls

2005-06
(1)
2008-09
(2)
2011-12
(3)
2014-2015
(4)
Year-to-Year
Change
(5)
Mean SD Mean SD Mean SD Mean SD Mean SD
A. District-Level Variables n=464 n=493 n=489 n=492 n=1,266
Transfer and Vacancy
Measures

Goldhaber et al.
Voluntary Transfers
No Seniority 0.293 0.456 0.215 0.411 0.200 0.401 0.215 0.412 -0.025 0.283
Seniority Considered 0.239 0.427 0.288 0.453 0.303 0.460 0.305 0.461 0.017 0.286
Seniority All Else Equal 0.409 0.492 0.469 0.500 0.466 0.499 0.449 0.498 0.015 0.306
Seniority Determining
Factor
0.058 0.234 0.028 0.166 0.031 0.173 0.030 0.172 -0.007 0.129
Involuntary Transfers
No Seniority 0.265 0.442 0.205 0.404 0.186 0.390 0.171 0.377 -0.029 0.284
Seniority Considered 0.188 0.391 0.217 0.413 0.225 0.418 0.222 0.416 0.007 0.259
Seniority All Else Equal 0.401 0.491 0.414 0.493 0.382 0.486 0.388 0.488 -0.003 0.328
Seniority Determining
Factor
0.147 0.354 0.164 0.371 0.207 0.405 0.220 0.414 0.025 0.237
Anzia and Moe 1.015 0.825 1.075 0.829 1.086 0.815 1.087 0.811 0.030 0.473
Koski & Horng 2.800 1.046 3.142 0.982 3.166 0.964 3.260 0.979 0.147 0.807
Transfer and Vacancy Restrict. 0.006 0.259 -0.036 0.281 0.031 0.269 0.014 0.258 0.003 0.284
Overall Restrict. -0.289 0.438 -0.056 0.447 0.106 0.453 0.186 0.445 0.154 0.292
Other District-Level Controls
Ln (Enrollment) 8.627 1.129 8.635 1.120 8.710 1.086 8.747 1.044 0.000 0.111
% FRL 42.758 25.269 47.612 25.169 51.474 24.829 54.469 25.690 3.641 9.782
B. School-Level Variables n=5804 n=6365 n=6496 n=6381 n=18289
Outcome Variables
% Experienced 88.545 9.985 90.673 9.498 94.224 9.427 89.286 10.338 0.409 12.072
% Masters 17.967 15.271 18.773 17.069 22.260 21.254 23.772 20.983 1.901 12.384
ELA Achievement 0.210 0.836 0.206 0.801 0.193 0.802 0.066 0.938 -0.046 0.340
Math Achievement 0.233 0.873 0.228 0.849 0.228 0.841 0.078 0.957 -0.050 0.540
School-Level Controls
Ln (Enrollment) 6.431 0.736 6.416 0.710 6.440 0.693 6.426 0.652 -0.021 0.179
Teacher to Pupil Ratio 20.905 3.354 18.999 2.972 24.037 3.839 24.606 18.833 1.080 13.292
% Minority 59.306 28.857 62.694 27.977 62.491 28.608 63.287 28.613 4.086 27.718
% FRL 52.088 30.266 55.626 29.751 60.137 28.893 61.012 29.137 0.868 5.022
Urban school 0.430 0.495 0.431 0.495 0.430 0.495 0.405 0.491
Suburban/town school (ref) 0.483 0.500 0.462 0.499 0.466 0.499 0.529 0.499
Rural school 0.087 0.282 0.107 0.309 0.103 0.304 0.066 0.249
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 144

Table 16. Regression of School-Aggregated Teacher Experience and Teacher Education on CBA Transfer and Vacancy
Provisions
% Experienced % Master’s Degree
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District
Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
A. Voluntary Transfers
% Minority
-0.079**
(0.029)
-0.071***
(0.012)
-0.068***
(0.016)
-0.032*
(0.014)
0.004
(0.022)
-0.017
(0.013)
Considered (ref- No Seniority)

0.118
(1.050)
2.482
(2.642)
All Else Equal

-0.422
(0.967)
1.632
(2.462)
Determining Factor

0.126
(2.349)
0.609
(3.001)
% Minority x Considered
-0.026
(0.034)
0.012
(0.015)
-0.014
(0.019)
-0.079
(0.060)
-0.057+
(0.033)
-0.038
(0.026)
% Minority x All Else Equal
0.028
(0.038)
0.007
(0.012)
0.003
(0.015)
0.004
(0.022)
-0.039
(0.032)
0.001
(0.016)
% Minority x Determining Factor
0.031
(0.039)
-0.001
(0.029)
0.024
(0.026)
0.079*
(0.034)
0.069
(0.072)
0.018
(0.045)
B. Involuntary Transfers
% Minority
-0.086*
(0.038)
-0.068***
(0.017)
-0.080***
(0.019)
0.004
(0.032)
-0.014
(0.034)
-0.019
(0.023)
Considered (ref- No Seniority)

0.288
(1.071)
-0.518
(2.624)
All Else Equal

0.147
(1.186)
0.914
(2.947)
Determining Factor

1.376
(1.212)
1.629
(3.051)
% Minority x Considered
0.021
(0.041)
-0.007
(0.019)
0.002
(0.024)
-0.024
(0.036)
-0.021
(0.039)
-0.003
(0.026)
% Minority x All Else Equal
0.019
(0.044)
0.007
(0.019)
0.016
(0.026)
-0.038
(0.035)
-0.010
(0.038)
-0.002
(0.024)
% Minority x Determining Factor
-0.000
(0.048)
0.007
(0.019)
0.014
(0.024)
-0.115+
(0.067)
-0.017
(0.043)
-0.016
(0.035)
C. Combined Provision Measures
% Minority
-0.096***
(0.025)
-0.069***
(0.010)
-0.082***
(0.012)
-0.044+
(0.024)
-0.018
(0.027)
-0.035*
(0.014)
Anzia & Moe

-0.130
(0.567)
0.746
(1.393)
% Minority x Anzia & Moe
0.020
(0.019)
0.004
(0.009)
0.010
(0.010)
-0.000
(0.015)
-0.007
(0.020)
0.008
(0.008)
% Minority
-0.047
(0.045)
-0.046*
(0.018)
-0.056*
(0.023)
0.034
(0.062)
-0.012
(0.041)
-0.008
(0.031)
Koski & Horng

0.196
(0.324)
0.045
(0.752)
% Minority x Koski & Horng
-0.009
(0.013)
-0.005
(0.005)
-0.004
(0.006)
-0.026
(0.024)
-0.005
(0.013)
-0.005
(0.011)
% Minority
-0.072***
(0.014)
-0.062***
(0.009)
-0.068***
(0.010)
-0.044**
(0.017)
-0.025*
(0.010)
-0.026**
(0.009)
TV Restrict.

1.902
(1.196)
1.107
(2.181)
% Minority x TV Restrict.
-0.068
(0.047)
-0.024
(0.022)
-0.012
(0.024)
-0.037
(0.057)
-0.022
(0.044)
0.011
(0.027)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All models
control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural, suburban/town)
and percent minority. The district and year fixed effect models also include district-level controls for district enrollment
(natural log) and percent free and reduced price lunch.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 145

Table 17. Regression of School-Aggregated Math and ELA Student Achievement on CBA Transfer and Vacancy Provisions
ELA Score (Standardized) Math Score (Standardized)
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
A. Voluntary Transfers
% Minority
-0.028***
(0.001)
-0.027***
(0.001)
-0.027***
(0.001)
-0.024***
(0.002)
-0.022***
(0.001)
-0.023***
(0.001)
Considered (ref- No Seniority)

-0.011
(0.063)
0.045
(0.071)
All Else Equal

-0.023
(0.059)
0.054
(0.070)
Determining Factor

0.069
(0.113)
0.022
(0.153)
% Minority x Considered
0.000
(0.002)
-0.000
(0.001)
-0.000
(0.002)
0.001
(0.002)
-0.001
(0.001)
-0.000
(0.002)
% Minority x All Else Equal
0.002
(0.002)
-0.000
(0.001)
0.000
(0.001)
0.001
(0.002)
-0.001
(0.001)
-0.001
(0.001)
% Minority x Determining Factor
-0.004
(0.005)
-0.001
(0.002)
-0.000
(0.003)
-0.005
(0.007)
-0.001
(0.002)
0.001
(0.004)
B. Involuntary Transfers
% Minority
-0.027***
(0.001)
-0.027***
(0.001)
-0.027***
(0.001)
-0.021***
(0.002)
-0.022***
(0.001)
-0.022***
(0.002)
Considered (ref- No Seniority)

-0.090
(0.057)
0.064
(0.081)
All Else Equal

-0.002
(0.044)
0.092
(0.061)
Determining Factor

0.019
(0.058)
0.117+
(0.071)
% Minority x Considered
0.004+
(0.003)
0.001
(0.001)
0.002
(0.002)
-0.001
(0.003)
-0.001
(0.001)
-0.001
(0.003)
% Minority x All Else Equal
-0.000
(0.001)
0.000
(0.001)
-0.000
(0.001)
-0.003
(0.002)
-0.001
(0.001)
-0.002
(0.002)
% Minority x Determining Factor
-0.002
(0.002)
-0.001
(0.001)
-0.002
(0.001)
-0.004
(0.003)
-0.003*
(0.001)
-0.003+
(0.002)
C. Combined Provision Measures
% Minority
-0.026***
(0.002)
-0.026***
(0.001)
-0.026***
(0.001)
-0.023***
(0.002)
-0.022***
(0.001)
-0.022***
(0.001)
Anzia & Moe

0.018
(0.028)
0.040
(0.036)
% Minority x Anzia & Moe
-0.001
(0.001)
-0.000
(0.000)
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
% Minority
-0.025***
(0.002)
-0.026***
(0.001)
-0.024***
(0.002)
-0.019***
(0.003)
-0.020***
(0.001)
-0.019***
(0.002)
Koski & Horng

0.017
(0.016)
0.059**
(0.019)
% Minority x Koski & Horng
-0.001
(0.001)
-0.000
(0.000)
-0.001+
(0.000)
-0.001
(0.001)
-0.001*
(0.000)
-0.001*
(0.001)
% Minority
-0.027***
(0.001)
-0.027***
(0.001)
-0.027***
(0.001)
-0.023***
(0.001)
-0.023***
(0.001)
-0.023***
(0.001)
TV Restrict.

0.067
(0.067)
0.141
(0.089)
% Minority x TV Restrict.
-0.005
(0.003)
-0.001
(0.001)
-0.003+
(0.002)
-0.009*
(0.004)
-0.002+
(0.001)
-0.005*
(0.002)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All
models control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural,
suburban/town) and percent minority. The district and year fixed effect models also include district-level controls for district
enrollment (natural log) and percent free and reduced price lunch.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 146

Table 18. Regression of School-Aggregated Teacher Experience, Teacher Education, ELA, and Math Achievement on CBA
Overall Restrictiveness Measure
A. % Experienced % Master’s Degree
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
% Minority
-0.074***
(0.016)
-0.064***
(0.010)
-0.065***
(0.010)
-0.047*
(0.018)
-0.032**
(0.012)
-0.026*
(0.011)
Overall Restrict.

0.967
(1.249)
-0.963
(3.047)
% Minority x Overall Restrict.
-0.013
(0.040)
-0.005
(0.016)
-0.020
(0.018)
-0.029
(0.030)
0.027
(0.035)
0.009
(0.015)
B. ELA Score (Standardized) Math Score (Standardized)
% Minority
-0.028***
(0.001)
-0.026***
(0.001)
-0.026***
(0.001)
-0.024***
(0.001)
-0.022***
(0.001)
-0.022***
(0.001)
Overall Restrict.

0.157**
(0.052)
0.329***
(0.075)
% Minority x Overall Restrict.
-0.007***
(0.002)
-0.002**
(0.001)
-0.005***
(0.001)
-0.008**
(0.002)
-0.005***
(0.001)
-0.006***
(0.001)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All models
control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural, suburban/town)
and percent minority. The district and year fixed effect models also include district-level controls for district enrollment (natural
log) and percent free and reduced price lunch.
Table 19. Regression of School-Aggregated Teacher Experience, Teacher Education, ELA, and Math Achievement on TV
Restrictiveness Controlling for CBA Overall Restrictiveness
A. % Experienced % Master’s Degree
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
% Minority
-0.072***
(0.015)
-0.062***
(0.009)
-0.065***
(0.010)
-0.045**
(0.017)
-0.030*
(0.012)
-0.027**
(0.010)
TV Restrict.

1.799
(1.223)
1.814
(2.094)
Overall Restrict.

0.689
(1.329)
-1.346
(2.956)
% Minority x TV Restrict
-0.071+
(0.043)
-0.023
(0.025)
-0.001
(0.023)
-0.026
(0.059)
-0.035
(0.042)
0.007
(0.031)
% Minority x Overall Restrict.
0.007
(0.037)
-0.001
(0.017)
-0.020
(0.019)
-0.022
(0.031)
0.034
(0.035)
0.008
(0.018)
B. ELA Score (Standardized) Math Score (Standardized)
% Minority
-0.028***
(0.001)
-0.026***
(0.001)
-0.026***
(0.001)
-0.024***
(0.001)
-0.022***
(0.001)
-0.022***
(0.001)
TV Restrict.

0.017
(0.066)
0.024
(0.089)
Overall Restrict.

0.155**
(0.053)
0.325***
(0.077)
% Minority x TV Restrict
-0.002
(0.003)
-0.000
(0.001)
-0.001
(0.002)
-0.005
(0.004)
-0.000
(0.001)
-0.002
(0.002)
% Minority x Overall Restrict.
-0.006**
(0.002)
-0.002*
(0.001)
-0.005***
(0.001)
-0.006**
(0.002)
-0.005***
(0.001)
-0.006***
(0.001)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All models
control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural, suburban/town)
and percent minority. The district and year fixed effect models also include district-level controls for district enrollment (natural
log) and percent free and reduced price lunch.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 147

Appendix

Appendix Table 1. Coded Legislation Topics
Topic
Describes the substantive topic of the law (Categories are not mutually
exclusive).
Collective Bargaining Negotiations
Laws pertaining to the legality of the collective bargaining and the
scope, and applicability of collective bargaining agreements for
teachers.
Union Certification/ Management
Laws focused on the rights of teachers to join teachers’ unions and the
ability of local teachers to organize new employee organizations.
Membership/Right-to-Work
Laws relating requirements that employees join unions as a condition
of employment.
Strikes
Laws relating to public sector employee strikes or other concerted
activities.
Political Contributions
Laws relating to financial contributions from public employee unions
for political purposes.
Union Employee Rights
Laws pertaining to specific rights of teachers’ union members,
including binding arbitration for the resolution of disputes.
Membership Dues Laws relating to the payment and use of union dues.
Teacher/District CBA Law
Laws focused explicitly on state legal protections for provisions found
in teacher CBAs.
Class Size
Laws relating to student/teacher ratios, class size reduction, class size
reduction funding.
Evaluation
Laws relating to the implementation or revision of teacher evaluation
systems.
Grievances Laws pertaining to the teacher grievance process.
Layoff/Discipline/Dismissal
Laws relating to layoff, discipline, and dismissal procedures for
teachers.
Leave
Laws relating to the amount of job leave given to teachers (i.e.
bereavement, sabbatical, personal, etc.).
Non-Teaching Duties
Laws pertaining to adjunct and other non-teaching duties of full-time
teachers, including preparation and collaboration time.
Retirement
Laws relating to teacher retirement systems, retirement benefit
amounts, and early retirement incentives.
Salary and Benefits
Laws relating to teacher salaries and health benefits, including the
implementation of new pay pans.
Performance Pay
Laws relating to performance incentives and merit pay plans that link
teacher performance with teacher compensation. (A subset of salary
and benefit laws).
School Days and Hours
Laws relating to the length of the school day, school year, and the
number of instructional minutes in the school day/year.
Tenure
Laws governing teacher tenure including the revocation of
tenure/lengthening or requiring additional probationary periods.
Transfer and Vacancies
Laws relating to hiring, teacher assignments, and the filling of
vacancies including voluntary and involuntary teacher transfers.
Working Conditions Laws pertaining to teacher safety and classroom conditions.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 148

Appendix Table 2. Bivariate Correlations between Nine Union Influence Variables (2010 and 2011)

2010 2011
Membership/Resources Election Influence Lobby Strength Membership/Resources Election Influence Lobby Strength
Dues Spend Mem Cov Win Per Prop Den Per Dues Spend Mem Cov Win Per Prop Den Per
Dues Per Teacher (ln) 1.00
Spending Per Student (ln) 0.95 1.00
Member Rate 0.84 0.87 1.00
CBA Coverage 0.74 0.77 0.80 1.00
Win Rate 0.33 0.22 0.27 0.28 1.00
Contrib Per Candidate
(ln)
0.22 0.13 0.25 0.20 0.48 1.00
Prop. Contrib
Teachers’ Union
0.28 0.25 0.19 0.03 0.39 0.64 1.00
Lobby Density 0.15 0.17 0.26 0.27 0.08 -0.13 0.00 1.00
Lobbyist Per Lawmaker 0.23 0.13 0.32 0.29 0.46 0.17 0.09 0.45 1.00
Dues Per Teacher (ln) 0.99 0.95 0.84 0.75 0.35 0.22 0.27 0.15 0.22 1.00
Spending Per Student (ln) 0.94 0.98 0.85 0.79 0.21 0.18 0.24 0.15 0.11 0.95 1.00
Member Rate 0.84 0.87 0.99 0.82 0.26 0.24 0.15 0.27 0.31 0.84 0.86 1.00
CBA Coverage 0.73 0.75 0.78 0.98 0.31 0.19 0.02 0.24 0.29 0.74 0.77 0.80 1.00
Win Rate 0.34 0.24 0.28 0.29 0.97 0.48 0.40 0.03 0.46 0.36 0.24 0.27 0.31 1.00
Contrib Per Candidate
(ln)
0.19 0.11 0.24 0.18 0.48 0.99 0.64 -0.14 0.17 0.19 0.15 0.23 0.18 0.45 1.00
Prop. Contrib
Teachers’ Union
0.31 0.27 0.21 0.06 0.37 0.64 0.99 -0.01 0.10 0.29 0.26 0.18 0.04 0.41 0.62 1.00
Lobby Density 0.07 0.02 0.10 0.09 0.06 -0.10 0.13 0.76 0.28 0.06 0.02 0.09 0.06 0.03 -0.11 0.12 1.00
Lobbyist Per Lawmaker 0.17 0.08 0.24 0.28 0.38 0.09 -0.02 0.33 0.84 0.17 0.07 0.24 0.27 0.38 0.10 -0.01 0.32 1.00
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 149

Appendix Table 3. Eigenvalues and Component Loading on Components from Principal Component Analysis on Opposition Group
Influence Variables (Year=2011)
Panel A. Overall Opposition Group Strength Influence

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 3.368 1.720 0.337 0.337
Factor 2 1.648 0.253 0.165 0.502
Factor 3 1.395 0.210 0.140 0.641
Factor 4 1.186 0.387 0.119 0.760
Factor 1 Factor 2 Factor 3 Factor 4 Uniqueness
Proportion of Open Seats Won by Business Group Allies 0.516 0.432 0.309 -0.259 0.385
Business Group Contributions Per Candidate (ln) 0.781 0.486 0.178 -0.059 0.119
Proportion of Contributions from Business Groups 0.369 0.653 0.417 0.068 0.259
Proportion of Open Seats Won by School Choice Group Allies 0.674 -0.491 0.233 0.228 0.199
School Choice Group Contributions Per Candidate (ln) 0.824 -0.398 0.047 0.046 0.159
Proportion of Contributions from School Choice Groups 0.494 -0.512 0.461 0.170 0.253
Business Group Lobby Organization Density -0.116 0.127 0.002 0.783 0.357
# Business Group Lobbyist Per Lawmaker 0.608 0.250 -0.579 0.311 0.136
School Choice Group Lobby Organization Density 0.411 -0.241 -0.213 -0.563 0.411
# School Choice Group Lobbyist Per Lawmaker 0.654 0.046 -0.666 0.019 0.126
Panel B. Opposition Group Election Influence Measure (Year=2011)

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 2.758 1.156 0.460 0.460
Factor 2 1.602 0.899 0.267 0.727
Factor 1 Factor 2 Uniqueness
Proportion of Open Seats Won by Business Group Allies 0.571 0.506 0.418
Business Group Contributions Per Candidate (ln) 0.780 0.511 0.131
Proportion of Contributions from Business Groups 0.462 0.669 0.339
Proportion of Open Seats Won by School Choice Group Allies 0.744 -0.500 0.197
School Choice Group Contributions Per Candidate (ln) 0.812 -0.396 0.185
Proportion of Contributions from School Choice Groups 0.630 -0.481 0.371
Panel C. Opposition Group Lobbying Influence Measure (Year=2011)

Eigenvalue Difference
Proportion
Variance
Explained
Cumulative
Variance
Explained
Factor 1 1.841 0.664 0.460 0.460
Factor 2 1.177 0.389 0.294 0.755
Factor 1 Factor 2 Uniqueness
Business Group Lobby Organization Density -0.105 0.789 0.366
# Business Group Lobbyist Per Lawmaker 0.841 0.390 0.141
School Choice Group Lobby Organization Density 0.474 -0.632 0.376
# School Choice Group Lobbyist Per Lawmaker 0.948 0.058 0.098

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 150

Appendix 4- Items Included in Transfer and Vacancy Measures
Anzia and Moe measure

2. What role does seniority play in voluntary transfer teacher assignments?
a. Seniority is not determinative = 0
b. Seniority is determinative = 1
3. What role does seniority play in involuntary transfer teacher assignments?
a. Seniority is not determinative = 0
b. Seniority is determinative = 1

Items included in Goldhaber et al. measure

2. Voluntary Transfers
a. Seniority not addressed
b. Seniority one of several factors
c. Seniority is tiebreaker
d. Seniority is only factor
3. Involuntary Transfers
a. Seniority not addressed
b. Seniority one of several factors
c. Seniority is tiebreaker
d. Seniority is only factor

Items included in the Koski & Horng measure
1. What role does seniority play in voluntary transfer teacher assignments?
a. No seniority language = 0
b. Seniority a factor, but not determinative = 1
c. Seniority determinative = 2
d. Displacement of other teachers based on seniority permitted (bumping) = 3
2. What role does seniority play in selecting a teacher to involuntarily transfer?
a. No seniority language = 0
b. Seniority a factor, but not determinative = 1
c. Seniority determinative = 2
3. What role does seniority play in receiving a teacher who is being involuntarily
transferred?
a. No seniority language = 0
b. Seniority a factor, but not determinative = 1
c. Seniority determinative = 2
d. Displacement of other teachers based on seniority permitted (bumping) = 3
4. How are outside applicants considered relative to inside applicants?
a. No preference for inside applicants = 0
b. Inside applicant is factored into decision, but not determinative = 1
c. Inside applicant is determinative = 2
5. When is the district required to provide reasons for denying a transfer request?
a. Not required at all = 0
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 151

b. Required on request = 1
c. Required in every instance = 2
6. What position must a teacher be given on returning from long-term paid leave?
a. a. Not guaranteed prior assignment = 1
b. b. Guaranteed prior assignment = 2

Appendix Table 4. Items Included in New Measure of Transfer and Vacancy Restrictiveness
Transfer and Vacancies
CBA priorities which members get transfers
CBA does not hold that voluntary transfers get preference ahead of involuntary transfers
CBA holds that those being involuntary transferred get preference ahead of voluntary transfers
CBA holds that members who have been displaced from site/are at risk of layoff get preference
CBA holds that teachers returning from leaves get preference
CBA does not state that the district must consider clear and compelling program needs/ student welfare in determining
transfers
CBA does not state that the district must consider credentials in relations to position requires
CBA does not state that the district must consider teacher qualifications
CBA states that the district must consider teacher preference
CBA states that the district must consider seniority in the building
CBA does not state that district can consider evaluations/quality at all in transfer decisions
CBA does not state that members major/minor fields of study can play a role in the decision
CBA does not state that members special job-related skills/talents can play a role in the decision
CBA addresses seniority as a factor in deciding who is voluntarily transferred
Seniority more than just considered in voluntary transfer decisions
Seniority deciding factor in voluntary transfer decisions
CBA does not hold that members can be involuntarily transferred after the start of the school year
CBA says that members who are involuntarily transferred for non-cause reasons retain preferential rights to return to
that site/position if spot reopens
CBA says that members who are involuntarily transferred for non-cause reasons retain preferential rights to return to
that site/position if spot reopens for more than 1 year?
CBA addresses seniority as a factor in deciding who is involuntarily transferred
Seniority is more than just considered in involuntary transfer decisions
Seniority is the deciding factor in involuntary transfer decisions
CBA places a limit on the frequency with which members may be involuntarily transferred
CBA restricts involuntary transfers for a teacher to no more than 1 time in at least 3 years
CBA outlines that a teacher may be transferred for the best operational needs of the district
CBA places restrictions on causes for which a member may be involuntarily transferred
CBA specifies treatment of involuntarily-transferred members’ preferences for vacancies
CBA specifies that involuntary transferee can at least indicate preferences
If involuntarily transferred, district honors request among list of vacancies if meets requirements
CBA specifies that seniority rights allow most senior involuntary transferred teacher first choice of vacant spots
CBA specifies that if mems are involuntary transferred due to school closing, teachers from that site receive priority in
filling vacancies
CBA specifies the order in which district can consider new employees for vacancies
Current employed teachers can at least apply/and or will be considered for a vacant position before new personnel are
considered for the assignment
If position opened within the school year is filled with a probationary/temporary teacher, will be re-opened the
following year to members seeking transfer/reassignment
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 152

Appendix Table 5. Correlation between Continuous Transfer and Vacancy Measures in Base Year
(2005-06)

Anzia & Moe Koski & Horng TV Restrict.
Overall
Restrict.
Anzia & Moe 1.000
Koski & Horng 0.551 1.000
TV Restrict. 0.412 0.578 1.000
Overall Restrict. 0.334 0.373 0.311 1.000

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 153

Appendix Table 6. Regression of School-Aggregated Teacher Experience and Teacher Education on CBA Transfer and
Vacancy Provisions, Limited Sample
% Experienced % Master’s Degree
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District
Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
A. Voluntary Transfers
% Minority
-0.063*
(0.029)
-0.067***
(0.014)
-0.061***
(0.016)
-0.028+
(0.016)
-0.018
(0.020)
-0.013
(0.015)
Considered (ref- No Seniority)

0.517
(1.225)
2.706
(2.861)
All Else Equal

-0.110
(1.283)
2.633
(2.737)
Determining Factor

3.098
(2.409)
1.627
(3.500)
% Minority x Considered
-0.030
(0.032)
0.003
(0.018)
-0.021
(0.019)
-0.089
(0.061)
-0.030
(0.033)
-0.049+
(0.027)
% Minority x All Else Equal
0.020
(0.038)
0.004
(0.018)
-0.005
(0.020)
0.003
(0.025)
-0.007
(0.029)
-0.006
(0.019)
% Minority x Determining Factor
0.023
(0.043)
-0.056
(0.044)
0.003
(0.034)
0.096**
(0.036)
0.036
(0.062)
0.080+
(0.042)
B. Involuntary Transfers
% Minority
-0.050
(0.034)
-0.066***
(0.018)
-0.063**
(0.019)
0.003
(0.039)
-0.062
(0.042)
-0.014
(0.026)
Considered (ref- No Seniority)

0.681
(1.365)
-4.917+
(2.925)
All Else Equal

0.157
(1.199)
-1.095
(3.574)
Determining Factor

1.633
(1.370)
1.508
(3.149)
% Minority x Considered
-0.004
(0.039)
-0.013
(0.023)
-0.020
(0.024)
-0.019
(0.043)
0.055
(0.044)
-0.007
(0.029)
% Minority x All Else Equal
-0.014
(0.041)
0.003
(0.020)
-0.004
(0.023)
-0.029
(0.043)
0.019
(0.050)
-0.011
(0.029)
% Minority x Determining Factor
-0.020
(0.046)
0.006
(0.022)
0.001
(0.025)
-0.125+
(0.071)
0.042
(0.050)
-0.024
(0.038)
C. Combined Provision Measures
% Minority
-0.075**
(0.023)
-0.070***
(0.012)
-0.079***
(0.013)
-0.050
(0.032)
-0.030
(0.023)
-0.035*
(0.016)
Anzia & Moe

-0.096
(0.643)
1.383
(1.590)
% Minority x Anzia & Moe
0.011
(0.020)
0.004
(0.011)
0.008
(0.011)
0.003
(0.016)
0.001
(0.018)
0.006
(0.010)
% Minority
-0.006
(0.040)
-0.044+
(0.026)
-0.045+
(0.025)
0.065
(0.078)
-0.081+
(0.048)
0.002
(0.038)
Koski & Horng

0.290
(0.412)
-0.348
(0.854)
% Minority x Koski & Horng
-0.018
(0.012)
-0.006
(0.007)
-0.007
(0.007)
-0.035
(0.029)
0.015
(0.016)
-0.009
(0.013)
% Minority
-0.062***
(0.015)
-0.064***
(0.008)
-0.068***
(0.009)
-0.043*
(0.020)
-0.032**
(0.011)
-0.028*
(0.012)
TV Restrict.

1.304
(1.440)
1.233
(2.257)
% Minority x TV Restrict.
-0.016
(0.055)
-0.014
(0.025)
-0.011
(0.026)
-0.075
(0.068)
0.035
(0.038)
0.007
(0.035)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All models
control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural, suburban/town)
and percent minority. The district and year fixed effect models also include district-level controls for district enrollment
(natural log) and percent free and reduced price lunch. The sample is limited to districts with at least four elementary schools
and districts where the median school has between 15 percent and 85 percent minority students.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 154

Appendix Table 7. Regression of School-Aggregated Math and ELA Student Achievement on CBA Transfer and Vacancy
Provisions, Limited Sample
ELA Score (Standardized) Math Score (Standardized)
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
A. Voluntary Transfers
% Minority
-0.027***
(0.002)
-0.027***
(0.001)
-0.026***
(0.002)
-0.023***
(0.002)
-0.023***
(0.001)
-0.023***
(0.002)
Considered (ref- No Seniority)

-0.003
(0.086)
0.027
(0.088)
All Else Equal

-0.006
(0.081)
0.076
(0.094)
Determining Factor

0.092
(0.164)
0.108
(0.206)
% Minority x Considered
-0.000
(0.003)
0.000
(0.002)
-0.001
(0.002)
0.000
(0.003)
-0.000
(0.002)
-0.000
(0.002)
% Minority x All Else Equal
0.002
(0.002)
0.000
(0.001)
0.000
(0.002)
0.000
(0.002)
-0.001
(0.002)
-0.001
(0.002)
% Minority x Determining Factor
-0.007
(0.005)
-0.001
(0.004)
-0.002
(0.006)
-0.012
(0.008)
-0.003
(0.005)
-0.005
(0.006)
B. Involuntary Transfers
% Minority
-0.027***
(0.001)
-0.027***
(0.001)
-0.027***
(0.001)
-0.021***
(0.002)
-0.022***
(0.001)
-0.022***
(0.002)
Considered (ref- No Seniority)

-0.188**
(0.071)
0.007
(0.122)
All Else Equal

-0.074
(0.051)
0.021
(0.062)
Determining Factor

0.024
(0.063)
0.086
(0.072)
% Minority x Considered
0.006*
(0.002)
0.003*
(0.001)
0.003+
(0.002)
-0.001
(0.003)
-0.000
(0.002)
-0.000
(0.003)
% Minority x All Else Equal
0.001
(0.002)
0.001
(0.001)
0.001
(0.001)
-0.002
(0.002)
-0.000
(0.001)
-0.000
(0.002)
% Minority x Determining Factor
-0.003
(0.002)
-0.002
(0.001)
-0.002
(0.002)
-0.004
(0.003)
-0.003+
(0.001)
-0.003
(0.002)
C. Combined Provision Measures
% Minority
-0.025***
(0.002)
-0.026***
(0.001)
-0.026***
(0.001)
-0.022***
(0.002)
-0.022***
(0.001)
-0.022***
(0.001)
Anzia & Moe

0.037
(0.033)
0.052
(0.042)
% Minority x Anzia & Moe
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
% Minority
-0.023***
(0.003)
-0.024***
(0.002)
-0.022***
(0.002)
-0.015***
(0.003)
-0.018***
(0.002)
-0.017***
(0.002)
Koski & Horng

0.032
(0.021)
0.068**
(0.025)
% Minority x Koski & Horng
-0.001
(0.001)
-0.001
(0.000)
-0.001*
(0.001)
-0.003*
(0.001)
-0.001**
(0.000)
-0.002**
(0.001)
% Minority
-0.026***
(0.001)
-0.026***
(0.001)
-0.026***
(0.001)
-0.022***
(0.001)
-0.023***
(0.001)
-0.023***
(0.001)
TV Restrict.

0.151+
(0.085)
0.239*
(0.105)
% Minority x TV Restrict.
-0.008+
(0.005)
-0.003+
(0.001)
-0.004*
(0.002)
-0.013**
(0.004)
-0.004*
(0.002)
-0.007**
(0.002)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models All
models control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural,
suburban/town) and percent minority. The district and year fixed effect models also include district-level controls for
district enrollment (natural log) and percent free and reduced price lunch. The sample is limited to districts with at least four
elementary schools and districts where the median school has between 15 percent and 85 percent minority students.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 155

Appendix Table 8. Regression of School-Aggregated Teacher Experience and Teacher Education on CBA Transfer and Vacancy
Provisions Using an Alternative Measure of School Disadvantage
% Experienced % Master’s Degree
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
A. Voluntary Transfers
% FRL
-0.050*
(0.020)
-0.044***
(0.011)
-0.045***
(0.012)
-0.047***
(0.012)
-0.007
(0.021)
-0.025*
(0.012)
Considered (ref- No Seniority)

1.184
(0.997)
1.929
(2.310)
All Else Equal

0.048
(0.957)
0.682
(2.463)
Determining Factor

0.523
(2.233)
3.494
(3.918)
% FRL x Considered
-0.022
(0.025)
-0.007
(0.012)
-0.015
(0.014)
-0.029
(0.048)
-0.053+
(0.030)
-0.019
(0.020)
% FRL x All Else Equal
0.034
(0.025)
0.000
(0.011)
0.002
(0.013)
0.012
(0.019)
-0.028
(0.029)
-0.005
(0.015)
% FRL x Determining Factor
0.038
(0.030)
-0.011
(0.028)
0.023
(0.023)
0.052*
(0.024)
0.027
(0.055)
0.007
(0.036)
B. Involuntary Transfers
% FRL
-0.049
(0.030)
-0.048**
(0.015)
-0.050**
(0.018)
-0.046
(0.028)
-0.015
(0.021)
-0.030
(0.019)
Considered (ref- No Seniority)

-0.745
(1.001)
0.133
(2.263)
All Else Equal

0.459
(1.021)
1.479
(2.367)
Determining Factor

2.075+
(1.086)
0.937
(2.574)
% FRL x Considered
0.028
(0.032)
0.013
(0.018)
0.008
(0.021)
0.028
(0.030)
-0.038
(0.029)
0.012
(0.021)
% FRL x All Else Equal
0.016
(0.032)
0.003
(0.018)
0.006
(0.022)
0.005
(0.030)
-0.022
(0.024)
-0.007
(0.019)
% FRL x Determining Factor
-0.022
(0.036)
-0.004
(0.018)
-0.003
(0.021)
-0.035
(0.054)
-0.007
(0.030)
-0.001
(0.028)
C. Combined Provision Measures
% FRL
-0.060***
(0.018)
-0.044***
(0.009)
-0.052***
(0.009)
-0.051*
(0.021)
-0.033
(0.024)
-0.030*
(0.013)
Anzia & Moe

0.246
(0.508)
0.331
(1.247)
% FRL x Anzia & Moe
0.016
(0.012)
-0.002
(0.008)
0.004
(0.008)
0.003
(0.013)
-0.000
(0.017)
-0.002
(0.008)
% FRL
-0.022
(0.033)
-0.020
(0.016)
-0.030
(0.019)
-0.013
(0.055)
-0.011
(0.038)
-0.025
(0.027)
Koski & Horng

0.330
(0.300)
0.158
(0.602)
% FRL x Koski & Horng
-0.007
(0.010)
-0.008+
(0.004)
-0.005
(0.005)
-0.012
(0.021)
-0.007
(0.011)
-0.002
(0.009)
% FRL
-0.041***
(0.010)
-0.045***
(0.008)
-0.046***
(0.008)
-0.047***
(0.014)
-0.032***
(0.010)
-0.032***
(0.009)
TV Restrict.

0.857
(1.051)
0.247
(1.912)
% FRL x TV Restrict.
-0.024
(0.035)
-0.008
(0.019)
-0.009
(0.019)
-0.016
(0.044)
-0.011
(0.040)
0.003
(0.022)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All models
control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural, suburban/town)
and percent minority. The district and year fixed effect models also include district-level controls for district enrollment (natural
log) and percent free and reduced price lunch. The district and year fixed effect models also include district-level controls for
district enrollment (natural log) and percent free and reduced price lunch.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 156

Appendix Table 9. Regression of School-Aggregated Math and ELA Student Achievement on CBA Transfer and Vacancy
Provisions Using an Alternative Measure of School Disadvantage
ELA Score (Standardized) Math Score (Standardized)
2005-06 Over Time 2005-06 Over Time

District Fe

(1)
District
and Year
FE
(2)
District X
Year FE

(3)
District Fe

(4)
District
and Year
FE
(5)
District X
Year FE

(6)
A. Voluntary Transfers
% FRL
-0.024***
(0.001)
-0.022***
(0.001)
-0.023***
(0.001)
-0.019***
(0.001)
-0.017***
(0.001)
-0.019***
(0.001)
Considered (ref- No Seniority)

0.036
(0.059)
0.088
(0.072)
All Else Equal

0.019
(0.056)
0.037
(0.073)
Determining Factor

-0.089
(0.147)
-0.193
(0.184)
% FRL x Considered
0.000
(0.002)
-0.001
(0.001)
-0.001
(0.001)
-0.001
(0.002)
-0.002+
(0.001)
-0.001
(0.001)
% FRL x All Else Equal
0.001
(0.001)
-0.001
(0.001)
-0.001
(0.001)
0.002+
(0.001)
-0.001
(0.001)
-0.000
(0.001)
% FRL x Determining Factor
-0.000
(0.005)
0.001
(0.002)
0.002
(0.003)
-0.000
(0.007)
0.002
(0.003)
0.003
(0.004)
B. Involuntary Transfers
% FRL
-0.022***
(0.001)
-0.022***
(0.001)
-0.022***
(0.001)
-0.016***
(0.001)
-0.016***
(0.001)
-0.017***
(0.001)
Considered (ref- No Seniority)

-0.026
(0.069)
0.045
(0.072)
All Else Equal

0.061
(0.048)
0.124*
(0.056)
Determining Factor

0.111+
(0.059)
0.209**
(0.073)
% FRL x Considered
0.000
(0.002)
0.000
(0.001)
-0.001
(0.002)
0.000
(0.002)
-0.001
(0.001)
-0.001
(0.002)
% FRL x All Else Equal
-0.002
(0.001)
-0.000
(0.001)
-0.001
(0.001)
-0.003+
(0.001)
-0.001
(0.001)
-0.002
(0.001)
% FRL x Determining Factor
-0.003+
(0.002)
-0.002*
(0.001)
-0.002+
(0.001)
-0.004+
(0.003)
-0.004**
(0.001)
-0.004*
(0.002)
C. Combined Provision Measures
% FRL
-0.023***
(0.001)
-0.022***
(0.001)
-0.023***
(0.001)
-0.018***
(0.001)
-0.017***
(0.001)
-0.018***
(0.001)
Anzia & Moe

0.035
(0.023)
0.043
(0.031)
% FRL x Anzia & Moe
-0.001
(0.001)
-0.001
(0.000)
-0.000
(0.000)
-0.000
(0.001)
-0.001
(0.001)
-0.001
(0.001)
% FRL
-0.020***
(0.002)
-0.021***
(0.001)
-0.020***
(0.002)
-0.014***
(0.003)
-0.015***
(0.001)
-0.014***
(0.002)
Koski & Horng

0.048*
(0.020)
0.072***
(0.021)
% FRL x Koski & Horng
-0.001
(0.001)
-0.001*
(0.000)
-0.001*
(0.000)
-0.001
(0.001)
-0.001**
(0.000)
-0.001*
(0.001)
% FRL
-0.023***
(0.001)
-0.023***
(0.000)
-0.024***
(0.000)
-0.018***
(0.001)
-0.018***
(0.001)
-0.019***
(0.001)
TV Restrict.

0.076
(0.058)
0.076
(0.077)
% FRL x TV Restrict.
-0.001
(0.003)
-0.001
(0.001)
-0.002
(0.001)
-0.002
(0.003)
-0.001
(0.001)
-0.003
(0.002)
Notes: *p<.05 **p<.01 ****p<.001; Results for each TV measure are derived from separate regression models. All
models control for school-level student enrollment (natural log), student-to-teacher ratio, school location (urban, rural,
suburban/town) and percent minority. The district and year fixed effect models also include district-level controls for
district enrollment (natural log) and percent free and reduced price lunch. The district and year fixed effect models also
include district-level controls for district enrollment (natural log) and percent free and reduced price lunch.
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 157

Appendix 10- Creation of Overall Contract Restrictiveness Measure

To generate a measure of overall contract restrictiveness, I followed the PIIR method
pioneered by Strunk and Reardon (2010). The original contract strength measure produced from
the PIIR method was constructed with the objective of maximizing the ability of the measure to
discriminate between contracts. As described in depth in Strunk and Reardon (2010), earlier
work using a PIIR-generated measure of contract restrictiveness began with a close content
analysis of 100 randomly-selected CBAs in place in the 2005-06 school year. A single coder read
all 100 CBAs and constructed a coding rubric that recorded every single provision in the CBAs.
These provisions were then checked against the California State Education Code, and provisions
that simply echoed the Education Code were removed from the rubric. There remained 639
provisions recorded in the rubric, each coded as “questions” “asked” of California collective
bargaining agreements regarding the presence of specific contractual provisions (e.g. does the
CBA specify a maximum class size?). The list of 639 represent the defined set of items over
which teachers’ unions and school district bargain in California and are analogous to items on a
survey or test with each contract serving as the answer sheet.
A set of five coders then coded the remaining CBAs, with the original single coder
training the new coders and auditing approximately one in every seven CBAs to ensure
consistency between coders. Once the CBAs were coded for these 639 “questions,” the number
of items was then reduced to 334 by selecting provisions based on the conditional probability of
a positive response (i.e. the provision was found in the contract) across all CBAs. In short,
because the CBA restrictiveness measure is built on a conditionally structured framework in
which items with a conditional probability of responding “yes” at or near .5 provide more
information to the measure, we reduced the set of provisions from 639 to 334 based on a given
item’s proximity to the .5 threshold. Additionally, some of the contract questions generate non-
binary responses, which are not immediately usable in the PIIR model. In a similar manner as
mentioned above, these items were transformed into dichotomous indicators in a way that
maximizes the discriminatory power of the CBA strength measure. Cutoff levels were
established such that the conditional probability of a contract responding affirmatively to a given
item is between 0.20 and 0.80. (See Strunk and Reardon (2010) for a more detailed explanation.).
To this point, only items that were covered by the state Education Code or that lacked variation
amongst the CBAs were excluded (meaning that the item would provide no signal about CBA
restrictiveness).
From this set of contract provisions, the final items included in the PIIR model were
selected according to standard test item selection methods. Exploratory Cronbach’s alpha
analysis was performed such that items that were not associated with the underlying trait of
contract restrictiveness at about 0.10 were removed. Then the alpha item correlations were re-run
and items that were not associated with contract restrictiveness at above 0.25 were removed. The
total number of items used in the PIIR model is reduced substantially by using this method. For
example, Strunk (2011) ended up with a final set of 39 items for use in her PIIR model. This
method of selection avoids one of the fundamental problems with previous contract-based
measures by ensuring that items were selected objectively and not based on preconceived beliefs
about which provisions were particularly constraining or flexibility-enhancing for administrators.
In other words, rather than “cherry-pick” the items believed to be most related and representative
of contract restrictiveness, standard test selection methods were used to winnow out items that
The Effects of Teachers' Unions on Congress, Statehouses, and Schools 158

are only weakly related to latent levels of contract restrictiveness while also maintaining enough
content to ensure that the resulting measure still has a high degree of face validity.
Given that the current study uses contracts from four different bargaining cycles (2005-
06, 2008-09, 2011-12, and 2014-15, each contract in each year was coded for the 639 provisions
mentioned above), I now expand on the exploratory Cronbach’s alpha analysis for item selection
defined above. As CBAs change over time, it is conceivable that a different set of contract items
will emerge as representative of the latent level of contract strength. To take into account the
multiple years of data, I modified the process by stacking the item-level data, treating each CBA-
year as independent, and then proceeded with the item selection process. This method strikes a
balance between maintaining some degree of continuity in the measure over time (e.g. keeping
items at the same cut points) while also allowing the measure to incorporate information from
more than one year of contracts (e.g. using all years in the alpha item correlations). After going
through the process outlined above, the final measure includes 34 contract items in all three years
of CBAs (listed in Appendix 11). These items were then employed in the PIIR model to create
the overall measure of contract restrictiveness.

The Effects of Teachers' Unions on Congress, Statehouses, and Schools 159

Appendix Table 11. Items in the overall CBA restrictiveness measure
Item Number Provision
Association Rights
1 Association members or presidents are promised leave.
2 The contract specifies an amount of release time for the association per year.
3 The contract specifies who pays for general association release time.
4 The association president (or designee) gets additional leave time.
5 The contract specifies who pays for the association president’s leave.
6 The contract specifies the total number of days of release time the association president receives
per year.
7 ≥ 10 days
8 ≥ 20 days
9 ≥ 40 days
10 The association president receives full-time leave.
Compensation
11 Members receive a bonus for having a PhD/EdD.
Class Size
12 The contract addresses class size.
13 The contract specifies a particular class size.
14 The district must balance class sizes within a specific period of time.
15 The district must balance class sizes within three weeks of the stat of the year or semester.
16 The district must take action if the class size is exceeded.
17 The district must take action by a specific time if class size is exceeded.
18 Class size actions be taken within three weeks
19 Specific actions must be taken if class size is exceeded.
Evaluation
20 Permanent members to use an alternative evaluation process with satisfactory prior performance.
21 Permanent members can use an alternative evaluation process.
Grievances
22 The board does not make final/binding decisions on grievances.
23 Grievances do not go to the board.
24 Grievances can go to arbitration.
25 Arbitration is the final stage in the grievance process.
26 Grievance arbitration is binding.
Non-Teaching Duties
27 There are restrictions on the length and/or number of faculty meetings.
28 There are time constraints on faculty meetings.
29 There are constraints in the number of faculty meetings.
Transfers & Vacancies
30 Seniority is addressed as a factor in who is voluntarily transferred.
31 Seniority is a factor in who is voluntarily transferred at least when all else is equal.
32 There are limits on the frequency with which members may be involuntarily transferred.
33 The CBA outlines specific causes for which a member may be involuntarily transferred.
School Days & Hours
34 The CBA specifies the length of the school day in instructional minutes.

Abstract (if available)

Abstract The ongoing public debate in the United States regarding teachers’ unions recently culminated in a foray of unprecedented legislative and court challenges that may fundamentally alter the way organized labor operates in education. Even as these policy changes threaten teachers’ unions, the small but growing literature that examines unions’ impact on the U.S. education system provides few definitive conclusions to guide policymaker’s judgments on whether to reduce, expand, or preserve the role of unions in public education. As the reforms have outpaced the evidence, it is increasingly clear that these new reforms are based more on political rhetoric and anecdote than on evidence about the appropriate role of teachers’ unions. As a result, the recent policy changes may have null or even adverse impacts on K-12 education. Consequently, through three empirical studies, this dissertation extends the small body of literature on how teachers’ unions influence federal, state, and school policy and operations with hopes of building a stronger research base to guide policymakers’ decisions.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Loaded questions: the prevalence, causes, and consequences of teacher salary schedule frontloading

PDF

The role of the timing of school changes and school quality in the impact of student mobility: evidence from Clark County, Nevada

PDF

School resource allocation in times of economic boom and bust

PDF

The local politics of education governance: power and influence among school boards, superintendents, and teachers' unions

PDF

No place like home: a three paper dissertation on K-12 student homelessness & housing affordability

PDF

Teacher evaluation reform in situ: three essays on teacher evaluation policy in practice

PDF

Exploring threats to causal inference in empirical education research

PDF

Teacher perceptions of evaluation policy in Hawaii

PDF

Building networks for change: how ed-tech coaches broker information to lead instructional reform

PDF

Titrating the solution: the diffusion and institutionalization of the logic of continuous improvement

PDF

Tough conversations and missed opportunities: implementing district policies for racial equity

PDF

State policy as an opportunity to address Latinx transfer inequity in community college

PDF

The influence of globalization and student participation in science fairs on 21st-century skill development, school leadership, instructional practices, and female students’ interest in science, ...

PDF

Choosing wisely: a three paper dissertation exploring how parents evaluate and choose schools

PDF

Who learns where: understanding the equity implications of charter school reform in the District of Columbia

PDF

The impact of Connecticut legislators on teacher diversity

PDF

How can I help you? A study of onboarding and ongoing supports for new teachers

PDF

Exploring heterogeneous effects of performance-based funding: Implications for equity and policy

PDF

Factors impacting the effectiveness of mentor teachers in a national teacher residency

PDF

Disruptions to the traditional textbook narrative: lessons from district leaders and teachers in California

Asset Metadata

Creator Marianno, Bradley David (author)

Core Title Levels of interest: the effects of teachers' unions on Congress, statehouses, and schools

School Rossier School of Education

Degree Doctor of Philosophy

Degree Program Urban Education Policy

Publication Date 08/03/2018

Defense Date 08/02/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag collective bargaining,Education Policy,OAI-PMH Harvest,Politics,teacher quality,teachers' unions

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Strunk, Katharine (committee chair), Cowen, Joshua (committee member), Marsh, Julie (committee member), Painter, Gary (committee member)

Creator Email bdmarianno@gmail.com,marianno@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-48913

Unique identifier UC11668672

Identifier etd-MariannoBr-6623.pdf (filename),usctheses-c89-48913 (legacy record id)

Legacy Identifier etd-MariannoBr-6623.pdf

Dmrecord 48913

Document Type Dissertation

Format application/pdf (imt)

Rights Marianno, Bradley David

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA