University of Southern California Dissertations and Theses

Does intense probation monitoring of truants work? An empirical econometric analysis of its effect on school and individual attendance, grades, and behavior

(USC Thesis Other)

Does intense probation monitoring of truants work? An empirical econometric analysis of its effect on school and individual attendance, grades, and behavior

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy subm itted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. ProQuest Information and Learning 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 800-521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DOES INTENSE PROBATION MONITORING OF TRUANTS WORK? AN EMPIRICAL ECONOMETRIC ANALYSIS OF ITS EFFECT ON SCHOOL AND INDIVIDUAL ATTENDANCE, GRADES, AND BEHAVIOR by Charles Edward Pell A Thesis Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF ARTS (ECONOMICS) May 2000 Copyright 2000 Charles Edward Pell Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 1405220 ___ ® UMI UMI Microform 1405220 Copyright 2001 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. Bell & Howell Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U NIVERSITY O F S O U T H E R N C A LIFO R N IA T H E GRADUATE S C H O O L U N IV ER SITY P A R K L O S A N G E L E S . C A L IFO R N IA 0 0 0 0 7 This thesis, written by CHARLES E D W A R D PELL_______________________ under the direction of h is Thesis Committee, and approved by all its members, has been pre sented to and accepted by the Dean of The Graduate School, in partial fulfillm ent of the requirements fo r the degree of M A STER OP ARTS IN ECO N OM ICS Dtmu THESIS COM MITTEE Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C o n t e n t s TABLES.................................................................................................................................iii FIGURE................................................................................................................................. iv INTRODUCTION ...............................................................................................................1 I. THE TRUANCY-CRIME PROBLEM........................................................................ 4 II. PROBLEM DESCRIPTION...................................................................................... 7 A . A n t i-T r u a n c y A p p r o a c h e s .............................................................................................................. 7 B. C o m m u n it y E a r l y Id e n t if ic a t io n P r o g r a m (CEEP).......................................... 15 C. T h e S c h o o l s ............................................................................................................................................. 23 D. T h e sis Q u e s t i o n ..................................................................................................................................... 26 HI. ANALYSIS................................................................................................................27 A . D a t a ................................................................................................................................................................28 1. Individual Characteristics........................................................................................ 29 2. Attendance.................................................................................................................. 29 3. Classroom Grade Marks.......................................................................................... 33 4. Behavior......................................................................................................................34 5. Special Education Status.......................................................................................... 35 6. CEIP-Supervised Students....................................................................................... 35 7. Dummy variables...................................................................................................... 38 B. M e t h o d s t o A n a l y z e CEIP’s E f f e c t s ...................................................................40 1. Choice o f Dependent and Independent Variables..................................................42 2. Between and Fixed Effect Regressions.................................................................... 43 3. Differences-in-Differences....................................................................................... 44 C . R e s u l t s .................................................................................................................... ...46 1. Between-Group (Student) Regressions.................................................................... 46 2. Ordinary Least Squares and Fixed Effect Regressions.........................................48 3. Differences-in-Differences Results...........................................................................64 4. Supervision Results...................................................................................................68 D . I n t e r p r e t a t i o n - I n t u i t i o n ............................................................................................................... 69 1. School Effect ..............................................................................................................69 2. Individual Effect........................................................................................................ 70 CONCLUSION................................................................................................................... 71 BIBLIOGRAPHY...............................................................................................................73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLES T a b l e 1: E t h n ic C o d e L i s t ......................................................................................................................... 2 9 T a b l e 2: A t t e n d a n c e C o d e s ....................................................................................................................2 9 T a b l e 3: B E T W E E N -G r o u p C o e f f ic ie n t E s t im a t io n s............................................................4 7 T a b l e 4: i n (T o t a l A b s e n c e s ) M odel-i .....................................................................................................4 9 T a b l e 5: l a (T o t a l A b s e n c e s )M odel-2 .....................................................................................................5 0 T a b l e 6: l n ( T o t a l A b s e n c e s ) M odel-i ( I n t e r m e d ia t e S c h o o l s O n l y ) ............................5 2 T a b l e 7: l a (T o t a l A b s e n c e s )M odel-2 (In t e r m e d ia t e S c h o o l s O n l y ) ............................53 T a b l e 8: T R U A N C Y A b s e n c e Sm odel-i ..................................................................................................5 4 T a b l e 9: T R U A N C Y A b s e n c e s M odel-2..................................................................................................5 5 T a b l e 10: T R U A N C Y A b s e n c e Sm odel-i (In t e r m e d ia t e S c h o o l s O n l y ) .......................5 7 T a b l e 11: T R U A N C Y A b s e n c e s M odel-2 (In t e r m e d ia t e S c h o o l s O n l y ) ...................... 5 8 T a b l e 12: ^ ( A v e r a g e C it iz e n sh ip)M odel-i........................................................................................ 5 9 T a b l e 13: l n ( Av e r a g e C itiz e n s h ip )M odel-2........................................................................................ 6 0 T a b l e 14: G P A m odel-i ........................................................................................................................................61 T a b l e 15: G P A m odel-2........................................................................................................................................62 T a b l e 16: B G I p e r io d- b y - p e r io d a v e r a g e s .................................................................................... 6 5 T a b l e 17: S U V A p e r io d - b y - p e r io d a v e r a g e s ...............................................................................6 5 T a b l e 18: D IF F E R E N C E S in p e r io d - b y - pe r io d a v e r a g e s (B G I - S U V A ).................. 6 6 T a b l e 19: D if f e r e n c e s- in - D if f e r e n c e s E s t im a t e s o f C E IP ’s im p a c t ...........................6 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE F ig u r e 1: BGI— S U V A A t t e n d a n c e C o m p a r is o n Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. INTRODUCTION 1 Southeast and East Los Angeles experience intensive juvenile gang and drug infestation problems, with corresponding high juvenile crime rates.1 Predictably, these areas have some of the highest incidents of truancy and dropout rates within Los Angeles County. For example, although one-third of all students who enroll in high school • y nationally drop out before graduation,- in Los Angeles County, this rate reaches as high as sixty-three percent for some minority students, and averages approximately forty percent.3 In the Montebello Unified School District (MUSD),4 the overall minority high school dropout rate reaches as high as fifty percent.5 Although most educators and law enforcement personnel realize the long-established connection between juvenile crime and school attendance (or truancy), few law enforcement or educational programs have attempted to link, and none has succeeded in linking, truancy prevention with juvenile crime prevention.6 Cognizant of this lack, the Los Angeles County Probation Department (LACPD) created and implemented the 1. See Interview with Eddie Velasquez, Administrative Assistant to the Superintendent of MUSD (former Principal of Bell Gardens Intermediate School (BGI); former LASD deputy sheriff) in Montebello, Cal. (Nov. 20, 1998). 2. See John Wirt, Tom Snyder, Jennifer Sable, Susan P. Choy, Yupin Bae, Janis Stennet, Allison Gruner & Marianne Perie, U.S. Dep’t o f Education, The Condition o f Education 1998, at 30-43 (Rebecca Pratt & Ginger Rittenhouse eds., 1998). Nationwide, eight percent o f public high school students were absent daily. See Nat’I Ctr. for Educ. Statistics, U.S. Dep’t o f Education, Schools and Staffing Survey, in The Condition o f Education 1996, Supplemental tbl. 42-1 (1996). 3. See Information Analysis Unit, Los Angeles County Office o f Education, The Condition o f Public Education in Los Angeles County 1996, at tbl. 4.2 [hereinafter L.A. schools 1996] (showing graduation rates for Los Angeles County Public School students for 1981 -96). See generally Charles Edward Pell, The Effects in California o f the Legislative Cuts in Adult Education on High School Student Achievement and Satisfaction (1996) (unpublished M.Ed. thesis, Whittier College School of Education) (on file with Whittier College Main Library) (listing the different drop-out rates for the Montebello Unified School District). 4. The MUSD has 18 elementary schools, 6 intermediate schools, 4 high schools, and 4 adult schools. See Finger Tip Facts, 1997-1998 Montebello Unified Sch. District (1998). 5. See Pell, supra note 3, at 46. 6. See, e.g., U.S. Dep’t o f Education & U.S. Dep’t o f Justice, Manual to Combat Truancy 1 (1996) (“In Minneapolis, daytime crime dropped 68 percent after police began citing truant students. In San Diego, 44 percent of violent juvenile crime occurs between 8:30 a.m. and 1:30 p.m.”). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 Community Early Identification Program (CEIP), an innovative program that uses a unique combination of state statutes to combat truancy, and consists primarily of school- based, pre-offense monitoring of potential juvenile offenders.7 In three sections, this Thesis details the truancy-juvenile crime problem, explains how CEIP functions, describes two schools within the MUSD at which CEIP operated, and reports and interprets the attendance, grade mark, and behavioral data at these schools between 1995 and 2000. This Thesis begins in Part I by reviewing the contemporary problems of juvenile crime and truancy and their connection. As a background to the problem that this Thesis examines, Part II first briefly examines some of the main truancy-prevention programs used in California, such as the use of “contributing to the delinquency of a minor” statutes, and then identifies each program’s major shortcomings. Part II then continues by introducing our program, Los Angeles County’s CEIP, which the LACPD currently operates in the MUSD, and then describes each of its different components. Although most people envision probation officers entering the scene after a person has been adjudicated guilty (which is the most common method of their involvement), CEIP employs probation officers to battle children’s truancy, by use of informal probation, before any criminal involvement has occurred. That is, once a school-aged child qualifies as a truant under state law (defined as three or more truancies in one school year), the probation officer can officially file against the child in juvenile court as a “truant” and/or choose to file against the parent for “contributing to the 7. Incidentally, CEIP won the “Most Innovative” program award from the LACPD in 1995. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 delinquency” o f his child. In lieu of this formal (and drastic) filing, CEIP voluntarily places the child on a six-month period of probation supervision, with the hope that such supervision will prod the child to get “back on the right track.” This Part of the Thesis then describes the two schools, Bell Gardens Intermediate (BGI) and Suva Intermediate School (SUVA), from where the data were collected. Last, this Part describes the question that this Thesis aims to answer, which is whether CEIP-supervision had lasting effects on either the supervised group’s attendance, behavior, or grades, or alternatively, on BGI’s (the school at which it primarily operated) attendance, behavior, or grades (whether or not it had individualized effects). Part III describes the empirical data that I collected for the 1995-2000 time period, which are really of two types. First, I collected the attendance, behavioral, and GPA data for every child at both of the individual schools, BGI and SUVA. Among the attendance data are 14 different types of absences, such as dental, medical, illness, tardies (late), or truancies- The behavioral data include ratings by each child’s teacher, and the number of formal and informal suspensions. Furthermore, I also collected each child’s age, ethnicity, gender, and special education status. Second, among the approximately 450 supervised individuals, in addition to the above information, data include interesting facts such as, inter alia, the city of residence, the parents’ dates of birth, the number of siblings, whether the family collects welfare, the number of people in the home, and the number of parents living in the home. Last, I collected data on the types of contacts that the supervising probation officer had with the child, such as number of scheduled home calls and number of behavioral-based contacts {e.g., whether the probation officer was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 called by the parent for assistance or if the child experienced problems in school, such as fighting) — information that I hoped would help me to conclude whether more of one type o f contact (e.g., home visits) had better effect than another. After explaining each type o f data, Part m continues by explaining the methods that I used to evaluate the data, which include using differences-in-differences (DD) and fixed effect and between-group regressions to measure the school effects, and ordinary least squares (OLS) regression to measure the effect of the program on the individuals themselves. Finally, Part HI describes in the results and attempts to interpret them. This Thesis concludes that because of CEIP’s positive school effects on attendance and student behavior, as a policy matter, CEIP should be continued, even though its effects on the supervised individuals themselves was not significant at any reasonable level (5%) o f significance. I. THE TRUANCY-CRIME PROBLEM “Overall, crime and crime victimization rates have fallen sharply since at least 1980 (albeit with fluctuations) but, within that decline, crime by... young people has * o proportionately increased.’ Accordingly, juvenile crime represents a major issue in contemporary society, especially in view of the dramatic increase in violent juvenile crime over the past decade.1 0 The number of juvenile Violent Crime Index1 1 arrests in 8. T h e H a n d b o o k o f C r im e & P u n is h m e n t 8 (Michael Tonry ed., 1998). 9. See, e.g., Hunter Hurst, A Closer Look at Juvenile Violence?, Juv. & Fam. Just. Today, Spring 1993 (“There is something about youth violence [today] that is far more frightening to all of us than at any time in recent history”). 10. See Patricia McFall Torbet, U.S. Dep’t of Justice, State Responses to Serious and Violent Juv. Crime 1 (1996) (citing the increasing violent crime arrest rate of juveniles). See also Anne L. Stahl, Drug Offense Cases in Juvenile Court, 1986-1995, U.S. Dep’t of Justice, Fact Sheet No. 81, June 1998 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 1996 was sixty percent above the level measured approximately a decade earlier in 1987. “ For example, an estimated 2,851,700 juveniles were arrested in 1996. Between 1978-93, juvenile arrests for murder rose 177 percent, whereas that rate for adults fell 7 percent.1 4 Furthermore, the 26 percent anticipated increase in the number of juveniles by 2003 has heralded predictions of a juvenile crime explosion during the next century.1 5 Although juveniles continue to commit more serious crimes, probation supervision is often the most severe sanction in juvenile cases.1 6 It necessarily follows that the number (“Juvenile arrests and juvenile court cases involving drug offenses more than doubled between 1991 and 1995”); Melissa Sickmund, Howard N. Snyder & Eileen Poe-Yamagata, U.S. Dep’t of Justice, Juvenile Offenders and Victims: 1997 Update on Violence: Statistics Summary, August 1997 (between 1985 and 1994, juvenile “[cjases involving crimes against persons were up 98%, Violent Crime Index offenses (a subset of person offenses) were up 98%, and weapons law violations were up 156%”); Eric J. Fritsch & Craig Hemmens, An Assessment o f Legislative Approaches to the Problem o f Serious Juvenile Crime: A Case Study o f Texas 1973-1995, 23 Am. J. Crim. L. 563, 564 & n.l (1996) (“Increased numbers o f juveniles are being arrested, especially for violent offenses.”); George E. Furtado, Esq., Juvenile Hearing Boards: Communities Respond to Juvenile Crime, 44-MAY R.I. B.J. 17, 19 n.21 (1996) (“According to the National Fire Protection Association, children and teenagers constituted 55% o f arson arrests in 1994, up about 40% over the last 15 years.”); Kelly Keimig Elsea, The Juvenile Crime Debate: Rehabilitation, Punishment, or Prevention, 5-FALL Kan. JJ-. & Pub. Pol’y 135, 135 (1995) (“Crimes o f murder, rape, robbery, and assault committed by children have increased by 93% over the last decade.”). But see Laureen D’Ambra, A Legal Response to Juvenile Crime: Why Waiver o f Juvenile Offenders is Not a Panacea, 2 Roger Williams U. L. Rev. 277, 277 (1997) (“Some national studies suggest that as low as five percent of violent acts are committed by young people”). It is interesting to note that although the nationwide number o f overall arrests for the FBI Crime Index decreased by approximately 45,000 between 1991 and 1995, the number o f juvenile arrests for the same period increased by approximately 35,000. See Easy Access to FBI Arrest Statistics: 1991-1995 (visited Feb. 5, 1999) <http://www.ncjrs.org/ojjhome.htm>. Los Angeles County’s juvenile violent crime arrest and teen violent death rates are above both the national and state levels. See L.A. SCHOOLS 1996, supra note 4, at tbl. 7.1. Gang crime has increased 34 percent since 1990. See id. at tbl. 7.2 (showing the “Children’s Score Card” for Los Angeles County). The violent crime victimization rate for persons ages 12 and greater has increased 72 percent over the same time period. See id. at tbl. 7.3. 11. The FBI Violent Crime Index measures, by arrests per 100,000 population, the volume of violent crimes nationwide by monitoring four crimes: murder and nonnegligent manslaughter, forcible rape, robbery, and aggravated assault See Howard N. Snyder, Juvenile Arrests 1996, in Office o f Juv. Just & Delinq. Prevention Bulletin, Nov. 1997, at 4 (Dep’t o f Justice, Washington, DC). 12. See id. at 1. 13. See id. at 2. That same year, juveniles comprised 15 percent of all arrests nationwide for murder and aggravated assault, 24 percent for weapons 32 percent for robbery, and 37 percent for burglary. See id. at 4. 14. See Steven D. Levitt, Juvenile Crime and Punishment, 106 J. Pol. Econ. 1156, 1156(1998). 15. See Cheri Panzer, Reducing Juvenile Recidivism Through Pre-Trial Diversion Programs: A Community’ s Involvement, 19 J. Juv. L. 186, 192 (1997). See also Furtado, supra note 11, at 19 (“The projected rate of violent juvenile crim e. . . is expected to double in the next 15 years.”). Also, the number of children enrolled in Los Angeles County schools is projected to increase from approximately 1.6 million to 1.7 million by 2005. See Jim Parker & Lucely Escamilla, The Condition o f Public Education in Los Angeles County 1996 (1997). 16. See Melissa Sickmund, Ph.D., U.S. Dep’t o f Justice, Office of Juv. Just. & Delinq. Prevention Fact Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 o f juveniles placed on probation continues to skyrocket, and the danger of overload is increasingly salient.1 7 A connection between attendance in school (or truancy) and juvenile crime has I o consistently been established: “If they’re not in school, they’re going to be on the streets - and they’re going to learn all the wrong things.”1 9 Recognition of this truancy- crime connection led to the creation and implementation of various truancy-prevention programs. Sheet #54, The Juvenile Delinquency Probation Caseload, 1985-1994 (Mar. 1997) (“In 1994 courts with juvenile jurisdiction handled 1.6 million delinquency cases. Probation supervision was the most severe disposition in nearly 539,000 o f these cases.”). 17. See id. (“[C]ases placed on probation grew 32% between 1985 and 1994.”); Jeffrey A. Butts, Ph.D., U.S. Dep’t of Justice, The National Juvenile Court Data Archive: Collecting Data Since 1927 (Aug. 1997) (“Between 1960 and 1994, juvenile court delinquency caseloads increased 280%”). 18. See, e.g., Alvin W. Cohn, Juvenile Focus, Fed. Probation, Sept. 1997, at 71, 72 (1997) (stating that “ juveniles who skip school are more likely to commit crimes.. . . Truancy is a stepping stone to delinquent and criminal activity.” (quoting Eileen M. Garry, Official, Office of Juvenile Justice and Delinquency Prevention)); Marcia Johnson, Juvenile Justice, 17 Whittier L. Rev 713, 768-69 (1996) (evaluating the different juvenile justice programs that are utilized and arguing that “academic failure is a key variable in the characteristics of delinquent juveniles” and that without proper educational tools, the probability increases that “children will acquire serious criminal tendencies and criminal records”); Elsea, supra note 11, at 141 (“there appears to be a correlation between juvenile delinquency and. . . educational deficiencies.”). It is clear that truancy is a key risk factor for future delinquency. See James W. Payne, Our Children’ s Destiny, Trial, Jan. 1999, at 83, 84 (“failure to correct the truancy problem leads very predictably to more serious delinquent and sometimes violent behavior.”). See also Shay Bilchik, Breaking the Cycle o f Juvenile Crime, Trial, Jan. 1999, at 36, 38 (1999); Marc Le Blanc & Rolf Loeber, Developmental Criminology Updated, 23 Crime & Just 115, 129 (1998) (“Child developmental studies offer a firm consensus that particular conduct problems, aggression, lying, truancy, stealing, general problem behavior-are predictive of later delinquency, as are early educational problems.”); Jane Watson, Crime and Juvenile Delinquency Prevention Policy: Time fo r Early Childhood Intervention, 2 Geo. J. on Fighting Poverty 245, 250 (1995) (“[Sjchool-related factors suspected to be significant [predictors of delinquency] are grade retention and truancy.”). This delinquency includes sex offenses. See, e.g., Earl F. Martin & Marsha Kline Pruett, The Juvenile Sex Offender and the Juvenile Justice System, 35 Am. Crim. L. Rev. 279, 299 (1998) (“The majority [of juvenile sex offenders] have [sic] committed misdemeanors or less serious felonies such as . . . truancy. . . . ”). It also includes serious and violent offenses, e.g., homicide. See Lee N. Robins, Ph JD., What Determines Rates o f Homicide Over Time and Place: Can We Find Out?, 69 U. Colo. L. Rev. 1009, 1011 (1998) ("The prototypic portrait o f the serious offender-to-be is a boy whose conduct problems began by age eight to ten, consisting o f truancy” ) (emphasis added). See also Alison Mitchell, Clinton Likes Boston's Gains Against Youth Crime, Seattle Post-Intelligencer, Feb 20, 1997, at A3 (observing that Boston’s reduction in juvenile homicide was attributed to aggressive law enforcement against truancy). 19. Alison Fee, Note, Forbidding States From Providing Essential Social Services to Illegal Immigrants: The Constitutionality o f Recent Federal Action, 7 B.U. Pub. Int. L.J. 93, 112 (1998). Illustrative of the feelings o f most personnel in law enforcement, Florida’s State Attorney recently remarked that “‘we know that most crimes occur during school hours as a result of truancy . . . .’” Jill A. Lichtenbaum, Note, Juvenile Curfews: Protection or Regulation?, 14 N.Y.L. Sch. J. Hum. Rts. 677, 715 (1998) (quoting Harry L. Shorstein, State Attorney for Jacksonville, Fla.). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. O. PROBLEM DESCRIPTION This Part of the Thesis concisely describes the main programs that are currently used to battle truancy. Then, it describes in detail the program that is the focus o f this Thesis, the Community Early Identification Program (CEIP), outlining each o f its steps, and their relevance to this empirical study. This Part gives descriptive statistics on the schools—one with CEIP and one without— that were used in the “natural experiment” comparison, and the city and school district within which they are. It closes by succinctly stating the question that this Thesis answers. A. Anti-Truancy Approaches Many school districts, cities, and law enforcement agencies combat truancy by utilizing various programs that rely solely on various California statutes such as the Penal, Welfare and Institutions, and Education Codes.2 0 In order to give a background on the current efforts against truancy, this Part o f the Thesis briefly describes the most common anti-truancy programs, highlighting, when necessary, their major problems. Some approaches place responsibility for student truancy on parents. For example, in September 1988, the California State Legislature amended California Penal Code § 21 272 to hold parents criminally liable for the truancy of their children. Not long thereafter, California taxpayers brought suit, on vagueness and overbreadth grounds, to 20. See Julius Menacker, Legal Policy Affecting School Truancy, 119 Ed. Law. Rep. 763 (1997) (evaluating the different programs employed to combat truancy such as grade reductions, placing students in detention facilities, and holding parents liable). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 challenge the constitutionality of § 272’s ability to hold parents liable. In the seminal ‘ 7 9 1993 decision Williams v. Garcetti,~ the California Supreme Court upheld the 93 amendment that allows prosecution of parents under Penal Code § 272.' As a result of Garcetti, many California counties and cities have coordinated efforts to prevent truancy by prosecuting the parents for violation of § 272.2 4 The response to using § 272 this way predictably has been mixed, ranging from praise' to criticism.' CEIP posited that placing the responsibility solely on the parents— by prosecuting them under § 272 but doing nothing to the child— was not the best way to proceed with abolishing truancy. More often than not, the parent would make a legitimate attempt to send the child to school, but the child would refuse to cooperate. Unable to control their children, parents would complain that they were prosecuted even though they did the best they could under the circumstances.2 7 In response, CEIP sought to place some of 21. Cal. Penal Code § 272 (West Supp. 1998). 22. 853 P.2d 507 (Cal. 1993). See also Ginger Thompson, Gang M ember’ s Mother Denies Failure Charge, LA . Times, May 20, 1989, Metro, at 1; Jason Emilios Dimitris, Parental Responsibility Statutes- and the Programs that M ust Accompany Them, 27 Stetson L. Rev. 655, 680 n.154 (1997); Robert W. Welkos, Mother Seized Under Gang Law Cleared, L.A. Times, June 10, 1989, Metro, at 1. 23. See Williams, 853 P.2d at 508-09. 24. See, e.g., K id’ s Truancy Puts M other in County Jail, The News Serving Montebello, Pico Rivera and Santa Fe Springs 1, Mar. 5, 1998 (“A mother has been sentenced to 15 days in county jail and five years’ probation for contributing to the delinquency o f her 13-year-old son and 16-year-old daughter, who have missed hundreds of days of school. . . . ”). 25. See, e.g., Demitris, supra note 22, at 657 (“Parental responsibility statutes can be a useful tool in combating juvenile crime . . . .”); Gloria Molina, Law on Parental Responsibility, L.A. Times, July 11, 1989, Metro, at 2 (giving support). 26. See, e.g., Linda A Chapin, Out o f Control? The Uses and Abuses o f Parental Liability Laws to Control Juvenile Delinquency in the Untied States, 37 Santa Clara L. Rev. 621, 626-27 (1997) (“Punishing or training parents is not an effective solution when bad parenting is not a significant cause of the child’s delinquency.”); Tami Scarola, Note, Creating Problems Rather Than Solving Them: Why Criminal Parental Responsibility Laws Do N ot F it Within Our Understanding o f Justice, 66 Fordham L. Rev. 1029, 1033 (1997) (“[P]arental control laws are inconsistent with our concepts of justice.”); Catherine Clements, Summary, Williams v. Garcetti: The Constitutionality o f Holding Parents Criminally Liable fo r the Acts o f Their Children, 25 Golden Gate U. L. Rev. 417 (1995); Toni Weinstein, Note, Visiting the Sins o f the Child on the Parent: The Legality o f Criminal Parental Liability Statues, 64 S. Cal. L. Rev. 859, 862 (1991) (concluding that prosecuting parents under § 272 “unreasonably interferes with individual rights and should be invalidated as an unjustifiable expansion of state power.”). 27. See Interview with Maria Cardenas, mother o f student attending BGI, in Bell Gardens, Cal. (Mar. 29, 1999) (stating that “I can’t force him to go to school when he don’t want to go.”). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 the responsibility on the child by also imposing sanctions on him for his truancy, which it does by making the juvenile a party to the CEIP contract— rather than holding only the parent responsible. Alternative approaches emphasize use o f pre-trial diversionary programs targeted at diverting certain juvenile offenders away from the formal adjudicatory juvenile justice system into an informal process,' such as informal probation, community-based programs,2 9 or teen court.3 0 Teen court, the most common pre-trial diversionary program, uses peer pressure to alter the behavior of juvenile offenders.31 Within the 39 program, teens act as prosecutor, defense counsel, judge, and jury. ' The juvenile offender pleads his case before his peers, who then (as either a judge or jury) may mete out his punishment. The teen court program consistently produces recidivism rates below that o f other informal programs.3 3 A common complaint voiced against these informal programs is that they do not effectively combat truancy because almost all of 28. See Panzer, supra note 15, at 194-201. 29. The city of Bell Gardens’ Youth Services Bureau (YSB) provides after-school recreational activities for all youth and has specialized programs for at-risk youth, such as boxing classes, police helpers, and after-school tutoring. See Interview with Michael So, Police Officer for Bell Gardens Police Department, in Bell Gardens, Cal. (Mar. 20, 1999). 30. See Panzer, supra note 15, at 198. 31. See id. 32. See id. 33. See, e.g., id. at 198-99 (“The juvenile recidivism rate for Bay County is 33 percent, but the Bay County Teen Court has a recidivism rate of less than 5 percent”) (quoting Chris Patterson, B ay County Bar Forwards Teen Court as Juvenile Sanction Alternative, 69-Oct Fla. B.J. 95 (1995)). For a more complete discussion o f teen court, see Simon I. Singer, Criminal and Teen Courts as Loosely Coupled Systems o f Juvenile Justice, 33 Wake Forest L. Rev. 509, 522-25 (1998). Another relatively new program with great results in reducing truancy is the Youth Accountability Board (YAB), which operates in San Bernardino County, California. See Panzer, supra note 16, at 198-200 (“Recent statistics from the Youth Accountability Board reflect a 97% success rate, county-wide, for juveniles completing the program”). This program structures boards that receive police reports o f juveniles whom the probation officer deems most likely to benefit from the YAB. The board then explains the program to the juvenile and his family, holds a hearing where an investigator and the juvenile present their respective sides of the case, determines the appropriate disposition-restitution, community service, or educadon-and finally presents the disposition to the juvenile in the form of a legally-binding contract See id. at 199-200. Through this mechanism, the YAB gives the juvenile offender an opportunity to “mend his ways” before he becomes enmeshed within the criminal justice system. Rhode Island uses a similar program, entitled Juvenile Hearing Boards (JHB), as a weapon against juvenile crime and truancy. See Furtado, supra note 11, at 17. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 the participants in these programs have been arrested for crimes, not truancies. Likewise, most of the participants are at an age when school improvement intervention may not be effective either because they have already amassed a huge number of truancies (and are therefore far behind in their studies), or because the juveniles are already enrolled in some type of alternative education, for example continuation high school. CEIP addresses all o f these concerns by attacking the problem before it is too late: before the juvenile has been arrested or expelled from school. In contrast to the previous two truancy-prevention programs, a deputy probation officer (DPO) can herself file a petition in juvenile court to initiate prosecution against the child. The DPO has two different avenues to file this petition in juvenile court, both of which make the juvenile a “ward of the court.”3 4 First, California Welfare and Institutions Code (“WIC”) § 6013 5 contains the main provisions regarding minors who are placed within the jurisdiction o f the juvenile court for truancies3 6 or disobedience of 37 their parents. If the minor misses the number of school days constituting habitual truancy or continues to disobey his parents, “the minor is then within the jurisdiction of 34. Cal. Welf. & Inst Code §§ 601-02 (West 1982). Once the juvenile is a ward of the court, the court has the discretion to place him in juvenile hall, juvenile camps, or on probation. See Inten/iew with Stephonie Steinruck, DPO II for LACPD, in Bell Gardens, Cal. (Oct. 21, 1998). 35. Cal. Welf. & Inst Code § 601 (defining habitual truant). See also Cal. Educ. Code § 48262 (West 1993) (same). 36. See Cal. Welf. & Inst Code § 601. The WIC labels a minor as a habitual truant when she has ‘Tour or more truancies within one school year.” Id. at § 601(b). See also Cal. Educ. Code § 48260 (West Supp. 1999) (defining a truancy as “absent from school without valid excuse. . . or tardy or absent for more than any 30-minute period during the school day without a valid excuse”). It is interesting to note that there is an inconsistency between the WIC and Education Code with respect to what constitutes “habitual truancy.” In other words, WIC allows prosecution of a habitual truant after “ fo u r or m ore truancies within one school year.” Cal. Welf. & Inst. Code § 601(b) (emphasis added). However, the Education Code states that “[a]ny pupil is deemed an [sic] habitual truant who has been reported as a truant three or more times per school year.” Cal. Educ. Code § 48262 (emphasis added). 37. Parental disobedience is defined as one “who persistently or habitually refuses to obey the reasonable and proper orders or directions of his or her parents, guardian, or custodian, or who is beyond the control of that person.” See Cal. Welf. & Inst Code § 601(a). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the juvenile court, which may then adjudge the minor to be a ward of the court.”3 8 Second, WIC § 6023 9 places any minor into the jurisdiction of the juvenile court if that minor violates any city, state, or federal law or ordinance,4 0 and is the most common section that DPOs use to initiate prosecution in juvenile court against minors arrested for a crime 4 1 Pursuant to either of the aforementioned sections, a DPO simply enters juvenile court and files a petition alleging that the juvenile falls within the confines of one of the sections, either as a habitual truant or as a criminal law violator.42 The juvenile court then decides whether to declare the juvenile a ward of the court4 3 If it does, the court then metes out the appropriate sanction, which often is only probation.4 4 The drawback with this method of fighting truancy is that by the time the DPO files either of these types of petitions, the goal is not intervention, but rather suppression.4 5 In other words, the DPO seeks not some type of remedial program for the child, but rather incarceration— either in juvenile hall or juvenile camps.4 6 On the other hand, CEIP identifies the child at a time when intervention, not suppression, is the goal, because it targets those youth who either have a limited number of truancies or are at a younger age. For the past decade, Los Angeles County has generally used the ACT program,4 7 which exclusively prosecutes the parents under Penal Code § 272.4 8 ACT is a 38. Cal. Welf. & Inst Code § 601(b) (West 1982). 39. Cal. Welf. & Inst Code § 602 (Minors violating laws defining crime; Ward of court). 40. See id. 41. See Interview with Eddie Velasquez, supra note 2. 42. See Cal. Welf. & Inst Code § 650. The DPO files the petition when proceeding under WIC § 601, whereas the prosecuting attorney files the petition when proceeding under WIC § 602. 43. Technically, a juvenile is not guilty, but rather adjudged to be a ward of the court 44. See Sickmund, supra note 17, at 1. 45. See Interview with Stephonie Steinruck, supra note 35. 46. LACPD has 3 juvenile detention camps and 4 juvenile halls. The camps employ a military-like environment, e.g., awaking at 6 a.m, while the juvenile halls are more akin to traditional jails. 47. See generally Sickmund, supra note 17, at 1. See also Joe Mozingo, D Jl. 's Program Tries to H ead O ff Truancy Early, L.A. Times, Aug. 30, 1998, at B l; Douglas P. Shuit, £L4. Steps Up Pressure on Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 collaborative effort, established by the Los Angeles County district attorney’s office in 1991, and used in sixteen Los Angeles County school districts.4 9 ACT targets twelve- to sixteen-year old students who have three or more unexcused absences from school.5 0 ACT has three steps.5 1 Similar to the approach discussed above, ACT also prosecutes the parent under § 272, but in addition, it institutes some preliminary steps before actually initiating the prosecution. Rather than simply file against the parents once the child amasses three or four truancies,5 2 ACT tries to motivate the parent to reduce the truancies by means of a mass parent-deputy district attorney meeting. If this meeting does not produce the desired increased attendance, the ACT program then provides for a Parents o f Truants, L.A. Times, Dec. 5, 1997, at B1. 48. See generally Interview with Eddie Velasquez, supra note 1. 49. See Mozingo, supra note 49, at B 1. 50. See Susan Gaertner, Three Strikes Against Juvenile Crime: Pre\>ention, Intervention and Detention, The Prosecutor, Nov/Dec. 1996, at 18,20. 51. First, the school refers the student to the prosecutor as a habitual truant, see Cal. Educ. Code § 48262 (West 1993) (“Any pupil is deemed an habitual truant who has been reported as a truant three or more times per school year”); see also supra Part II.A, which prompts the Los Angeles County District Attorney’s office to inform the student’s parents that their child has a truancy problem. Shortly thereafter, the parent is required to attend a meeting held by the District Attorney’s office and the school, at which a deputy district attorney explains their obligations as parents under the California Penal Code § 272, see supra Part II.A (giving complete discussion o f § 272’s requirements) and California Education Code § 48200, see Cal. Educ. Code § 4-8200 (stating that “each parent, guardian, or other person having control or charge o f the pupil shall send tlie pupil to the public full-time day school . . . .”). See also Cal. Dep’t of Educ, School Attendance Review Boards Handbook 17-21 (Sheila Burton ed , 1995) [hereinafter SARB Handbook]. This is not a one-on-one meeting between the parent and deputy district attorney, but rather is a mass meeting where the deputy district attorney informs fifty or more parents of the possible truancy problem and the personal ramifications of their child’s continued truancy. Second, if attendance does n ot improve within a prescribed time period, usually three to four months, a School Attendance Review Team (SART) meets, at which time a deputy district attorney and school officials work together with the parents to formulate a contract to improve the student’s attendance. The school assistant principal in charge of discipline or attendance, a DPO, the student, his parents, a deputy district attorney, and school attendance officer attend the SART meeting. See id. at 1-2. This meeting is the second chance the parent(s) and juvenile have to improve the attendance. Also, the school may impose remedial measures at this meeting, such as after-school tutoring, parenting classes, or weekend study. Additional solutions include recommending changes in the student’s program, arranging for changes in teachers’ assignments, or assisting students in securing employment See id. at 2. The parents and student sign the SART contract, which stipulates that attendance will improve and that both the parent and child understand prosecution will follow if attendance does not improve. Finally, if the student continues to be absent without an excuse and the parents have not established that they cannot control the student, the deputy district attorney will charge the parents with “contributing to the delinquency of a minor” under Penal Code § 272. See supra Part II.A (giving complete discussion o f using California Penal Code § 272 to file charges against the parents). This carries a penalty of up to one year in jail and a 52,500 fine. See generally Cal. Penal Code § 272. 52. Without any precursor steps, the prosecutor can file against the parents once the three or four truancies occur. See supra Part II A . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 contract, formulated during the SART meeting, that gives the parent and child another opportunity to improve attendance before finally being prosecuted under Penal Code § 272. ACT had great success in bringing children back into school.5 3 Unfortunately, the former truants, who did not want to attend school in the first place, then caused disruptions in school by fighting with other students or swearing at teachers, and the school understandably suspended or expelled them.54 ACT thus exacerbated the problem by leading ultimately to more school absences, the exact evil against which it was designed to combat. Furthermore, a child expelled from school is not technically truant, so California Penal Code § 272 is inapplicable against the parents.5 5 CEIP addresses both flaws: Both parents and child have incentives to reduce the child’s truancies, and CEIP can prosecute a child for additional acts not reachable under other programs, such as not following school rules. Many communities leave the problem of truancies, as they (regrettably) often leave the education, supervision, and parenting of their children, to school districts. These school districts often use nothing more than the Education Code56 to impose penalties directly upon the truant.5 7 For example, upon the second truancy within the same school year, the school, using only the Education Code, may assign the juvenile to an after- 53. See Interview with Eddie Velasquez, Reserve DPO for LACPD, in Bell, Cal. (Jan. 10, 1999). 54. See id. 55. In other words, the juvenile is not technically truant because, although he is not present in school, the absence is not a truancy because the school expelled him. See Cal. Educ. Code § 48260 (West Supp. 1999). 56. Section 48264.5 o f the California Education Code delineates the penalties available for the first through the fourth truancies within one school year. See Cal. Educ. Code § 48264.5 (West Supp. 1999). 57. Occasionally, whereas the Penal Code may only have a general provision that applies, the Education Code will often have a specific provision that is exactly tailored to the specific violation. See. e.g., Michael D. Harris, Lawyer Charged Over Doctored UCLA Transcript, L-A. Daily J., Feb. 12, 1999, at 1 (discussing use of Education Code to prosecute for “fraudulently preparing transcripts,” instead o f the Penal Code’s general forging document provision); Michael D. Harris, UCLA Student Pleads to Ruse to Get Hired, L.A. Daily J., Feb. 9, 1999, at 1 (same). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 58 school or weekend study program. Likewise, upon the fourth truancy within the same school year, the school must classify the juvenile as a “habitual truant” If he is thereafter adjudged a ward of the court under WIC § 601, the school may impose sanctions, including: 1) imposition o f twenty to forty horns of community service; 2) payment of a $ 100 fine (for which the parent may be jointly liable); or 3) suspension or revocation of driving privileges according to § 13202.7 of the California Vehicle Code.3 9 Some teachers have also tailored their classes to make attendance comprise a non- negligible portion of grading, such as conducting pop-quizzes or simply allocating a certain percentage of the class grade to attendance points.6 0 Thus, if the student is absent, he cannot get these points, and his grade in the class will suffer.6 1 These school- based programs have some problems when they act solely using the Education Code to combat truancy. If assigned to after-school or weekend study, the truants were intractable, and simply refused to comply; if driving privileges were suspended or 62 revoked, they nevertheless continued driving. The juveniles simply are not frightened by the sanctions the Education Code affords the schools.6 3 The only school-based programs that succeed are those that use incentives such as field trips, and generally even 58. See Cal. Educ. Code § 48264.5(b). 59. See id. § 48264.5(d). The last option only applies if the pupil has attended an attendance review program conducted by the school or probation office. See id. § 48264.5(c). But see Cal. Veh. Code § 13202.7 (West Supp. 1999) (allowing the court to suspend the habitual truant’s driver’s license without requiring a SART meeting). 60. See Interview with Roshan Shah, Classroom Teacher for Montebello High School, in Montebello, Cal. (Nov. 29, 1998). 61. See id. 62. See Interview with Eddie Velazquez, supra note 2. 63. See id. Likewise, reducing grades has not been very successful with high school and older intermediate school students because they tend not to care as much about their grades as younger students do. See Interview with Roshan Shah, supra note 62. Furthermore, some students have brought suits claiming that lowering grades as a result o f truancy violates their due process rights. See, e.g., Julius Menacker, Legal Policy Affecting School Truancy, 119 Ed. Law Rep. 763 (1997). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 these are successful only with younger children.6 4 As discussed in the preceding five subsections, the major complaint voiced against most of the truancy-prevention programs is that they do not begin until too late in the child’s journey down the wrong road. For example, when proceeding against the parent pursuant to Penal Code § 272 for the child’s truancy (either using ACT or directly under § 272), the deputy district attorneys would often wait until the child amassed an inordinate number of absences. By then it was often too late to help the child: she had already developed bad habits, committed other crimes in the interim, was so far behind that returning to school would be futile, or had moved out of the area. Moreover, many people felt that holding parents liable for their child’s truancies was unjust or ineffective.6 5 Last, school-based programs had limited effects on truancy because many children would not follow the discipline meted out by the schools. B. Community Early Identification Program (CEIP) In 1995, the Los Angeles County Probation Department (LACPD),6 6 in conjunction with the County District Attorney’s Office, Office of Education, and the MUSD, 67 initiated CEIP in order to combat school truancy and juvenile delinquency. CEIP is an 68 education-based prevention and intervention program that targets at-risk youth. 64. See, e.g., Interview with Alex Duarte, Classroom Teacher at Caesar Chavez Elementary School for the MUSD, in Bell Gardens, Cal. (Mar. 28, 1999). 65. See, e.g., Scarola, supra note 27, at 1074 and accompanying text. 66. LACPD was established in 1903. See Los Angeles County Probation Department, Los Angeles County Probation Department Homepage (visited Mar. 10, 2000) <http://probation.co.la.us>. It has approximately 4,200 employees, with an annual budget of over S300 million. See id. 67. Steinruck states that she initially began the CEIP because the huge caseload of each DPO prevented her and others from doing much more than supervise the very violent offenders, with the obvious implication that the lesser crimes, e.g., truancies, were not addressed until they reached serious proportions. See id. See also Patricia McFall Torbet, U.S. Dep’t of Justice, Juvenile Justice Bulletin, Juvenile Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 “The [California] Legislature intended to address juvenile delinquency at its inception and at the earliest signs o f delinquency.”6 9 CEEP attempts to meet the legislature’s intention by identifying, as early as possible, youths who may have a high risk o f school failure, a precursor of juvenile delinquency. Thus, it proceeds one step further than the other anti-truancy programs, which usually wait too long to act.7 0 Where traditional probation programs do not begin until the minor has been arrested or has an inordinate number of truancies,7 1 CEIP attempts to identify children at risk before they fail or become involved within the criminal justice system.7 2 Once the child enters 73 CEIP, he and his family receive a family case plan, close supervision by both a reserve and a full-time DPO, and additional resources, such as tutoring, to ensure student success Probation: The Workhorse o f the Juvenile Justice System (Mar. 1996) (“[T]he number of juvenile cases placed on probation (either formally or informally) increased 21 percent, from 428,500 in 1989 to 520,600 in 1993.”). 68. “Los Angeles County’s children are much more likely to be at risk on many levels than children in the state or the nation: whether for low income, crime dropping out of school or joblessness.” See L.A. Schools 1996, supra note 4, at ch. 7. See also id. at tbl. 7.1 (comparing profiles o f children in Los Angeles County and the United States, including low birth-weight babies, teen violent death rate/ages 15-19, juvenile violent crime arrest rate/ages 15-17, and teen birth weight). 69. People v. Adam R. (In re Adam R.), 57 Cal. App. 4th 348,352 (Cl App. 1997). 70. See Johnson, supra note 19, at 768-69 (showing connection between school failure and crime). 71. See infra Part II.C.3. (discussing ACT program). 72. See CEIP Handbook (available at CEIP office, room N-2). 73. The LACPD has approximately 60 reserve DPOs. See Telephone Interview with Greg McCovey, Reserve Program Coordinator for the LACPD, in Downey, Cal. (Mar. 20, 1999). Reserve DPOs must normally possess at least a four-year university degree and pass a rigorous seven-step process, including an in-depth background process. See, e.g., Cal. Gov’t Code §§ 103 l(d)-(e) (West 1995). The selection process to become a reserve DPO has seven steps. First, the reserve coordinator evaluates the completed application to ensure that the educational and/or experiential requirements have been met Second, if the coordinator preliminarily approves the application, the applicant is invited to a personal interview with the coordinator and one or two other reserve DPOs. If accepted, the applicant must then complete the background packet, which includes a comprehensive evaluation of the applicant’s background. A criminal background check, including, inter alia, traffic, and credit checks, is then performed. If satisfactory, the fourth step is a psychological exam conducted by a licensed psychologist Next, a physical exam is given. If the applicant passes all of the preceding steps, he will be scheduled for the next available academy. Presently, the reserve program has three or four academies per year. The academy consists o f approximately 150 hours o f instruction, ranging from visits to juvenile halls and probation camps to the procedures to use law enforcement computers. Additionally, different instructors will detail their individual programs and the attendant laws. If the applicant passes the academy, he will receive his badge and begin his first assignment This is the final step - the six-month period o f probation. During this time period, foe reserve DPO is evaluated by either a full-time DPO or a senior reserve DPO at least three separate times. If no problems have been encountered, foe provisional reserve DPO becomes permanent Reserve DPOs must work a minimum o f 20 hours per month. As stated supra, possible assignments vary widely from CEIP to gang unit assistance to courtroom participation. For more information on LACPD’s Reserve DPO Program, see Los Angeles County Probation Department (last Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 in school. Further, CEIP exploits sections o f the California WIC and Education Codes that other anti-truancy programs neglect. For example, WIC § 236 allows the probation department to establish “activities designed to prevent juvenile delinquency.”7 4 Accordingly, CEIP employs after-school tutoring, parenting classes, and weekend sports, such as basketball leagues.7 5 WIC § 654,7 6 the key section for CEIP, allows the DPO, 77 with parent consent, to initiate a six-month period o f supervision o f the minor. A child may enter CEIP through a number of possible avenues. School officials, 78 parents, the probation or police department, Juvenile Courts, or any community member7 9 may initiate CEIP by completing the service referral form. Each of these referrals is coded in the data section as a dummy variable. CEIP “236 WIC Service Referral” form, which CEIP distributes to the schools and parents, allows anyone to refer to CEIP “any minors that you feel are ‘at risk’” o f truancy, gang involvement, or drug 80 usage. Once the CEBP team receives a referral, its members investigate it, progressing to the next step of CEIP. modified Oct. 18, 1999) <http://www.co.la.ca.us/probation>. 74. Cal. Welf. & Inst. Code § 236 (West Supp. 1998). WIC § 601.5, also gives the county the option to establish an at-risk youth early intervention program. The program’s purpose is to “provide a swift and local service response to youth behavior problems so that fixture involvement with the justice system may be avoided.” Id. 75. In essence, CEIP is an activity “designed to prevent juvenile delinquency,” and is therefore permissible under § 236. See id. Moreover, WIC § 652 explicitly allows the probation officer to make an investigation, “as he or she deems necessary to determine whether proceedings in the juvenile court should be commenced.” See id. at § 652 (West Supp. 2000) (stating that “[wjhenever the probation officer has cause to believe that there was or is within the county, or residing therein, a person within the provisions o f Section 601 or 602, the probation officer shall immediately make such investigation as he or she deems necessary to determine whether proceedings in the juvenile court should be commenced.”). 76. Cal. Welf. & Inst Code § 654 (West Supp. 1998) [hereinafter WIC § 654], 77. See id., stating that “the probation officer may, in lieu o f filing a petition to declare a minor a ... ward o f the court and with consent o f the minor and the m inor's parent or guardian, delineate specific programs o f supervision fo r the minor, fo r not to exceed six m onths ” (emphasis added). This type of supervision is often referred to as “informal probation.” See, e.g., Paul D. v. Superior Court, 158 Cal. App. 3d 838 (1984) (“informal supervision”). 78. See, e.g., id. 79. See Cal. Welf. & Inst Code § 653 (West Supp. 1998) (allowing any person to apply to the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 The CEIP site team, usually comprised of the school administrator or attendance coordinator and a reserve or full-time DPO, will examine the individual circumstances of the youth and perform a skeletal gathering of information that it places into the CEEP student information packet. This packet should include past attendance patterns, any criminal conduct records, and school behavioral records. Furthermore, a reserve DPO will conduct a home call at the child’s residence to ascertain the conditions of the juvenile’s home life. If the child fits within CEIP’s parameters, the CEIP team will 81 advance her to the next step. The primary criteria utilized by the CEIP team to determine whether to admit one into CEIP consist of no violent convictions (because then the juvenile would probably be under formal probation), a limited number of 83 absences, “ and parental cooperation. Please note that this is not a selection effect, i.e., this is not the case that only children whose parents strongly want them are on the program. What is meant by “cooperation” is that the parents sign the forms agreeing to the six-month period of CEIP probation. However, if they choose not to sign, then they could be subject to “contributing to the delinquency” of their child under § 272, because they are not doing what they can to ensure that their child attend school. Thus, most parents sign when presented with the full information of what alternatives they have. probation officer requesting that proceedings be commenced against a juvenile, which then requires the probation officer to investigate to determine whether proceedings in juvenile court should be pursued). 80. CEIP Referral Form (on file with author). 81. As a side note, the choice whether or not to admit the minor into CEIP, rather than forward the case to the court for filing may not be in the total discretion of the CEIP team. See, e.g., John O. v. Superior Court, 169 Cal. App. 3d 823 (CL App. 1985) (juvenile court erred by denying informal supervision to minor on grounds that county lacked adequate funds to provide such supervision); Paul D., 158 Cal. App. 3d at 838 (holding illegal probation department’s refusal to consider minor for informal probation solely because he denied the charges pending against him). 82. Because CEIP concentrates on early intervention, if the student has more than ten absences, absent extenuating circumstances, the CEIP team will not admit the student into CEIP. 83. See CEIP referral evaluation form 3 (1995). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 84 85 A DPO, reserve DPO, school administrator, parent or guardian, police officers), and the minor herself will meet and discuss the conditions of probation that the CEIP 86 team will impose on the juvenile during CEIP supervision. This meeting is the true beginning of CEIP supervision. CEIP conditions of probation can include many things such as obeying all federal and state laws, obeying all instructions and directions of the reserve and full-time DPO, obeying all instructions and orders of parents and/or 87 guardians, urine testing upon request of" a DPO, and reporting to the DPO or CEIP 88 team as directed. Another standard condition of CEIP probation is that the child 89 submit to search by a (reserve) DPO at any time— without probable cause. The CEEP conditions of probation also may require the parent to participate in a counseling or educational program with the minor.9 0 Thereafter, if the child does not meet these conditions, she has violated her CEIP probation, and the DPO may file a petition against her in juvenile court9 1 84. See Cat Welf. & Inst. Code § 283 ( W e s t 2000) (“Every probation officer, assistant probation officer, and deputy probation officer shall have the powers and authority conferred by law upon peace officers”). 85. If the juvenile has an arrest or criminal record, the CEIP team invites the arresting officer to the CEIP initial meeting to add his input regarding this juvenile. See Interview with Dan Garcia, Reserve DPO for LACPD, in Montebello, Cal. (Mar. 15, 1999). 86. Often a Spanish translator is required, see Interview with Eddie Velasquez, supra note 2 (stating that “in more times than not, either another person or I must translate” at the meeting), with the CEIP bilingual forms being utilized. 87. See Cal. Welf. & Inst Code § 729.3 (West 1998) (“[A] condition o f probation may require the minor to submit to urine testing upon the request of a peace officer or probation officer for the purpose of determining the presence of alcohol or drugs.”) . 88. See In re Abdirahman S., 68 Cal. App. 4th 963 (1997) (“[A] condition o f probation which requires or forbids conduct which is not itself criminal is valid if that conduct is reasonably related to the crime of which the defendant was convicted or to future criminality.”). 89. See Interview with Eddie Velazquez, supra note 2. See also Cal. Welf. & Inst. Code § 654.6 (allowing imposition of at least 10 hours o f community service time, with the possibility o f the minor’s parent bearing the cost of the program, if she has the financial capacity to pay). Moreover, any specific needs of the student, such as drug counseling o r after-school tutoring, may be added to the “conditions of supervision,” and if so added, have the force o f law. 90. See Cal. Welf. & Inst Code § 661 (“ [A] parent guardian, or foster parent may be required to participate in a counseling or education program with the minor concerning [against] whom the petition has been filed.”). 91. See Cal. Welf. & Inst Code § 654 (“ If the probation officer determines that the minor has not Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 9 * > After completing CEIP’s supervision training program, " a reserve DPO usually conducts supervision that includes weekly contact with the minor, her family, or the 93 school. Each officer typically supervises ten to fifteen CEIP clients. This close supervision is a key step for CEIP, ensuring a swift response if problems arise.9 4 “Regular home visits and home observation are essential components of a successful delinquency prevention program.”9 5 The reserve DPO visits the home of the child at least bi-weekly.9 6 “Early identification [of truancy problems] begins in the classroom.”9 7 Because most of the CEIP participants have either missed school or experienced problems there, the reserve DPO will initiate contact with the probationer’s teacher(s) within the first week of CEIP supervision. The teacher will voice any concerns, educational or behavioral, that she feels CEIP should address. After completing the family and teacher interviews, the reserve DPO formulates a plan with involved himself or herself in the specific programs the probation officer shall immediately file a petition or request that a petition be filed by the prosecuting attorney.”). 92. Before an officer may supervise a juvenile CEIP probationer, they must obtain the “CEIP Juvenile Supervision Certificate.” This certificate program includes four components, including completing LACPD’s Reserve DPO Academy. 93. Compare this ratio with the national urban median caseload o f 47 probationers per officer. See, e.g., Patricia McFall Torbet, U.S. Dep’t of Justice, Juvenile Justice Bulletin, Juvenile Probation: The Workhorse o f the Juvenile Justice System (Mar. 1996). In Los Angeles County, the average DPO’s caseload surpasses 500 active supervision cases. See Interview with Stephonie Steinruck, supra note 69. 94. See Interview with Edward Tapia, Reserve DPO for LACPD, in Bell, Cal. (Jan. 5, 1999). 95. Watson, supra note 19, at 256. 96. Usually the reserve DPO will schedule one visit with the parent and child, then make a second visit unannounced. The latter visit gives the probation officer an opportunity to obtain a realistic view o f the home life of the juvenile. See id. (“Home observations are particularly useful for the assessment of psychological and emotional problems, especially for either difficult clinical problems or hard-to-reach families.”). During the scheduled visit, the reserve DPO records the home conditions and questions the parent and child regarding any problems or special needs. See Interview with Edward Tapia, supra note 21. Mr. Tapia states that home visits are essential to the CEIP, because they give the CEIP team a realistic view of the problems the child faces on a daily basis, e.g., gang coercion. Furthermore, the home calls give the CEIP “face value” because the community knows that someone is present to help with problems. See id. Often the reserve DPO will also interview neighbors to determine whether or to what extent the juvenile associates with gang members or if any other problems exist that are not apparent from speaking with the juvenile’s family. See id. 97. School Attendance Review Boards Handbook, Operations and Resources 3, Cal. Dep’t o f Educ. (1995) [hereinafter SARB Handbook].. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 98 the CEIP team to assist the juvenile to either “catch up” or focus on problem areas. This aspect o f CEIP supervision coincides with the conditions o f probation, because a common condition of probation is attendance at after-school tutoring programs. The reserve DPO also maintains contact with the ancillary educational providers to ensure that the juvenile attends on a regular basis and to promptly meet any additional need that may arise." Additionally, if the child cannot attend school without causing problems or getting involved with gangs, CEIP may employ the use of County Schools.1 0 0 “The training of parents in discipline and supervision is a successful delinquency prevention device.”1 0 1 Just as the children have certain areas, as discussed in the above paragraph, that need work, parents often also have either parenting,1 0 2 language,1 0 3 or educational deficiencies1 0 4 that additional schooling will assist Recognizing that these needs exist, the CEIP team counsels the parent so that she can obtain the necessary help. The reserve DPO records every action that he takes with respect to the CEIP probationer, beginning with the first meeting. These entries in the record of supervision 98. Additionally, because the teacher now knows that the juvenile is on probation, she can use this knowledge as another tool to ensure compliance with homework assignments and other requests. For example, if the student begins to misbehave, the teacher may remind him that she may have to speak to the DPO about his behavior, this warning often quiets down the student See Telephone Interview with Melodie Santana, Teacher for BGI for the MUSD, in Downey, Cal. (Jan. 6, 1999). Also, the reserve DPO gives the teacher a supervision form to record any behavioral problems, so that the DPO can address them at the CEIP bi-monthly meeting. 99. See, e.g., Interview with Eddie Velazquez, supra note 2. For example, if the child continues to have problems in mathematics, CEIP has an agreement with Montebello Kumon and Bell Gardens Adult School to provide additional mathematics tutoring. See Interview with Jaime Quintero, Teacher for Bell Gardens Adult School for the MUSD, in Bell Gardens, Cal. (Mar. 22, 1999). 100. See, e.g., Larry Springer, Vision 2000, Los Angeles County Office o f Educ., Division of Juv. C t & Community Schools (1998). Los Angeles County operates the largest juvenile court school system in the United States, serving over 40,000 students annually. See id. County schools are schools run by the county, as opposed to the usual school district or city, which are located in juvenile halls, probation camps, community educational centers, and within other schools. See id. 101. Watson, supra note 19, at 255. 102. See L.A. Schools 1996, supra note 4, at tbl. 24-3. 103. See D istrict Profile, Fiscal Year 96/97, Montebello Unified Sch. DisL (visited Feb. 26, 1999) <http://www.ed-data.kl2.ca.us> (showing that 51% of the MUSD has limited English proficient students, as compared to 25% statewide). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 consist of the date of occurrence, action taken, and further recommended action.1 0 5 The DPO records everything he learns from the home and school visits, including the exact language of the probationer. This supervision record is indispensable to CEEP because the reserve DPO refers to it during the CEIP bi-monthly meeting to monitor the probationer’s progress. If the probationer does not succeed under CEIP, the record of supervision acts as a more complete record upon which a court relies when evaluating further delinquent conduct of the juvenile.1 0 6 This record of supervision is what I used to code in the different type of contacts that each probation officer had with the supervised probationers. If the minor has completed the period of supervision to the satisfaction of the CEIP team, that is, no problems have been encountered and the attendance of the minor has improved, the case will be closed and no further action will be taken.1 0 7 For example, if the student had twenty absences in the time period prior to the initiation of the CEEP supervision and at the end of the six months she has had only four absences, although 108 she is technically a “habitual truant,” the great improvement will be considered by the CEIP team when making the determination whether to file a petition against the child. Also, nothing in WIC § 654 proscribes initiating another, possibly shorter, period of 104. See L A Schools 1996, supra note 4, at tbl. 24-3. 105. See infra App. A (LACPD record o f supervision). 106. In other words, if the DPO chooses to file a petition against the juvenile, the record o f supervision gives the judge a more complete picture o f the person with whom he deals. Additionally, if the juvenile is later arrested for a crime, he cannot plead that this is his first offense or that he is an “angel,” especially in light of the copious notes that the probation officer compiled on his truancies, home or school problems, or alternative criminal activity. 107. Compare People ex rel. Kottmeier v. Superior Court, 194 Cal. App. 3d 1536 (C t App. 1987) (petition cannot be brought against minor more than six months after probation agreement for informal supervision) with Cal. Welf. & Inst Code § 654 (West 1998) (“Nothing in this section shall be construed to prevent the probation officer from filing a petition or requesting the prosecuting attorney to file a petition at any time within the six-month period or a 90-day period thereafter. ) (emphasis added). 108. See supra Part III.B. 1. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23 109 supervision. Last, the DPO must prepare a follow-up report of the program, delineating which measures were used.1 1 0 C. The Schools This Part of the Thesis describes the sample, which is the two schools at which I collected and compared the data: Bell Gardens Intermediate School (BGI) and Suva Intermediate School (SUI). Both of these schools are located in the Montebello Unified School District (MUSD),11 1 which is in Los Angeles County, California.1 1 2 Los Angeles County11 3 has the largest population o f any county in the United States, comprising more than twenty-eight percent of California’s residents,11 4 and is exceeded by only eight states.1 1 5 It has a total land area o f4,083.21 square miles, including 1,875 square miles of mountains and 132 square miles of islands. As of January, 2000, the ethnic makeup of Los Angeles County is 43.9% Hispanic, 33.5% White, 12.3% Asian, 10.0% Black, and 0.3% American Indian.1 1 6 In 1999, the County had an unemployment rate of 5.9%, with a per capita income of $25,719, and average earnings per job of 117 $35,626, in 1997. The County leads the state of California in most measures, including such things as percent of housing stock (26.9% in 1999), automobile registrations (4,825,512 in 1998), and manufacturing (31.5% in 1992). 109. See Cal. Welf. & Inst. Code § 654. 110. See id. at § 654(c) (requiring “followup report o f the actual program measures taken.”). These reports are on file at BGI, CEIP office, room A-9. 111. MUSD has twenty-eight schools, six that use CEIP. 112. Los Angeles County has approximately 1,700 schools. 113. See County o f Los Angeles Statistical Data 1 (visited Feb. 15, 2000) <http://www.co.la.ca.us/statistics.htm>. 114. See id. 115. See About Los Angeles (visited Mar. 11, 2000) <http://www.co.la.ca.us/overview.htm>. Los Angeles County’s 1997-98 budget was SI2.6 billion. See id. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24 O f a population o f9,790,000 persons living in eighty-eight cities,11 8 Los Angeles County has a total enrollment of 1,549,833 students, dispersed among 94 public school districts, including 1,134 elementary, 233 intermediate, 165 high, and 90 continuation schools.11 9 Los Angeles Unified School District is the nation’s second-largest school district, with 707,986 students enrolled for the 1999-2000 school year.1 2 0 MUSD serves over sixty thousand students, dispersed among four adult schools,1 2 1 three traditional high schools, one continuation high school, six intermediate schools (two of which are BGI and SUVA), and eighteen elementary schools.1 2 2 The number of K-12 students is approximately 35,000, with an average daily attendance of approximately 32,000. These students predominately come from the cities of Bell Gardens, Commerce, East Los Angeles, Montebello, Monterey Park, Pico Rivera, and South San Gabriel. Fifty-two percent of MUSD’s students are described as “limited 19 1 English proficient.” “ Both of the schools that are the subject of this empirical study are located in the MUSD. As of 1999, the ethnic breakdown of the district is 0.1% American Indian, 5.1% Asian or Pacific Islander, including Filipino, 91% Hispanic, 0.4% black, and 3.5% White.1 2 4 These rates are much more Hispanic-heavy, when compared with the state 116. See County o f Los Angeles Statistical Data, supra note 114, at 2. 117. See id. 118. This figure is for July, 1999. See id. at 1. 119. See County Schools Facts, 1996-97 (visited Mar. 10, 2000) <http://www.laeoe.edu/schooIs/facts.html>. 120. See Kristina Sauerwem & Karima A. Haynes, A Dispatch From the Front Lines, L A Times, Oct. 15, 1999, at B2. 121. It is interesting to note that Montebello Adult Schools serve more students, 36,000, than the entire K-12 program. See Montebello Unified School District, supra note 133. 122. See id. 123. See Montebello Unified Sch. Dist. Finger Tip Facts 1997-1998, supra note 133, at I. 124. See District Profile, M ontebello Unified Sch. Dist. 2 (visited April I, 2000) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 ethnic enrollment rates of 41.3% Hispanic and 37.8% White, and the district’s own numbers in 1992, which were 86.9% Hispanic and 5.3% White. Additionally, over 75% of the district’s students qualify for Free/Reduced Price meals, up from 59% in 1992. Each of the schools is within the 2.5 square mile boundaries of the city of Bell Gardens, California, which is approximately ten miles southeast of downtown Los Angeles. Additionally, Bell Gardens has 9 other schools from the MUSD—six elementary, one high school, and two adult schools. Bell Gardens was the 72d city incorporated in Los Angeles and currently has a population of over 45,000, with a p5 median age o f twenty-two years. “ This population has increased from 42,315 in 1990. “[T]he city’s mostly blue-collar residents have an average income that makes Bell Gardens the third-poorest suburb in the nation.”1 2 6 Especially relevant to CEIP is the fact that approximately twelve thousand of Bell Gardens’ residents (approximately one- fourth) are younger than eleven years. In other words, these younger residents are (or will be soon) in the school system, so some program should be in place to ensure that they attend school. Hispanics and Whites comprise, respectively, 87.5% and 10.5% of the city’s residents, while the median income is a meager $23,819.1 2 7 Less than 50% of the city is comprised o f married couples with children. <http://www.ed. data.k 12.ca.us/>. 125. See About the City o f B ell Gardens (visited Mar. 11, 2000) <http://www.ci.bell- gardens.ca.us/about.htm>. 126. See Cecelia Rasmussen, Community Profile: Bell Gardens, L.A. TIMES, Jan. 3, 1997, at Metro 1 (emphasis added). 127. See id. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26 Bell Gardens Intermediate (BGI) is the largest of the six intermediate schools in the | - J O MUSD, with an average daily attendance (ADA) of approximately 1,900 students. “ Intermediate schools in the MUSD have students in grades five through eight, with approximate ages ten through fourteen. The ethnic distribution of the students at BGI has increased from 93.68% Hispanic in 19921 2 9 to 97.8% in 1999-2000. CEIP first began to supervise BGI students in 1995, with most supervision between 1996-1997. Suva Intermediate School (SUVA) mirrors BGI in most aspects: for example, ethnicity— SUVA has gone from 96.8% Hispanic in 1995 to 97.7% in 2000. Likewise, its size is comparable to BGI, as it has approximately 1,750 students per year. Both BGI and SUVA feed into the same high school, Bell Gardens High School (BGH), which has approximately three thousand students per year. We did not collect ethnic and gender data for the students at BGH that did not come from either BGI or SUVA. D. Thesis Question The goal of this Thesis is to identify the effect of (CEIP’s) intense probation officer supervision of these truants on either the supervised individuals’ attendance, grades, or behavior or alternatively, what effect, if any, CEIP had on each school’s attendance, grades, or behavioral rates. As described above, this school district gives us a great 128. See Montebello Unified Sch. Dist. 1992-1998, Montebello Unified Sch. Dist, Educational Measurement 10 (1998). ADA is the measure of the average number o f students who attend. The state remunerates school districts, and by necessary implication schools, according to the number of students present at the school. See Pell, supra note 4, at 4-8 (discussing definition of ADA). 129. See Bell Gardens Intermediate 1992-1998, Montebello Unified Sch. Dist., Educational Measurement 2 (1998). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27 opportunity to compare the school effects of the program, because the two schools, BGI and SUVA, are a beautiful natural experiment, because they are in the same city and school district, and have virtually the same type o f students by most measures. O f course, each school will definitely have many unobservables, such as the principal’s governing style, etc., that we cannot add to the model, besides with the intercept. III. ANALYSIS This Part of the Thesis describes in detail each type of data that were collected and included in the study, differentiating between the data for the CEIP-supervised and unsupervised individuals. It then briefly explains the methods that were used to evaluate these data, including running between-group, normal ordinary least squares, and fixed- effect regressions, and the Differences-in-Differences estimation technique. It then shifts to the results portion, which returns the results of these regressions, focusing on any effect on attendance, grades, or behavior, both at the school-wide and individual level—findings that indicate that CEIP was successful at increasing overall attendance and reducing truancies at the school-wide level, but that the supervised individuals had positive, but minor (yet significant) effect on both total absences and truancies, and increased GPAs and improved behavior (as shown by a negative coefficient for the citizenship average). Last, this Part then interprets these findings, again differentiating between individuals and schools, and also attempts to explain why CEIP had both a great effect at the school, in addition to the effect at the individual, level. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A. Data In this section of this Thesis, I will describe the different types o f data that I collected for both the supervised and unsupervised individuals. It is important to note at this juncture that the data for the supervised individuals are much more extensive because these are the children with whom CEEP had the most direct contact. Although the LACPD officially began CEIP in late 1995, most o f the supervision and data-keeping of CEIP-supervised individuals began in 1996. Because CEIP involves a six-month period of supervision, I decided to separate all of the attendance, citizenship, and GPA data into six-month periods. Additionally, in order to obtain a more accurate view of the trends of the attendance, grade, and behavioral data, I started at January, 1995, and continued to present, April 2000 (breaking the time periods as follows: ti = 1/1/1995-6/30/1995, t2 =7/1/1995-12/31/1995,..., tu=l/l/2000-4/l/2000). The number of supervised persons varied by time period, with the heaviest supervision in the third (1/1/1996-6/30/1996) through sixth (7/1/1997-12/31/1997) time periods, which together accounted for approximately 85% of the supervised individuals (386/460). These data consist of a panel, which means that many different, individual students have been observed over time, as compared to a simple time-series (where only one individual is followed over time) or cross-sectional data (where many subjects are compared at only one time period). This panel consists of semi-annual performance observations and individual characteristics of approximately fifteen thousand students over five years, from 01/01/1995 to 04/01/2000. The average number o f observations Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 per student was approximately five, ranging from a minimum of one to a maximum of 11. 1. Individual Characteristics First, I collected the data for the individuals that did not change over time, which included each student’s gender, ethnicity, and date of birth. There exist five different types o f ethnic identifiers: Table 1: Ethnic Code List Eth A Asian or Pacific Islander, including Filipinos Eth B Black. non-Hisr»anic Eth H HisDanic Eth I American Indian or Alaskan Native Eth W White At BGI, the percentages of males was 49.53%, averaged over the entire time period. At SUVA, this same measure was 48.91%. 2. Attendance The attendance data include 14 different codes, which represent different types of absences: Table 2: Attendance Codes C.ndp D p srrin tin n D Dental Appointment E Field Trip F Funeral I Illness K Excused Tardv L Late M Medical Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 N Non-Annortionment P Permit S Suspension (Official) T Truancy U Unverified Y Suspension (Unofficial) Z Saturdav Make-up Dav Most of these attendance code descriptions are self-explanatory, but for the ones that are confusing, I will briefly illuminate. The N (Non-Apportionment) Code signals that the school did not receive payment for that day, i.e., the child did not attend at all or missed too much of the day to qualify for either state or federal funding. Likewise, attending a funeral of a person who is not immediate family (father, mother, sister, brother, grandparents, etc., who are covered under F code where the school gets paid) results in an N, because the state and/or federal government does not remunerate the school district for these funerals. Another example of an N would be where the student had to stay home to care for the family (e.g., babysitting) because the parents were unable—a situation where the school does not get paid, but knows what the child was doing. The P (Permit) Code means that the student obtained permission to miss the day (or a set number of periods in the day) for an approved-type activity, e.g., college fair. Another confusion could result from the two different types of suspensions— S and Y, official and unofficial, respectively. A suspension is defined as when the child is removed from the school for a period of one to five days as a result of negative behavior.1 3 0 In both types of suspensions, official and unofficial, the child is removed from the school. The difference, however, between an official and unofficial suspension Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 exists in the procedure that the schools uses to effect that removal and the type of action by the student that led up to the suspension. In other words, the school must follow a set procedure (notification, note sent home, and opportunity to be heard) when officially suspending a child, because that suspension will go into the student’s permanent “cum” (cumulative) file that follows him while he is in the district. On the other hand, for certain behavioral infractions that are not as serious as others, e.g., a fight/scuffle where no injury resulted, the school may decide to punish the child, but not permanently tarnish his record. In that case, the child will be “unofficially” suspended, and told to return the next day or two days. Alternatively, the school might use an official suspension when the student acted very badly, for example, brought a weapon to school. T (for truancy) is the type of absence that CEEP most tries to stop, which is where the student was absent from school for no legitimate reason (compare babysitting, where no official truancy would result, but would still be a negative-type absence, e.g., U, which we would include with our regressions under the dependent variable truancy-type absences, TRUabs). For example, the child may have attended one period, but choose to leave with friends to see a movie. As stated in the CEIP description section, supra, the probation officer may file against a child once he has amassed more than three truancies in any one school year. U, however, may encompass some absences that really are truancies. An Unverified absence is one that the child either had no reason for his absence (e.g., he was ditching) or had a reason that the school did not accept as valid (e.g., went to Mexico to visit 130 See Interview with Eddie Velasquez, supra note 1. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. family). In other words, many absences that could be officially called truancies are fit under this heading because the school does not have concrete evidence that the child was ditching, but the child has no valid excuse either (e.g., the babysitting example). To account for the somewhat subjective quality of the difference between a T(ruancy) and U(naccounted), when I ran some regressions, I included all of these four types together, T, U, N, and L. The last type of absence, Z, Saturday Make-up, describes the situation where the child may have ditched, and thus would have qualified for a truancy, but in lieu o f that has agreed to make up the time on a Saturday to avoid having the truancy on his record. Furthermore, the school can then gain apportionment for this attendance. Attendance is taken a little differently between the intermediate schools (BGI and SUVA) and the high schools (BGH), because the high schools take attendance period by period, while the intermediate schools take attendance on a daily basis. To account for this discrepancy, I had to multiply the intermediate school (day) absence codes by 6, except for L(ate), which legitimately was a one period occurrence. Among all eleven time periods, the total period absences (using variable TOTabs) in the intermediate schools ranged from 0 to 708, with a mean o f27.65 (36.67). The average for SUVA over all time periods was 24.46 (32.77), with a range from zero to 708 absences. The same numbers for BGI were 25.16 (29.50) and ranged from zero to 402. What this indicates is that, when taking the standard deviations into consideration, the average of the total absences is not significantly different between the two schools. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 For the persons who changed intermediate schools one or more times, their intermediate period absence average was 37.29 (39.13) absences, and 27.34 (46.85) period absences in the high school. It thus seems that those who changed schools had worse attendance in the average than those who stayed at one school. At the high school, the mean period absences were 29.99 (43.30), ranging from perfect attendance to a maximum o f564 periods missed! The average total number of absences for all time periods and across the intermediate-high school divide is 27.50 (32.18), ranging from a minimum of zero absences (perfect attendance) to a maximum of 708. 3. Classroom Grade Marks I collected the grade point averages (GPA) for all except the last six-month time periods, i.e., ti — tio. The highest grade that is given was an A, for a value of 4, and lowest was F, for a value of 0. The average GPA over all time periods for all of the schools, both intermediate and high school, was 2.46 (0.795). For BGI, the average GPA was 2.51 (0.729), and for SUVA, 2.79 (0.736). For the persons who changed schools one or more times, their intermediate school grade average was lower than both intermediate schools, resting at 2.356 (0.739). What these averages and standard deviations seem to indicate, however, is that the GPA is not significantly different from one school to the next, on the average. At the high school, the overall average GPA was 2.22 (0.787), with students originating from SUVA having a somewhat higher GPA, 2.258 (0.789), than those who Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. came from BGI, 2.176 (.776). For persons who changed intermediate schools one or more times, their high school GPA was much lower, averaging 1.725 (0.677). Thus, again, the students who changed schools had lower GPA than students who stayed at either school. 4. Behavior There are really at least two different measures that record the students’ behavior. The first that I extracted from the data was each individual class’ citizenship mark. This mark is given by the instructor and ranges from 1 (good) to 3 (extremely poor). I took an average for each, except the last, of the six-month time periods. Second, both types of suspensions are good measures for behavior, because the most often action that leads to a suspension is negative behavior. These suspensions are under the attendance category supra. For t\ through no, the citizenship mark average was 1.25 (0.363) at BGI, and 1.228 (0.403) at SUVA. This value for the students who changed was 1.374 (.493). Because the higher the average, the worse the behavior, school-changing students again had worse averages than those at either school. At the high school, the citizenship average was 1.289 (0.384). For the other indicator of behavior, formal suspensions (S), the average of periods suspended at BGI was 0.09 (0.48), while at SUVA, it was 0.11 (0.65). At the high school the average was 0.52 (4.90), with students that had come from SUVA having a slightly higher average, 0.51 (4.89), than those students from BGI, 0.49 (4.36). As with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 the other averages, the differences between or among the schools are not really significant. 5. Special Education Status Another indicator that is part of the data set is whether the student is special education or not. This variable, SPL_ED, varies by i and t, because sometimes children come on and off the status. These students are regularly kept in the same classroom with the same teacher for the whole day, unless they are able to be “mainstreamed,” which is where they have some classes with the regular students and some with the other special education students. Special education students comprised 0.7% of BGI’s student body, as compared to 0.566% o f SUVA’s students. 6. CEEP-Supervised Students This group (n=460) consists of those individuals whom a reserve or full-time deputy probation officer supervised. Among those, approximately one hundred have extensive personal data, including: a. probation officer («— 5) who supervised the CEIP participant b. parents’ date of birth c. number of siblings d. whether the family collected welfare (AFDC) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 e. student’s primarily language (English or Spanish) f. whether the student is bilingual g. referring agency: school (n = 453), police or probation (n = 2), or parents (n = I)1 3 1 h. number of people in the home i. the number of parents in the home j. the city of residence Among the students where data existed, each had 3.28 (1.91) siblings, 56.4% collected welfare, 86% were bilingual. An average of 5.39 (2.11) people live in each household, and over 65% of them lived with one or fewer of their parents. O f utmost importance is the type of supervision that was conducted. First, of course, much depends on the probation officer himself, because one could reasonably surmise that the different probation officers employ different methods of supervising the CEIP students. There were five different probation officers, each of whom I represented with a different dummy variable. Aside from those indicia, which are represented above, I also collected data for each type of contact that the probation officer had with the child. This included: a. scheduled home visits: visits that the probation officer announced beforehand 131. For some reason, four CEIP-supervised students did not have this datum. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37 b. UNSCHEDULED home visits: visits that the probation officer conducted without informing the parents or probationer. These visits tend to give the probation officer and CEEP team a more realistic view of the student’s home life c. scheduled school visits: contacts consisting of scheduled appointments with either the teacher d. UNSCHEDULED school visits e. Office appointments f. Behavioral contacts: any time the probation officer was contacted as a result of negative behavior by the student g. Miscellaneous contacts (anything else) Any conclusions that can be drawn from the data will be very important from a policy standpoint, because if certain types o f contacts have a significant effect, then perhaps these should be used in the future, and ones that are not significant should be scrapped. Especially in this time where many municipal law enforcement agencies are facing budget shortfalls, any methods to make each officer more efficient comes greatly appreciated. Although there were over 450 students whom CEIP supervised, officers only kept extensive records on approximately 100 of them. For this subgroup (on whom we later run normal OLS regressions to obtain the effect of these contacts), probation officers conducted an average of 3.11 (1.89) scheduled home visitations (ranging from 0 to 12), Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 3.32 (2.49) unscheduled home visitations (ranging from 0 to 15), 2.83 (3.00) scheduled school contacts (ranging from 0 to 23), 2.82 (1.65) unscheduled school contacts (ranging from 0 to 7), 2.56 (1.85) office appointments (ranging from 0 to 13), 3.57 (4.40) behavioral episodes (ranging from 0 to 24), and 4.78 (3.97) miscellaneous contacts (ranging from 0 to 25). These standard deviations are nice, because they indicate that there was not just a set number of each type of supervision, which allows us to differentiate among them (as compared to if there had only been 3 of each type for each student, where we could not conclude anything about each type of contact’s effect). 7. Dummy variables Obviously, the first dummy variable that I generated was whether the student was on supervision during the time period. I labeled this dummy, CEEP. Second, I generated a dummy for whether the child had previously been supervised, called EXCEIP. This dummy is important, because it allows examination on whether those supervised individuals had improved attendance, behavioral, or grade data over time. Likewise, I generated dummy /CEIP, which is the number of time periods since the student has been supervised, e.g., if he had been supervised in tj, then t % would be 1, /9 would be 2, and so forth. This variable is very important, because it will allow us to differentiate CEIP’s effect over time more exactly that simply EXCEIP, which does not account for the possibility of changing effects over time. To complement /CEEP, I also added /CEIPsq, which equals /CEIP*/CEIP. This allows us to gauge nicely any diminishing return to the program. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 I also generated dummies for which school the student attended, called SCHOOL, which was able to vary by / and t. It is important to note at this juncture that because of the close proximity of the two schools, BGI and SUVA (i.e., the whole city o f Bell Gardens is only 2.5 square miles), some students changed schools, either between or within periods. For those who changed between periods, it was easy to follow, because SCHOOL would simply be different from one period to the next. However, for that small group o f students who changed within a period, I generated another dummy called CHANGE, because this group may have changed for different reasons and should be separated. There are also dummies for each individual probation officer and whether the student is in high school, HIGH. This latter variable will allow us to determine whether there was any effect in the high school that was different than at the intermediate schools. Likewise, dummies were also generated for each time period, to allow me to account for any trends that were in effect without regard to CEEP, and for each grade, for the same reason. Last, I generated two interesting dummies to account for the subgroup of the school population that was truant, TRUANT and fmrTRUANT. The intuition behind the first, TRUANT (student is currently defined as truant), should be clear, because it allows me to not only acknowledge that some truants were not supervised by CEEP, but also to measure what effect if any they contributed to each dependent variable. It also follows that fmrTRUANT (one if the student had ever been truant, otherwise zero) would allow Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 me to differentiate between those who were truant in the current time period, and those who had been so in the past. Both of these variables, when added to the model, allows more specificity in measuring CEIP’s effect on the respective dependent variables. B. Methods to Analyze CEIP’s Effects As stated earlier in this Thesis, I primarily desired to answer two different, but perhaps correlated, questions: whether CEIP had any individual effect on the students whom it supervised, and second, whether CEIP had any effect on BGI’s overall school attendance, grades, or student behavior. Identifying this second, what I shall call “school effect,” required me to control for any differences in attendance between the two schools that are not due to (even if correlated with) the presence of CEIP at the school. To control for this, I utilized the differences-in-differences (DD) estimation procedure by comparing the means of the attendance between BGI and SUVA for the 1995-2000 time period. Because we have a panel, I first ran “between group” regressions because they allow me to obtain a rough (or primary), but nevertheless, consistent estimate of the program’s and other individual effects. However, to take account of heterogeneity among individuals, we needed to consider a more complicated model—one that allows me to consider the changes over time. Second, I utilized regular ordinary least squares on the model with each of the respective independent variables in which I was interested, CEIP, EXCEIP, /CEIP, /CEIPsq, time dummies, grade dummies, FALL dummy, SCHOOL dummy, ethnicity Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 dummies, gender dummy, special education dummy, high school dummy, current truant dummy, and former truant dummy. After running these regressions, for each independent variable, total absences, truancy absences, average citizenship, and GPA, to do this consideration over time, I compared these OLS results on the same table with regressions using the fixed-effect model. This model assigns an individual-specific fixed effect that is not observable to the econometrician, but that allows us to use the data to estimate the program’s effect more efficiently. The caveat of this model, however, is that even if it shows a positive effect for the program, it does not allow us to conclude definitively that the program was the cause for this because it does not rule out exogenous factors that were time specific. For example, suppose that the city o f Bell Gardens had become much richer and employed greater police supervision in the exact time periods that CEIP was operational. Then the positive effect on attendance, etc., may have been a result of this police presence rather than CEIP. To remedy this potential pitfall, I finally employed the differences-in-differences (DD) model. This model takes account of naturally-occurring trends by controlling for the variations over time, and the heterogeneity among individuals, as well as between schools. For example, DD acknowledges the differences in the attendance rates between BGI and SUVA, but measures whether there was any broadening or narrowing of these differences when CEIP was introduced at BGI. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1. Choice o f Dependent and Independent Variables A brief discussion is warranted here regarding how I chose which variables to use as either my independent or dependent variables. I first wanted to look at three main observables: attendance, grades, and behavior. What I did, therefore, was create some nice sums, TOTabs (equal to the sum o f all the absences) and TRUabs (equal to the sum of all o f the absences that qualify as truancies, such as tardies). Thus, to measure the effects that I wanted to see, I used GPA, TRUabs, TOTabs, and AVECIT (classroom citizenship mark average) as my dependent variables. The main independent variables were, depending on whether fixed effect regression or not (because fixed puts all the individual characteristics in one), were the ethnicity vector, SCHOOL, HIGH, MALE, AGE, SPL_ED, and the respective CEIP variables, CEIP, EXCEIP, tCEIP, and tCEIPsq. O f course, I performed regressions while alternating between the use of EXCEIP and the tCEIP—tCEIPsq duo. To determine if endogeniety bias would occur, I obtained the correlation between these variables. Because a relatively high correlation existed between them, and I was unable to say definitively which would be the cause of the other, e.g., AVECIT and GPA—which one causes the other to improve, and some of the same people would be good in one may be good in the other, I chose not to include these in my independent variables. Thus, each of these dependent variables will not be included in the independent side of the equation, i.e., when running a regression on AVECIT, I did not Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 include any other dependent variable, GPA, TOTabs, TRUabs, as an independent variable in the same regression. 2. Between and Fixed Effect Regressions First, I ran the between group regressions, which is a regression model that provides consistent estimates of individual effects by measuring individualized deviations from their own over-time average. This model can be written as: i ; = A + i3 ,( X ,-X () + /32Z( +i/, where, in this case, Y stands for our observed depended variables, AVECIT, GPA, TOTabs, TRUabs, and Suspensions). Vector X stands for the time-varying characteristics, such as attendance, school, etc. Z is the vector consisting of the observed non-time-varying individual characteristics, such as gender and ethnicity. X does NOT vary by individual, unlike the fixed-effect model below. This is not that completely plausible in our case, because each person predictably has many characteristics that are different from others, but does allow us to have a primary assessment o f individual effects without considering potential changes over time. Because we wish to run regressions on the individual students to measure the effect of different observables that may affect the grades, attendance, or behavior over time, I chose to employ the following regression equation for fixed effects: r„ = p , x , t » , Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 In this equation, i represents the individual, and t represents the time period (t = 1,2, ..., 11). All of the other variables have previously been discussed in the above section, including X, which is a vector of the observable absence and other codes. We do not have Z (individual non-changing variables, e.g., gender) in this model because they are all included in the fixed effect, Aand each cannot be separately estimated. Obviously, uu is the error term. To reiterate, each individual has his own unobserved fixed effect, which is denoted by A /, and is taken to be constant over time t and specific to the individual. To compare the utility of running these fixed effect regressions, I placed their results side-by-side with the same model regressed using only OLS, in the same table. As shown below, the Fixed Effect regressions show us nicer results than the simple OLS. 3. Differences-in-Differences “Differences-in-diflferences strategies are simple panel-data methods applied to sets of group means in cases when certain groups are exposed to the causing variable o f interest and others are not. This approach, which is transparent and often at least superficially plausible, is well-suited to estimating the effect of sharp changes in the I37 economic environment or changes in government policy.” “ The differences-in- differences method is often used, as shown from the commonly-discussed examples of 133 measuring the effects of minimum wages on employment, the effect of immigration 132. 3 H a n d b o o k o f La b o r E c o n o m ic s 1296 (1999) (O . Ashenfelter and D. Card, eds.). 133. See, e.g., Lester, Richard A , Shortcomings or Marginal Analysis fo r Wage-Emprcyment Problems, 36 AM. ECON. REV. 63 (1946). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45 on the employment o f natives,1 3 4 effects of taxes on labor supply,1 3 5 and the effects of Medicaid on health.1 3 6 In our case, we want to measure the effect of the sharp policy change of the addition o f CEIP to BGI versus SUVA, which did not have this change. Because these two schools are virtually identical, measuring the difference in the differences in the respective rates should assist us to identify CEIP’s effect on attendance, grades, or behavior, if any. To accomplish this, I first included school-by-period effects between BGI and SUVA, by measuring the differences in the differences between the attendance, course grades, and (negative) behavioral rates of these two schools. That is, had I forewent comparing BGI’s rates with those of SUVA, I might have well concluded that because every year since CEIP was instituted BGI’s respective rates improved, that CEIP must be responsible for this improvement. This would have been a common mistake that many social scientists make, because it fails to recognize that the program (or whatever policy change that they happen to measure) may have had no effect at all, because the rates were improving notwithstanding the existence of the program. DD rectifies1 3 7 this because by examining the trends in attendance, behavior, and grades at SUVA (which did not have CEIP, and is essentially the same school as BGI in most other respects, e.g., 134. See, e.g., Card, David E., The Impact o f the Mariel Boatlift on the M iami Labor Market, 43 INDUSTRIAL AND LABOR RELATIONS REV. 245 (1990). 135. See, e.g., N. E issa, Taxation and Labor Supply o f Married Women: The Tax Reform A ct o f 1986 A s a Natural Experiment, WORKING PAPER, HARVARD UNIV., DEP’T OF ECONOMICS (1993). 136. See, e.g., A.S. Yelowitz, The M edicaid Notch, Labor Supply and Welfare Participation: Evidence fro m Eligibility Expansions, Working Paper, Mass. Inst Of Tech., Dep’t of Economics (1994). 137. For an excellent discussion on the pitfalls o f DD estimation, see Bruce D. Meyer, Natural and Quasi-Experiments in Economics, 13 J. BUS. & ECON. STAT. 151 (1995). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 ethnicity o f students, city (same), income, approximate number of students/year ( « b g = 1900, h s u v a — 1800)), and comparing those rates to BGFs rates, we hope to eliminate natural trends that were occurring regardless o f CEIP’s presence. Thus, the treatment group here is comprised only of BGI students, because CEIP was present primarily at BGI. They are identified in the data by the dummy SCHOOL,7=1. Conversely, my control group is SUVA students, who are identified by S C H O O L S . Last, by use of the dummy CEIP, I also briefly examined the experimental group of those individuals whom were supervised to examine the direct effect o f the program. As written supra, the data on some of these persons are much more extensive, which militates towards more conclusions based on what, if anything, of the type of supervision had an effect, including, inter alia, the probation officer, type of contacts, etc. C. Results In this section, the Thesis both explains and attempts to interpret the results from running the school difference-in-differences and the other regressions for the individuals. 1. Between-Group (Student) Regressions The following tables represents the results o fbetween-group estimations, with the first regression of each type done on a model that has both CEIP and EXCEIP and the second model is without them, i.e., regressing our Y vector (GPA, /«AVECIT, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 /wTOTabs, TRUabs) on CEBP, EXCEIP, SCHOOL, AGE, ethnicity vector, gender, and high school: Table 3: BETW EEN-G roup Coefficient E stim ations1 3 8 Model Variable GPA, GPAz AVECITi AVECIT2 TOTabSi TOTabsz TRUabsi TRUabS2 CEIP -1.293 (-5.778) n/a 0.2816 (2.385) n/a 7.4021 (0.405) n/a 4.0268 (0.717) n/a EXCEIP -0.265 (-6.792) n/a 0.0609 (2.940) n/a 3.7951 (1.494) n/a 1.8583 (2.383) n/a SCHOOL -0.191 (-13.998) -0.214 (-15.91) 0.0241 (3.356) 0.0293 (4.141) -0.0313 (-0.198) 0.2033 (1.300) 0.2471 (5.079) 0.3647 (7.561) AGE -0.090 (-13.596) -0.097 (-14.654) -0.0271 (-7.122) -0.2501 (-6.628) 0.8282 (19.830) 0.8534 (20.427) 0.2271 (17.710) 0.2397 (18.590) Special Ed -0.145 (-1.945) -0.188 (-2.524) 0.9269 (2.313) 0.1027 (2.568) 8.6340 (5.375) 9.7247 (6.057) 3.4938 (7.085) 4.0400 (8.152) Asian 0.390 (1.930) 0.3907 (1.924) -0.2285 (-2.155) -0.2286 (-2.154) -1.0309 (-0.467) -1.2302 (-0.556) -0.3604 (-0.532) -0.4606 (-0.674) Black -0.329 (-1.535) -0.3446 (-1.601) 0.1346 (1.198) 0.1386 (1.232) 4.9974 (2.128) 4.9759 (2.113) 2.3616 (3.276) 2.3505 (3.233) Hispanic 0.1307 (0.705) 0.1106 (0.593) -.1404 (-1.441) -0.1356 (-1.390) 0.4706 (0.236) 0.4697 (0.235) 0.0599 (0.098) 0.0591 (0.096) White 0.1186 (0.062) -0.0197 (-0.103) -0.0619 (-0.617) -0.5455 (-0.543) 1.9547 (0.953) 2.0806 (1.011) 0.3831 (0.608) 0.4459 (0.702) Male -0.3259 (-24.43) -0.3245 (-24.196) 0.2121 (30.166) 0.2118 (30.096) -0.2574 (-1.671) -0.2686 (-1.739) 0.2086 (4.412) 0.2030 (4.257) High School 0.0073 (0.213) 0.0354 (1.027) 0.1286 (6.147) 0.1188 (5.711) 18.9303 (44.301) 18.85127 (44.058) 3.5089 (26.749) 3.4689 (26.267) R2 0.1564 0.1465 0.1094 0.1075 0.3939 0.3899 0.2511 0.2380 At first glance, the results might suggest that the program (CEIP) has a negative effect on GPA, because the coefficient is negative and significant at 5% level. However, we cannot conclude anything about CEIP’s effect from this result, because the students in the program generally have a lower GPA. Thus, no causal relationship between GPA 138. /-values shown in parentheses. I did not include standard deviations, nor did I run as complex regressions here, because I simply wanted to get a rough estimate. For the more intricate results, please see the infra tables under OLS and FE regressions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and CEIP can be concluded from the between-group regression estimates. This is an example where an inappropriate model can give misleading results, because between- group regressions treat individuals equally by ignoring the individual differences and only using the individual’s over-time average. We can, however, draw conclusions from the observables that do not change over time, namely the gender and ethnicity. These results suggest that in general, Asian students have a higher GPA, and black students have a lower GPA, both when compared to the average. Likewise, for gender, the regression results indicate that males in general have a lower GPA than females. Additionally, we can conclude that on the average, students at BGI have a lower GPA than the students at SUVA. Similar conclusions can be drawn with respect to citizenship and attendance. 2. Ordinary Least Squares and Fixed Effect Regressions In order to confirm that the fixed effect (FE) regression use was correct, I ran both fixed effect and normal ordinary least squares (OLS) on the same model for /«(total absences), truancy absences, GPA, and /«(average citizenship). The following tables list the results from the fixed-effect regressions, where for each dependent variable, I alternated between including CEEP with either EXCEIP (model-1) or alternatively, the combination of fCEEP and /CEEPsq (model-2). In each table, I include both the regression coefficient and standard error, in addition to the t-value, in order that we can measure the significance of each variables coefficient. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 I placed the OLS and fixed-effect regressions side-by-side so that we could compare the two, especially the sign changes. Table 4: //i(Total Absences)m o d ei-i Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 0.3641574 0.0642956 5.664 -0.4653647 0.0634559 -7.334 EXCEIP 0.2136206 0.0289581 7.377 -0.7210545 0.0484491 -14.883 (CEIP _ _ — — _ _ (CEIP2 — — — — — — fi 1.444074 0.0273904 52.722 2.430822 0.1453955 16.719 fe (dropped) (dropped) fe 1.255126 0.0261071 48.076 2.197803 0.1164761 18.869 U 0.0895738 0.0267293 3.351 -0.0549229 0.0360004 -1.526 ts 1.258512 0.0250964 50.147 2.053892 0.0885594 23.192 U 0.184426 0.0264374 6.976 -0.1192332 0.0611867 -1.949 tr 1.298065 0.0243293 53.354 1.904331 0.0610735 31.181 U -0.5863648 0.0259628 -22.585 -1.17825 0.0881743 -13.363 fe 0.5633688 0.0235894 23.882 0.8794874 0.0347506 25.309 fio -0.60503 0.0258889 -23.37 -1.454088 0.1163138 -12.501 5th grade -2.084114 1.272005 -1.638 -2.243665 1.155354 -1.942 6™ grade -2.206365 1.272011 -1.735 -1.978031 1.154274 -1.714 7™ grade -2.169417 1.272018 -1.705 -1.542357 1.153875 -1.337 8™ grade -2.150544 1.272023 -1.691 -1.264281 1.154195 -1.095 9™ grade 0.6275171 0.0344162 18.233 -0.2742017 0.089809 -3.053 10 grade 0.4334966 0.0359376 12.062 -0.0803001 0.0628781 -1.277 11™ grade 0.3110169 0.0388104 8.014 0.0907042 0.0418528 2.167 FALL 1.097427 0.0260085 42.195 1.894098 0.1166604 16.236 SCHOOL -0.06858 0.0110264 -6.22 (dropped) Asian -0.8013967 0.1728514 -4.636 (dropped) Black 0.0962726 0.1853899 0.519 (dropped) Hispanic -0.2949021 0.1604295 -1.838 (dropped) White 0.0634258 0.1648571 0.385 (dropped) MALE -0.0654713 0.0107572 -6.086 (dropped) Special Ed 0.2956405 0.0505419 5.849 -0.0058661 0.0631976 -0.093 High School -3.166798 1.272328 -2.489 -1.099109 1.163178 -0.945 Current Truant 1.524004 0.0136339 111.78 0.8310926 0.0168513 49.319 Former Truant 0.5346215 0.0168509 31.727 0.092506 0.0202199 4.575 Intercept 3.626785 1.282166 2.829 2.361242 1.156505 2.042 Within R2 — 0.2396 Between R2 — 0.1385 Overall R2 0.2882 0.1387 This table highlights why we should run the fixed effect regressions, because o f the respective shifts in the signs of CEIP and EXCEIP, which as stated above, was a dummy that represents that the student had been supervised earlier. This model seems to show Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 that both while the student is supervised and afterward, he has reduced total absences, which makes sense when looking at CEIP’s functioning. Also nice are the time effects that are evidenced by the inclusion of our time dummies. Almost all of the time periods have significant coefficients, with different time periods having much more strong effect, both positive and negative, on the total attendance. Table 5: //i(Total Absences)modei- 2 Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 0.3643967 0.0642772 5.669 -0.3630984 0.060596 -5.992 EXCEIP — — — — — — (CEIP 0.1765426 0.0201255 8.772 -0.222927 0.0233365 -9.553 (CEIP2 -0.0256372 0.0031065 -8.253 0.0129179 0.002997 4.31 fi 1.43533 0.0274136 52.358 2.341416 0.145997 16.037 h (dropped) (dropped) t> 1.245381 0.0261587 47.609 2.117084 0.1170258 18.091 U 0.0882172 0.0267258 3.301 -0.0476565 0.0360355 -1.322 ts 1.246603 0.0251954 49.477 1.979917 0.0890659 22.23 fc 0.1813229 0.0264394 6.858 -0.1031282 0.0612745 -1.683 (r 1.283989 0.0244435 52.529 1.851751 0.0614901 30.115 (a -0.5900527 0.0259795 -22.712 -1.127519 0.0384612 -12.746 (9 0.55393 0.0236281 23.444 0.856221 0.0349514 24.498 (to -0.6011517 0.0259102 -23.201 -1.376828 0.1168111 -11.787 5™ grade -2.082611 1.271761 -1.638 -2.205582 1.155115 -1.909 6th grade -2.205035 1.271767 -1.734 -1.966357 1.154027 -1.704 7™ grade -2.169666 1.271774 -1.706 -1.548321 1.153628 -1.342 8™ grade -2.149972 1.27178 -1.691 -1.281532 1.153951 -1.111 9™ grade 0.6346115 0.0344595 18.416 -0.2124361 0.0901625 -2.356 10 grade 0.4444965 0.0359851 12.352 -0.0335421 0.0631645 -0.531 11th grade 0.3241551 0.0388519 8.343 0.1169004 0.0420212 2.782 FALL 1.088329 0.0260555 41.77 1.815435 0.1172105 15.489 SCHOOL -0.0653782 0.0110068 -5.94 (dropped) Aslan -0.8019437 0.1728183 -4.64 (dropped) Black 0.0951095 0.1853544 0.513 (dropped) Hispanic -0.2935132 0.1603989 -1.83 (dropped) White 0.0669289 0.1648244 0.406 (dropped) MALE -0.0655393 0.0107554 -6.094 (dropped) Special Ed 0.2941977 0.0505337 5.822 -0.0098334 0.0631909 -0.156 High School -3.168169 1.272084 -2.491 -1.181842 1.162992 -1.016 Current Truant 1.524296 0.0136257 111.87 0.8248378 0.0168583 48.928 Former Truant 0.535379 0.0168369 31.798 0.0809394 0.0202297 4.001 Intercept 3.631701 1.281921 2.833 2.413395 1.156288 2.087 Within R2 — 0.24 Between R2 — 0.145 Overall R2 0.2885 0.1471 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 This second model, which shows the program’s effect over time (with rCEEP), also seems to portend well for the program with respect to total absences. In other words, because the /-CEIP value is negative (and significant) together with its square being positive, we have a result that squares well with the supervision program, because the effect of being off the program does decrease your total absences, but with declining marginal effect. It is also interesting to note the effect that the different dummies show us. For example, the FALL dummy variable shows us something that seems very clear. Across all of the models, we are able to see that students seem to have more absences in the fall, at a high level of significance (/-value»5). Likewise, the grade (year in school, e.g., fifth grade, sixth grade,...) dummies were all significant, with the general trend that at the intermediate school grades, the total absences were always lower. As expected, former truants also had worse attendance, which militates towards implementing CEIP as early as possible, to ensure that the child never has to deal with all of the problems (present and future) that truancy can cause. Because of the difference in the attendance-taking between the high school and the intermediate schools, BGI and SUVA, I also ran the same above regressions, but limited them to the intermediate schools. The following table shows us that CEIP’s effect was consistent at the intermediate schools, as well as the pooled regression showed us that CEIP improved attendance throughout all grades. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 Table 6: /«(Total Absences)m 0 d ei-i (Intermediate Schools Only) Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 0.5560976 0.0672937 8.264 -0.3357494 0.0657231 -5.109 EXCEIP 0.4564838 0.0384553 11.871 -0.4982746 0.0570087 -8.74 (CEIP _ _ _ _ _ _ (CEIP2 — — — — — — (i 1.291646 0.0314263 41.101 0.6394401 0.7676644 0.833 fc 0.3031125 0.0315671 9.602 -0.2007431 0.6145131 -0.327 (3 0.9808047 0.0317368 30.904 0.5785269 0.6144797 0.941 U 0.3048928 0.0317668 9.598 -0.025936 0.4612843 -0.056 ts 0.8498365 0.0320267 26.535 0.6017478 0.4613548 1.304 te 0.3515198 0.0320785 10.958 0.1590175 0.308381 0.516 tr 0.8480058 0.031999 26.501 0.7084615 0.3084513 2.297 u 0.0301696 0.0314446 0.959 -0.070764 0.155827 -0.454 t9 0.6466174 0.0316178 20.451 0.5902272 0.1559514 3.785 5th grade -1.837769 1.323094 -1.389 -1.51735 1.209011 -1.255 6™ grade -1.963487 1.323104 -1.484 -1.656714 1.179396 -1.405 7th grade -1.923785 1.323113 -1.454 -1.595285 1.169359 -1.364 8th grade -1.905477 1.32312 -1.44 -1.615091 1.179369 -1.369 9™ grade (dropped) (dropped) 10 grade (dropped) (dropped) 11™ grade (dropped) (dropped) FALL 0.5976548 0.031447 19.005 0.5566995 0.0245758 22.652 SCHOOL -0.0927683 0.0139385 -6.656 (dropped) Asian -0.9885243 0.2307883 -4.283 (dropped) Black 0.0765366 0.245218 0.312 (dropped) Hispanic -0.3916171 0.2148849 -1.822 (dropped) White 0.0046634 0.2199914 0.021 (dropped) MALE -0.0371933 0.0135511 -2.745 (dropped) Special Ed 0.3099603 0.0526975 5.882 -0.0225389 0.0832879 -0.271 Current Truant 1.464553 0.0176786 82.844 0.7238001 0.0232663 31.109 Former Truant 0.5497198 0.0243032 22.619 -0.0259091 0.0305639 -0.848 Intercept 3.636161 1.340608 2.712 3.39202 1.200713 2.825 Within R2 — 0.0915 Between R2 — 0.2229 Overall R2 0.2075 0.1329 These results basically mirror the regressions for the whole panel, i.e., former CEEP- supervised students had fewer total absences, whether looking at them as a general pool of former or over time. This is good, because it lets us separate out attendance trends having nothing to do with the program, such as, for example, the schools’ taking more rigorous attendance, increased police force, etc.— factors that our model might miss. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 Table 7: /«(Total Absences)modei-2 (Intermediate Schools Only) Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 0.5531358 0.0673041 8.218 -0.3040168 0.0640321 -4.748 EXCEIP — — — — — — (CEIP 0.3457747 0.0365059 9.472 -0.2697368 0.0402956 -6.694 (CEIP2 -0.0540739 0.0076879 -7.034 0.0285468 0.0071067 4.017 ft 1.285962 0.031484 40.845 0.5416167 0.7681452 0.705 b 0.3033996 0.031595 9.603 -0.2846381 0.6149272 -0.463 b 0.9755116 0.0318198 30.657 0.4957151 0.6149067 0.806 U 0.3063611 0.0318094 9.631 -0.0950783 0.4616405 -0.206 ts 0.8443909 0.0321578 26.258 0.5352488 0.46171 1.159 U 0.3553268 0.0321091 11.066 0.1051991 0.3086366 0.341 b 0.8407907 0.0321101 26.185 0.6658202 0.3087013 2.157 U 0.026555 0.0314706 0.844 -0.0910461 0.1559467 -0.584 f o 0.6421651 0.0316406 20.296 0.5723921 0.1560589 3.668 5th grade -1.83682 1.323367 -1.388 -1.480737 1.209026 -1.225 6™ grade -1.961726 1.323376 -1.482 -1.642048 1.179385 -1.392 7th grade -1.923466 1.323386 -1.453 -1.596797 1.169345 -1.366 8™ grade -1.900656 1.323395 -1.436 -1.631927 1.179361 -1.384 9™ grade (dropped) (dropped) 10 grade (dropped) (dropped) 11th grade (dropped) (dropped) FALL 0.5933827 0.0314658 18.858 0.5576804 0.0245888 22.68 SCHOOL -0.0896547 0.0139263 -6.438 (dropped) Asian -0.9883224 0.2308361 -4.281 (dropped) Black 0.0779321 0.2452687 0.318 (dropped) Hispanic -0.3894095 0.2149291 -1.812 (dropped) White 0.0097329 0.2200362 0.044 (dropped) MALE -0.0378231 0.0135555 -2.79 (dropped) Special Ed 0.3119951 0.0527124 5.919 -0.018867 0.0832888 -0.227 Current Truant 1.467021 0.0176757 82.997 0.7212869 0.0232739 30.991 Former Truant 0.5531218 0.0242936 22.768 -0.0332044 0.0305764 -1.086 Intercept 3.635922 1.340885 2.712 3.431241 1.200764 2.858 Within R2 — 0.0916 Between R2 — 0.1985 Overall R2 0.2072 0.1251 Both of these intermediate-school-only regressions have the same results for total absences—that CEIP had a positive and significant effect on reducing total absences both during and after supervision, as shown by EXCEIP’s negative (and significant) coefficient and the /CEIP-fCEIPsq duo’s separate signs (in right direction). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Because the total absence measure is a rough measure, as discussed above in the data description section, I also obtained the sum of all of the absences that would technically fit under truancy, even if the school did not record them as official truancies—L (late), N (non-apportionment), T (truancy), and U (unverified). Thus, below I run the same regressions as above for TRUabs. Table 8: TRUANCY Absencesmodei-i Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 1.659636 0.5662414 2.931 -1.242081 0.6503336 -1.91 EXCEIP 1.218332 0.2550298 4.777 -1.198762 0.4965349 -2.414 (CEIP _ _ _ _ _ — (CEIP2 — — — — — — fi 5.602547 0.2412233 23.226 9.575886 1.4901 6.426 fe (dropped) (dropped) h 3.840599 0.2299207 16.704 7.986274 1.193716 6.69 U 0.456843 0.2354004 1.941 -0.2788916 0.3689539 -0.756 ts 4.019776 0.2210204 18.187 7.256316 0.907609 7.995 (6 0.5268754 0.2328302 2.263 -0.9360036 0.6270773 -1.493 tr 3.37639 0.2142645 15.758 5.975257 0.625917 9.546 (s -0.2212336 0.2286502 -0.968 -2.826461 0.9036627 -3.128 (9 2.07522 0.2077481 9.989 3.579136 0.3561446 10.05 tio -0.3252541 0.2279994 -1.427 -4.124629 1.192053 -3.46 5™ grade -4.543146 11.20235 -0.406 -2.814569 11.84075 -0.238 6™ grade -5.104624 11.2024 -0.456 -1.827955 11.82968 -0.155 7™ grade -4.773177 11.20246 -0.426 -0.1091796 11.8256 -0.009 8™ grade -4.357444 11.20251 -0.389 1.316398 11.82888 0.111 9™ grade 2.488518 0.3030977 8.21 -2.540163 0.9204157 -2.76 10 grade 0.6414403 0.3164971 2.027 -2.119796 0.6444119 -3.29 11th grade 0.4945622 0.3417972 1.447 -0.794195 0.4289329 -1.852 FALL 4.131979 0.2290532 18.039 7.582531 1.195606 6.342 SCHOOL 0.7350573 0.0971074 7.57 (dropped) Asian -4.085894 1.522275 -2.684 (dropped) Black 1.246846 1.632699 0.764 (dropped) Hispanic -3.622173 1.412877 -2.564 (dropped) White -0.9735393 1.45187 -0.671 (dropped) MALE 0.1404095 0.0947372 1.482 (dropped) Special Ed 2.304019 0.4451147 5.176 0.4752005 0.647687 0.734 High School -9.896506 11.20519 -0.883 2.416897 11.92094 0.203 Current Truant 18.2948 0.1200717 152.366 14.92984 0.1727025 86.448 Former Truant 1.798753 0.148403 12.121 0.4512636 0.2072252 2.178 Intercept 6.671486 11.29183 0.591 -2.648032 11.85255 -0.223 Within R2 — 0.2228 Between Rz — 0.3586 Overall R2 0.335 0.2825 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 The intuition behind the running a separate regression for truancy-type absences is clear, for two reasons: one, if there is a different sign between the two (e.g., total absences went up, but truancies down), then maybe the CEIP-supervised students only became wiser at “playing the game” and maybe shifted absences that are really truancies into a non-truancy absence, such as illness. Second, it allows us to measure exactly the type of absences that CEIP was designed to eliminate or reduce. Table 9: TRUANCY AbsenceSmodei- 2 Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 1.661557 0.5661129 2.935 -1.430244 0.6210574 -2.303 EXCEIP — — — — — — (CEIP 1.087341 0.1772524 6.134 -0.4098872 0.2391792 -1.714 ICEIP2 -0.1676337 0.0273603 -6.127 -0.0005238 0.0307171 -0.017 f a 5.537114 0.2414422 22.933 9.091852 1.496343 6.076 fe (dropped) (dropped) b 3.765441 0.2303895 16.344 7.567907 1.199413 6.31 U 0.4455042 0.2353843 1.893 -0.2122047 0.3693331 -0.575 ts 3.928921 0.2219053 17.705 6.909302 0.9128487 7.569 fe 0.504256 0.2328618 2.165 -0.7869533 0.6280105 -1.253 tr 3.274743 0.215283 15.211 5.727665 0.6302208 9.088 h -0.2404746 0.228811 -1.051 -2.540076 0.906651 -2.802 h 2.009962 0.2081014 9.659 3.460857 0.3582213 9.661 * 1 0 -0.2902186 0.228201 -1.272 -3.71297 1.197214 -3.101 5™ grade -4.52696 11.20088 -0.404 -2.662308 11.83893 -0.225 6™ grade -5.091383 11.20093 -0.455 -1.783419 11.82778 -0.151 7™ grade -4.768655 11.20099 -0.426 -0.1421276 11.82369 -0.012 8™ grade -4.343608 11.20104 -0.388 1.231585 11.827 0.104 9™grade 2.551757 0.3034975 8.408 -2.227813 0.9240884 -2.411 10 grade 0.7278353 0.3169338 2.296 -1.891785 0.6473817 -2.922 11™ grade 0.5907665 0.3421832 1.726 -0.6623708 0.4306815 -1.538 FALL 4.061212 0.2294806 17.697 7.157378 1.201307 5.958 SCHOOL 0.7617472 0.0969412 7.858 (dropped) Asian -4.088277 1.522075 -2.686 (dropped) Black 1.238315 1.632486 0.759 (dropped) Hispanic -3.608662 1.412693 -2.554 (dropped) White -0.9428151 1.45167 -0.649 (dropped) MALE 0.1393375 0.0947267 1.471 (dropped) Special Ed 2.296427 0.4450692 5.16 0.4450108 0.6476527 0.687 High School -9.90553 11.20372 -0.884 1.980269 11.91966 0.166 Current Truant 18.29904 0.1200062 152.484 14.90712 0.1727828 86.277 Former Truant 1.806613 0.1482892 12.183 0.4156263 0.2073371 2.005 Intercept 6.697616 11.29035 0.593 -2.333126 11.85096 -0.197 Within R2 — 0.223 Between R2 — 0.3694 Overall R2 0.3351 0.2872 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 As above, the sign changes between OLS and the fixed-effect regressions emphasize the importance of the fixed effects o f the individuals and their effects on the absences. Moreover, we again obtained positive and significant results of CEIP and EXCEIP on truancy-type absences. Unfortunately, with the standard errors, the effects may not be as noteworthy as they appear. Thus, I ran the model-2 regressions to see if some better effect can be uncovered, and this model did give us some decent results with the CEIP dummies—both during CEIP supervision and after seems to cause a decline in truancies for the student As always, however, perhaps the program has a limited effect, or other trends were on course that were not caught by the model, aspects that the DD estimation procedure may help to uncover. However, as in most of the other models, we get significantly large coefficients with high /-values for the grade dummies, which means that some things are happening in different grades, such as teachers more lenient at younger grades to fewer distractions at the intermediate. Again, to ascertain whether the difference in taking attendance perhaps had any impact in the results, or alternatively whether CEIP had much different results at the intermediate schools as compared to the high schools (and effects that could not be captured by the dummy variable HIGH), I choose to also run the regressions for truancy- type absences, but limiting them to the attendance while at the intermediate schools. The tables below represent the OLS and fixed-effect regressions for both models one and two. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 Table 10: TRUANCY Absencesmodei-i (Intermediate Schools Only) Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-vaiue CEIP 1.553341 0.5649682 2.749 -0.5183137 0.6433762 -0.806 EXCEIP 2.129576 0.3228534 6.596 0.8365553 0.5580694 1.499 ICEIP _ _ _ — _ _ ICEIP2 — — — — — — fa 5.878139 0.2638408 22.279 12.80412 7.514814 1.704 (2 -0.3124421 0.2650231 -1.179 5.625027 6.015586 0.935 fa 3.165679 0.2664483 11.881 9.468895 6.015259 1.574 fa 0.0948796 0.2666999 0.356 4.762007 4.515601 1.055 fa 3.200804 0.268882 11.904 7.915276 4.516291 1.753 fa 0.3174814 0.2693165 1.179 3.396926 3.018801 1.125 fa 3.185416 0.2686494 11.857 6.478213 3.019489 2.145 fa -0.0041015 0.2639948 -0.016 1.619401 1.52542 1.062 fa 2.739007 0.2654487 10.318 4.589805 1.526638 3.006 5th grade -4.499913 11.1081 -0.405 -3.881416 11.83524 -0.328 6th grade -4.985627 11.10819 -0.449 -2.385918 11.54533 -0.207 7th grade -4.7348 11.10826 -0.426 -0.3031544 11.44708 -0.026 8th grade -4.38743 11.10833 -0.395 1.652943 11.54507 0.143 9™ grade (dropped) (dropped) 10 grade (dropped) (dropped) 11th grade (dropped) (dropped) FALL 4.834964 0.2640148 18.313 4.573546 0.2405775 19.011 SCHOOL 0.8616282 0.1170212 7.363 (dropped) Asian -4.902736 1.937595 -2.53 (dropped) Black 0.4196642 2.05874 0.204 (dropped) Hispanic -3.624618 1.804078 -2.009 (dropped) White -0.3452173 1.846949 -0.187 (dropped) MALE -0.0374281 0.1137689 -0.329 (dropped) Special Ed 1.885449 0.4424243 4.262 -0.2355714 0.8153208 -0.289 Current Truant 20.26848 0.1484212 136.561 16.07906 0.227758 70.597 Former Truant 1.01218 0.2040392 4.961 -0.9675007 0.2991954 -3.234 Intercept 6.061411 11.25515 0.539 -2.663088 11.75401 -0.227 Within R2 — 0.2344 Between R2 — 0.3074 Overall R2 0.3632 0.2921 Both of these intermediate-school only truancy models gave us interesting results. Unlike the total absence regressions that we ran for the intermediate schools only, here the results conflict with the pooled sample regression. These two models indicate that CEBP actually had a negative effect on reducing truancy-type absences. However, this coefficient was not significant, when evaluating their respective 7-values. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 Table 11: TRUANCY Absencesmodei-2 (Intermediate Schools Only) Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 1.546532 0.5649539 2.737 -0.4241061 0.6268231 -0.677 EXCEIP — — — — — — (CEIP 1.643724 0.3064319 5.364 0.6325916 0.3944618 1.604 (CEIP2 -0.251055 0.0645326 -3.89 -0.0724316 0.0695688 -1.041 fi 5.854242 0.2642781 22.152 13.01279 7.519526 1.731 fe -0.3089371 0.2652094 -1.165 5.807611 6.019645 0.965 fe 3.143302 0.2670969 11.768 9.642374 6.019445 1.602 U 0.1021853 0.2670092 0.383 4.90645 4.519091 1.086 ts 3.174287 0.2699342 11.759 8.04183 4.519771 1.779 (6 0.333543 0.2695249 1.238 3.496483 3.021305 1.157 fr 3.148791 0.2695335 11.682 6.552825 3.021938 2.168 U -0.022678 0.2641655 -0.086 1.653961 1.526593 1.083 ta 2.71598 0.2655929 10.226 4.619795 1.527692 3.024 5™ grade -4.499143 11.1084 -0.405 -3.950748 11.83539 -0.334 6th grade -4.981813 11.10848 -0.448 -2.41431 11.54523 -0.209 7™ grade -4.740226 11.10856 -0.427 -0.3028216 11.44695 -0.026 8™ grade -4.375442 11.10864 -0.394 1.683869 11.545 0.146 9™ grade (dropped) (dropped) 10 grade (dropped) (dropped) 11™ grade (dropped) (dropped) FALL 4.815466 0.264125 18.232 4.570157 0.2407048 18.987 SCHOOL 0.8699274 0.1168976 7.442 (dropped) Asian -4.902867 1.937648 -2.53 (dropped) Black 0.4262973 2.058796 0.207 (dropped) Hispanic -3.619257 1.804124 -2.006 (dropped) White -0.3321971 1.846994 -0.18 (dropped) MALE -0.0391934 0.1137853 -0.344 (dropped) Special Ed 1.888963 0.4424701 4.269 -0.2386757 0.8153309 -0.293 Current Truant 20.27485 0.1483705 136.65 16.07859 0.2278327 70.572 Former Truant 1.021181 0.2039213 5.008 -0.9641205 0.2993184 -3.221 Intercept 6.072544 11.25545 0.54 -2.750104 11.75452 -0.234 Within R2 — 0.2344 Between R2 — 0.3031 Overall R2 0.3632 0.2897 Because CEIP also targets the classroom behavior as well as the truancies of the students whom it supervises, we now examine the average of the citizenship marks for all of the students in our panel. This is a very important measure, because, as mentioned above in the CEIP explanation section, CEIP is unique in truancy prevention because it not only brings the children back to school, but holds a hammer above them (the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 possibility o f truancy filing) to ensure that they do not commit acts that cause the school to remove them. Table 12: //{(Average Citizenship)m 0 d ei-i R egression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 0.053829 0.022193 2.425 -0.0128718 0.0235996 -0.545 EXCEIP 0.0247451 0.0066241 3.736 -0.0339377 0.018397 -1.845 (CEIP _ _ _ _ _ _ (CEIP2 — — — — — (i -0.4563572 0.2398745 -1.902 -0.1605285 0.2095505 -0.766 fe 0.0173302 0.0054655 3.171 0.107731 0.0287109 3.752 fa -0.5074349 0.2398612 -2.116 -0.2164409 0.2081418 -1.04 U 0.0117288 0.0051745 2.267 0.0855649 0.0217094 3.941 U -0.5260504 0.2398533 -2.193 -0.2497346 0.2070086 -1.206 U -0.0050555 0.0049742 -1.016 0.05576 0.014822 3.762 tr -0.5369372 0.2398323 -2.239 -0.2738936 0.2060816 -1.329 ta -0.0004044 0.0045331 -0.089 0.0306614 0.0081382 3.768 ( 9 -0.5421891 0.2398297 -2.261 -0.3090615 0.2054134 -1.505 5™ grad e -0.0119084 0.2397027 -0.05 -0.0785918 0.2367097 -0.332 6™ grad e -0.0985836 0.2397047 -0.411 -0.1153778 0.2363787 -0.488 7™ grade -0.0819506 0.2396968 -0.342 -0.0618133 0.2362456 -0.262 8™ grade -0.116703 0.2396981 -0.487 -0.063618 0.2363502 -0.269 9™ grade 0.2004767 0.007712 25.996 0.0286226 0.0220458 1.298 10 grade 0.1379193 0.008036 17.163 0.0315267 0.0153367 2.056 11™ grade 0.0539693 0.0086032 6.273 0.0092319 0.009837 0.938 FALL -0.5019219 0.2398354 -2.093 -0.3205683 0.2049964 -1.564 SCHOOL 0.0120537 0.0025385 4.748 (dropped) Asian -0.0846593 0.0363796 -2.327 (dropped) Black 0.0558327 0.0401946 1.389 (dropped) Hispanic -0.0540085 0.0336126 -1.607 (dropped) W hite -0.0617676 0.0347214 -1.779 (dropped) MALE 0.120086 0.0023776 50.507 (dropped) S pecial Ed 0.0356413 0.012119 2.941 -0.0132407 0.0160522 -0.825 High School -0.2022581 0.2397842 -0.844 0.0517013 0.2389914 0.216 Current Truant 0.1553271 0.0030496 50.933 0.0558015 0.0039977 13.958 Former Truant 0.0548686 0.003622 15.149 0.0058884 0.0047241 1.246 Intercept 0.7153723 0.3406906 2.1 0.467679 0.3140578 1.489 Within R2 — 0.0611 Betw een R2 — 0.0158 Overall R2 0.1601 0.0285 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 Table 13: //{(Average Citizenship)m 0 d ci - 2 Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP 0.0530589 0.0221937 2.391 0.0029183 0.0219195 0.133 EXCEIP — — — — — — (CEIP 0.0084635 0.0048696 1.738 -0.0104343 0.0078543 -1.328 (CEIP2 -0.0006732 0.0008221 -0.819 0.0017704 0.0009682 1.829 fi -0.4551778 0.2401554 -1.895 -0.1193573 0.2101559 -0.568 (z 0.0174978 0.0054743 3.196 0.1145564 0.0288346 3.973 fa -0.5061152 0.2401471 -2.108 -0.1763399 0.2087218 -0.845 U 0.0121407 0.0051916 2.339 0.0910127 0.0218281 4.17 ts -0.5245092 0.2401456 -2.184 -0.2110172 0.2075616 -1.017 ts -0.0045134 0.0049959 -0.903 0.0595728 0.0149252 3.991 tr -0.5354807 0.2401234 -2.23 -0.2366494 0.2065904 -1.146 fa -0.0002677 0.0045445 -0.059 0.032793 0.0081898 4.004 fe -0.540967 0.2401083 -2.253 -0.2737561 0.2058649 -1.33 5™ grade -0.0118119 0.2397174 -0.049 -0.0817692 0.2367133 -0.345 6™ grade -0.0985133 0.2397194 -0.411 -0.1169324 0.2363796 -0.495 7™ grade -0.0815753 0.2397115 -0.34 -0.0614484 0.2362455 -0.26 8™ grade -0.116437 0.2397129 -0.486 -0.0624773 0.2363508 -0.264 9™ grade 0.2004803 0.0077206 25.967 0.0238738 0.0221317 1.079 10 grade 0.1377935 0.0080468 17.124 0.0280632 0.0154063 1.822 11th grade 0.0538842 0.0086101 6.258 0.0072527 0.0098728 0.735 FALL -0.5007308 0.2401072 -2.085 -0.2870815 0.2054009 -1.398 SCHOOL 0.012417 0.0025343 4.9 (dropped) Asian -0.0847311 0.0363818 -2.329 (dropped) Black 0.0558155 0.040197 1.389 (dropped) Hispanic -0.0538335 0.0336147 -1.601 (dropped) White -0.061431 0.0347236 -1.769 (dropped) MALE 0.120089 0.0023778 50.503 (dropped) Special Ed 0.0360392 0.0121202 2.973 -0.0128637 0.0160542 -0.801 High School -0.2023364 0.2397989 -0.844 0.0585875 0.2390066 0.245 Current Truant 0.1555659 0.0030495 51.013 0.0560754 0.0039993 14.021 Former Truant 0.0551881 0.0036208 15.242 0.0061677 0.0047277 1.305 Intercept 0.7136089 0.3408983 2.093 0.4284801 0.3144047 1.363 Within R2 — 0.0611 Between R2 — 0.0153 Overall R2 0.16 0.0282 After comparing both the OLS and fixed-effect regressions for both models, the conclusions that can be drawn is that CEIP did also have a positive effect on the supervised students’ citizenship, especially after the program. This makes sense intuitively from at least two different perspectives. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 First, students cannot usually “turn over a new leaf’ overnight in their citizenship. In other words, unlike absences, which CEIP should directly effect right away (because it should be quite easy to ensure the child comes to school), a child may take longer to leam how to actually behave correctly in the classroom. Second, because o f some subjectivity with the citizenship grades, i.e., the teachers mark them down (and thus may take a little time to “get used to” the child’s improvement, there may also be a lag-time caused by the teacher as well as the student. Table 14: GPAm o d ei-i Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP -0.3074245 0.0629452 -4.884 0.0673754 0.0565867 1.191 EXCEIP -0.178945 0.0184951 -9.675 0.0849543 0.0441582 1.924 (CEIP ____ ____ ____ — ____ ____ (CEIP2 — — — — — — 6 0.6218822 0.6861464 0.906 -0.0909145 0.5045095 -0.18 (2 -0.0026228 0.0155999 -0.168 -0.1813618 0.0685251 -2.647 f e 0.7135909 0.6861098 1.04 -0.0388867 0.5011782 -0.078 £4 -0.0018713 0.014752 -0.127 -0.145307 0.0518132 -2.804 ts 0.7315471 0.6860875 1.066 0.0031574 0.4984974 0.006 U 0.0283366 0.014186 1.998 -0.0991191 0.0353771 -2.802 tr 0.7160378 0.686023 1.044 0.0297235 0.4963047 0.06 U 0.0401802 0.0129291 3.108 -0.0051405 0.019447 -0.264 ta 0.7585254 0.68602 1.106 0.1336592 0.4947227 0.27 5th grade 0.9024829 0.6856662 1.316 0.3124034 0.570109 0.548 6th grade 0.993394 0.6856714 1.449 0.3177756 0.569327 0.558 7th grade 0.7374277 0.6856552 1.076 -0.0169037 0.569019 -0.03 8th grade 0.8287729 0.6856588 1.209 -0.0152665 0.5692665 -0.027 9™ grade -0.4177 0.0219648 -19.017 0.0835675 0.0526446 1.587 10 grade -0.3534469 0.0228914 -15.44 -0.0653369 0.0366276 -1.784 11th grade -0.0965294 0.024513 -3.938 0.0044916 0.0235301 0.191 FALL 0.5829348 0.6860357 0.85 0.0808846 0.4937366 0.164 SCHOOL -0.10982 0.0070949 -15.479 (dropped) Aslan 0.3099376 0.103934 2.982 (dropped) Black -0.1926122 0.1144077 -1.684 (dropped) Hispanic 0.1363987 0.0961456 1.419 (dropped) White 0.1560999 0.0992143 1.573 (dropped) MALE -0.2997042 0.0067213 -44.59 (dropped) Special Ed -0.1607758 0.0337624 -4.762 -0.2282685 0.0375875 -6.073 High School 0.8429914 0.6859013 1.229 -0.5836772 0.5755163 -1.014 Current Truant -0.5756586 0.00859 -67.015 -0.2080239 0.0094854 -21.931 Former Truant -0.2132006 0.0102989 -20.701 -0.0284868 0.0112493 -2.532 Intercept 1.371743 0.9745419 1.408 2.664039 0.7562943 3.522 Within R2 — 0.2019 Between R2 — 0.1094 Overall R2 0.2486 0.1245 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 62 Table 15: GPAm odei-2 Regression: Variable Coefficient OLS Std Error t-value Coefficient FIXED EFFECT Std Error t-value CEIP -0.3051191 0.0629563 -4.847 0.0615805 0.0523326 1.177 EXCEIP — — — — — — (CEIP -0.1025942 0.0135454 -7.574 0.0342159 0.0186323 1.836 (CEIP2 0.013151 0.002294 5.733 -0.0018648 0.0023079 -0.808 (i 0.7259799 0.6870488 1.057 -0.0103496 0.5058639 -0.02 fa -0.0012429 0.0156285 -0.08 -0.1682861 0.068814 -2.446 fa 0.8175619 0.6870264 1.19 0.0401367 0.5024735 0.08 fa -0.000797*1 0.0148036 -0.054 -0.1339556 0.0520935 -2.571 fa 0.835375 0.6870209 1.216 0.0801804 0.4997316 0.16 fa 0.029513 S 0.0142499 2.071 -0.0903667 0.0356233 -2.537 fa 0.820478S 0.6869569 1.194 0.1036583 0.4974426 0.208 u 0.0426275 0.0129634 3.288 -0.0011504 0.0195702 -0.059 t9 0.8622248 0.6869144 1.255 0.2030482 0.4957188 0.41 5™ grade 0.9011901 0.6858075 1.314 0.3061505 0.5700739 0.537 6t h grade 0.9920113 0.6858126 1.446 0.3149388 0.5692854 0.553 7™ grade 0.735582 0.6857963 1.073 -0.016305 0.5689755 -0.029 8™ grade 0.8270663 0.6858001 1.206 -0.0132676 0.5692244 -0.023 9™grade -0.4196612 0.0219945 -19.08 0.0735129 0.0528465 1.391 10 grade -0.3562877 0.0229265 -15.54 -0.0726716 0.0367925 -1.975 11™ grade -0.0993171 0.0245365 -4.048 0.0005816 0.0236153 0.025 FALL 0.6851139 0.6869109 0.997 0.1469797 0.4946231 0.297 SCHOOL -0.1119295 0.0070834 -15.802 (dropped) Aslan 0.310341S 0.1039554 2.985 (dropped) Black -0.1922372 0.1144312 -1.68 (dropped) Hispanic 0.1353609 0.0961655 1.408 (dropped) White 0.1540641 0.0992351 1.553 (dropped) MALE -0.2996873 0.0067229 -44.577 (dropped) Special Ed -0.1617553 0.0337706 -4.79 -0.2269909 0.0375883 -6.039 High School 0.8430715 0.6860426 1.229 -0.5708666 0.5755093 -0.992 Current Truant -0.57*5306 0.0085908 -67.11 -0.2075299 0.0094882 -21.872 Former Truant -0.2145734 0.0102964 -20.84 -0.0273956 0.011256 -2.434 Intercept 1.271174 0.9752749 1.303 2.588721 0.7570296 3.42 Within R2 — 0.2021 Between R2 — 0.1102 Overall R2 0.2483 0.1254 As with both the total absences and model-1 of the average citizenship, we obtained (significant) positive GPA effects for the CEIP-supervised students. Again, the squared time term was negative, which indicates that the program had a diminishing return. Special education students had a significantly large negative effect on GPA. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 These results thus indicate that CEIP has a positive effect on GPA o f the supervised students, both at the same time that they were supervised and afterward. First, however, note that the program effect during supervision was not significant. This is easily explained by the fact that when assigning the time period for CEIP supervision, I had to use a rough estimate because most times, CEIP supervision did not begin exactly at the beginning of a period. Thus, CEIP supervision really existed across periods. But because accounting for this would have been prohibitively difficult, I simply put the person as supervised in the period in which CEIP supervision began (except when less than one month remained in a period, when I placed the supervised in the next period). As a result of this, looking at fCEIP=l may give a more accurate measure o f CEIP’s effectiveness, which is borne out by the data, which show that CEIP did have a positive effect over time. Likewise, a somewhat longer time amount may be necessary for grade improvement, because unlike attendance, which can be immediately measured, grades often take a period or two for any improvements to surface. Second, the GPA effect diminished over time, which we were able to detect by use o f the variable /CEIPsq, which was the square of the time elapsed since on the program (/CEIP). This result also is quite plausible, because like any supervision program, the more time away from being supervised, the less one would expect its effect to last, or be as salient. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 In almost all of the models, the school had a significant effect—even if it was not always positive, i.e., the school-effect for total absences was positive, but for truancies was negative, and both citizenship and GPA were negative. We thus shift to differences-in-differences to see whether any naturally-occurring trends can be removed so that we can obtain CEIP’s effect. 3. Differences-in-Differences Results This section shows the DD estimation of the effect of CEIP presence on school attendance, behavior, and grades for BGI. First, I list below, the period-by-period changes in the main three indicators: attendance, behavior, and grades, for each school. Also included is the interesting datum of the Hispanic enrollment percentage. Next, a figure is given that graphs the relationship between the period attendance rates, in order to afford the reader an opportunity to actually see the trends in attendance between the two schools, a trend that seems to be similar (when one was going up, so was the other) except for the introduction of CEIP. Then, the difference-in-difference estimation is obtained, and the tables used to do so are given. Last, this section discusses briefly some of the intuition behind the result in the DD estimates, including an elaboration on why CEIP had a stronger immediate school effect on attendance than on citizenship. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 Table 16: BGI p e r i o d - b y - p e r i o d a v e r a g e s 139 Period Variable ti t2 fa U ts b t7 ta b tio tii Hispanic 0.953 0.961 0.960 0.964 0.963 0.961 0.960 0.966 0.969 0.977 0.978 Attendance% 92.92 n/a 93.53 n/a 93.64 n/a 93.73 n/a 94.07 95.60 96.03 Citizenship 1.288 1.272 1.266 1.256 1.203 1.181 1.173 1.309 1.227 1.270 n/a GPA 2.489 2.458 2.542 2.391 2.483 2.395 2.514 2.538 2.636 2.503 n/a Illness 4.298 2.742 3.643 3.014 3.400 3.105 3.535 2.113 2.640 2.312 1.520 Tardies 1.212 0.736 0.888 0.499 0.591 0.479 0.786 1.118 1.523 1.286 0.776 Non-Appor 0.988 0.776 1.168 1.032 1.317 1.105 1.486 1.312 1.555 1.275 0.555 Unverified 0.003 0.000 0.001 0.000 0.000 0.000 0.000 0.000 0.005 0.001 0.084 S a t Makeup 0.049 0.008 0.032 0.009 0.011 0.002 0.148 0.080 0.193 0.097 0.069 Truancies 0.131 0.063 0.073 0.026 0.091 0.040 0.105 0.039 0.061 0.026 0.032 Off. Suspend. 0.147 0.058 0.121 0.093 0.144 0.107 0.106 0.058 0.081 0.063 0.049 Un. Suspend. 0.012 0.002 0.009 0.002 0.059 0.013 0.002 0.003 0.000 0.005 0.002 / Table 17: SUVA p e r i o d - b y - p e r i o d a v e r a g e s Period Variable ti b b b b ts t7 b b tio tn Hispanic 0.968 0.969 0.972 0.975 0.975 0.973 0.974 0.977 0.977 0.978 0.978 Attendance% 93.99 n/a 92.37 n/a 92.34 n/a 92.23 n/a 92.94 93.03 94.07 Citizenship 1.277 1.226 1.218 1.234 1.217 1.239 1.210 1.201 1.205 1.250 n/a GPA 2.801 2.739 2.815 2.738 2.831 2.787 2.842 2.802 2.826 2.729 n/a illness 4.655 3.526 4.143 3.314 3.768 3.257 3.413 2.542 3.055 2.327 1.519 Tardies 0.213 0.362 0.507 0.539 0.617 0.529 0.743 0.571 0.816 0.788 0.493 Non-Appor 1.010 0.728 0.652 0.697 0.579 0.763 0.694 0.559 0.535 0.629 0.330 Unverified 0.000 0.000 0.000 0.000 0.002 0.004 0.002 0.018 0.007 0.001 0.060 S a t Makeup 0.000 0.000 0.010 0.018 0.015 0.018 0.026 0.029 0.039 0.033 0.037 Truancies 0.128 0.067 0.121 0.062 0.190 0.052 0.059 0.023 0.034 0.026 0.017 Off. Suspend. 0.201 0.150 0.163 0.105 0.178 0.129 0.097 0.042 0.104 0.048 0.015 Un. Suspend. 0.027 0.039 0.017 0.003 0.000 0.001 0.004 0.004 0.006 0.006 0.003 In order to easily view the attendance trends, he following figure shows graphically, the attendance rate relationship for the schools over the time period 1995-2000. 139. First, note that the attendance data in these tables are in day units, not periods, because the comparison here is between the two intermediate schools, where there was no measurement problems. Note that the higher the Citizenship average, the worse the behavior, because 1 is best and 3 worst Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 1: BGI— SUVA Attendance Comparison 6 6 BGI v ersu s SUVA attendance trends Time Period (*6 month s each) As stated above, for differences-in-differences, we must first calculate the differences in the percentages of each effect that we want to explain by CEIP, namely the attendance, GPA, citizenship, and truancies: Table 18: DIFFERENCES in period-by-period averages (BGI - SUVA) Period Variable t i fa fa t« t s t s t7 fa fa tio t n Hispanic -0.015 -0.008 -0.012 -0.011 -0.012 -0.012 -0.014 -0.011 -0.008 -0.001 0 Attend. % -1.070 n/a 1.160 n/a 1.300 n/a 1.500 n/a 1.13 2.57 1.96 Citizenship 0.011 0.046 0.048 0.022 -0.014 -0.058 -0.037 0.108 0.022 0.02 n/a GPA -0.312 -0.281 -0.273 -0.347 -0.348 -0.392 -0.328 -0.264 -0.19 -0.226 n/a Illness -0.357 -0.784 -0.5 -0.3 -0.368 -0.152 0.122 -0.429 -0.415 -0.015 0.001 Tardies 0.999 0.374 0.381 -0.04 -0.026 -0.05 0.043 0.547 0.707 0.498 0.283 Truancies 0.003 -0.004 -0.048 -0.036 -0.099 -0.012 0.046 0.016 0.027 0 0.015 O. Suspend. -0.054 -0.092 -0.042 -0.012 -0.034 -0.022 0.009 0.016 -0.023 0.015 0.034 U. Suspend. -0.015 -0.037 -0.008 -0.001 0.059 0.012 -0.002 -0.001 -0.006 -0.001 -0.001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 Now, to complete the differences-in-differences model, we need to find the differences in these differences based on time. In other words, we want to see whether the addition of CEIP made any changes in the naturally-occurring differences in the attendance, GPA, and behavioral rates for BGI and SUVA. Concentrating on the differences in attendance, grades, citizenship (including both formal and informal suspensions), and illnesses, we obtain the following table that illustrates the DD estimation of CEIP’s effects on BGI. That is, each cell contains the difference between the cells in the foregoing table, e.g., for GPA, ft — 0.031 here represents the difference between 0.312 (which was the GPA difference between SUVA and BGI for ft) and 0.281 (which was the GPA difference between SUVA and BGI for ti). Table 19: Differences-in-Differences Estimates of CEIP’s impact Period Variable ti t2 t3 U ts ts t7 ts ts tio tn Attendance% - -1.070 2.230 0.140 0.200 -0.370 1.440 -0.610 Citizenship - 0 .035 0.002 -0.026 -0.036 -0.044 0.021 0.145 -0.086 -0.002 n/a GPA - 0.031 0.008 -0.074 -0.001 -0.044 0.064 0.064 0.074 -0.036 n/a Illness - -0.427 0.284 0.2 -0.068 0.216 0.274 -0.551 0.014 0.4 0.016 Truancies - -0.007 -0.044 0.012 -0.063 0.087 0.058 -0.03 0.011 -0.027 0.015 Off. Suspend. - -0.038 0.05 0.03 -0.022 0.012 0.031 0.007 -0.039 0.038 0.019 Un. Suspend. - -0 .022 0.029 0.007 0.06 -0.047 -0.014 0.001 -0.005 0.005 0 Although the program first officially began in operation at BGI in 1995, we take the difference in the differences estimation using the difference between tz and ft, because as explained above, the bulk of the supervision began between periods two and three. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 8 The DD estimates indicate that CEEP was successful at the school-wide level, especially in attendance. First, the attendance difference-in-differences-in-differences is 2.23 percent—this is the DD estimation of improvement in attendance that we can attribute to the CEIP. Moreover, formal truancies went down 4.4%, which makes sense when noticing that the main problem that CEIP was designed to combat is truancy. Unfortunately, according to the DD estimates, CEIP had almost no effect on citizenship, with it actually worsening by 0.2%. It did, however, seem to have a positive DD estimate in the next time period of 2.6%. As explained below, this may best be understood as an example of when one aspect of a program has an immediate effect and another has a lagged effect, i.e., taking more time to become realized. Last, BGI student GPAs increased 0.8%. 4. Supervision Results In this section, the thesis briefly discusses the results from running normal OLS regressions on the effects of certain types of supervision on attendance, citizenship, and GPA. First, the reader must note that because of the limited number o f observations for these data, the results are at best only a question to be posited. Among supervised probationers, the data show that unscheduled home calls had the most positive effect on attendance, while scheduled school calls had the most effect on citizenship and GPA. Unscheduled school contacts had limited effect, as did office visits. However, behavioral-based contacts did have a positive effect on total absences. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 69 Unlike the effects in the many tables above, unfortunately, none of these positive results was significant Last, the probation officer (as measured using their respective dummies) did not seem to have a significant effect on any of the dependent variables. D. Interpretation-Intuition The data appear to show that CEIP had positive effects both on a school-wide level for attendance and truancy reduction, and on the individuals whom it supervised for GPA and average citizenship. This is a very nice result, because it seems to indicate that CEIP is successful at the main aspects of school life that it seeks to impact—overall attendance, especially removal of truancies at a school-^vide level, yet with the individuals whom it supervises, it still is successful at making them behave, once it does get them back to school. 1. School Effect The improvements in behavior and attendance sho-wn by the data at the school-wide level are consistent with the way in which CEIP operates. In other words, CEIP deputy probation officers monitor the child by making both announced and unannounced home and school visits to ensure that the child comes to school. This accounts for the positive effect on attendance, because not only does the individual being supervised know that the probation officer watches him, but also do the other students know that if they step out of line, they could be next on the program, which presumably they do not want to happen. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 70 Likewise, individuals being supervised are cognizant that if they act out in class, that they are violating their conditions o f CEIP probation and could be subject to filing in juvenile court. As a result, their behavior in class often improved, as shown by the data. The small (almost negligible) positive effect on grades also makes sense, because although CEIP does offer referrals for school assistance (with such resources as after school tutoring), its main focus is on getting the child into school, and ensuring that he does not do anything to cause the teacher or school to have him removed. With this, the data show that CEIP was successful at the school-wide level. Moreover, it is much more difficult to monitor whether a child is performing adequately with his school work (i.e., his GPA), besides ensuring that he do it, as compared to making sure that he attends and behaves in class—which CEIP did have positive effects. Thus, the data are plausible when one understands exactly how CEIP works. 2. Individual Effect CEIP’s positive impact on the supervised individuals’ attendance (both total and truancy-type), GPA and citizenship reflects yet again the way in which CEIP operates. It is plausible that for a certain group o f students, no matter what is done, they simply will not come to school—thus helping to explain some of the data for CEIP’s less-than-stellar results for truancy in the intermediate-school only regressions. However, for those students on the fringe, i.e., would come to school if appropriately nudged, CEIP is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 successful at ensuring they do not misbehave once they do come back, with the threat of filing in court against them if they choose not to comply. Although the data militate toward finding that CEIP had no good effect on truancy absences at the intermediate school (shown by the positive coefficients on truancy), this fact can also be understood as a function o f the relative rigor of checking excuses in the intermediate versus the high school. That is, normally, it seems that illness excuses in the intermediate school (e.g., the child is sick and come with a note from “Mom”) are not checked as carefully as those in the high schools. Thus, many true truancies may have been masquerading as illnesses at the intermediate. But, because CEIP supervised these children, each excuse was checked much more carefully, and as a result, more real truancies were probably detected, thereby the slight increase in truancies. However, this is also a positive effect, because when looking at this result with the fact that total absences overall went down, we see that for the true truancies may have still have gone done in the aggregate, because they no longer could hide in illness, and the reduction in overall absences was greater than the coefficient for the increase in truancies. Thus, even the seemingly negative intermediate-only truancy results, when evaluated with the other data, appear not to harm CEIP’s success. CONCLUSION No simple solution exists to the interrelated problems of truancy and juvenile delinquency-crime. But, this Thesis’ evaluation o f the empirical attendance, grade, and behavioral data buttresses the conclusion that LACPD’s CEIP is more successful at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 72 fighting truancy than other programs, such as only holding the parents liable. Rather than simply bringing children who are already far behind back into school (where they may fail or force their way out), CEIP works to keep them there by letting them know that not only will it prosecute their parents, but it will also prosecute the juveniles themselves if they miss school, especially if they commit acts, such as fighting, that force the school to expel them. This is bome out by the improvement of average citizenship since CEIP’s beginning at BGI. Additionally, aside from the predictable result (bome out by the data) that the supervised individuals would have positive improvement in some of their measured variables, grades and behavior, the positive results at the school-wide level on attendance (especially truancy reduction) justify CEIP—maybe even more so, because at the school- wide level, more persons are impacted by the program. Last, the limited data on the supervised group indicate that unscheduled home and scheduled school visits had the greatest positive effect, with behavioral-based contacts coming in as next-best. Thus, this Thesis also recommends that these types of behavior should be used more extensively, and the less efficient actions should be discontinued, or reduced. Although arguably in its nascent stage, CEIP has shown such encouraging results in reducing truancy and improving school behavior that it should be expanded throughout Los Angeles County. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. BIBLIOGRAPHY 73 About Los Aneeles. Los Angeles County Office of Information. 11 Mar. 2000 <http://www.co.la.ca.us/overview.htm>. About the City of Bell Gardens. City of Bell Gardens. 11 Mar. 2000 <http://www.ci.bell- gardens.ca.us/abouLhtm>. Ashenfelter, O. and D. Card, eds. Handbook of Labor Economics. 3 vols. New York: Elsevier Science, 1999. Bilchik, Shay. “Breaking the Cycle of Juvenile Crime.” Trial 35-Jan. (1999): 36-41. Butts, Jeffrey A. The National Juvenile Court Data Archive: Collecting Data Since 1927. US Dept, of Justice. Washington: GPO, 1997. CA Education Code. Sec. 48200 (2000). CA Education Code. Sec. 48260 (1999). CA Education Code. Sec. 48262 (1993). CA Education Code. Sec. 48264.5 (1999). CA Government Code. Sec. 1031 (1995). CA Penal Code. Sec. 272 (1998). CA Vehicle Code. Sec 13202.7 (1999). CA Welfare and Institutions Code. Sec. 236 (1998). CA Welfare and Institutions Code. Sec. 283 (2000). CA Welfare and Institutions Code. Sec. 601 (1982). CA Welfare and Institutions Code. Sec. 601.5 (1999). CA Welfare and Institutions Code. Sec. 602 (1982). CA Welfare and Institutions Code. Sec. 650 (1998). CA Welfare and Institutions Code. Sec. 652 (2000). CA Welfare and Institutions Code. Sec. 653 (1998). CA Welfare and Institutions Code. Sec. 654 (1998). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 74 CA Welfare and Institutions Code. Sec. 654.6 (1999). CA Welfare and Institutions Code. Sec. 661 (1996). CA Welfare and Institutions Code. Sec. 729.3 (1998). California. Dept, of Education. School Attendance Review Boards Handbook. Operations and Resources. Sacramento: CA Office of Pubs., 1995. Card, David E. “The Impact of the Mariel Boatlift o f the Miami Labor Market.” Industrial and Labor Relations Review 43 (1990): 245-57. Cardenas, Maria. Personal Interview. 29 Mar. 1999. Chapin, Linda A. “Out of Control? The Uses and Abuses of Parental Liability Laws to Control Juvenile Delinquency in the United States.” Santa Clara Law Review 37 (1997): 621-72. Clements, Catherine. “Williams v. Garcetti: The Constitutionality of Holding Parents Criminally Liable for the Acts of Their Children.” Golden Gate University Law Review 25 (1995): 417-33. Cohn, Alvin W. “Juvenile Focus.” Federal Probation 61-Sept. (1997): 71-73. County of Los Angeles Statistical Data. Los Angeles County. 15 Feb. 2000 <http://www.co.la.ca.us/statistics.htm>. County Schools Facts. 1996-97. Los Angeles County Office o f Education. 10 Mar. 2000 <http://www.lacoe.edu/schools/facts.html>. D’Ambra, Laureen. “A Legal Response to Juvenile Crime: Why Waiver of Juvenile Offenders Is Not a Panacea.” Roger Williams University Law Review 2 (1997): 277-303. Dimitris, Jason Emilios. “Parental Responsibility Statutes — and the Programs that Must Accompany Them.” Stetson Law Review 27 (1997): 655-98. District Profile. Fiscal Year 96/97. Montebello Unified School District. 26 Feb. 1999 <http://www.ed-data.kl2.ca.us>. Duarte, Alex. Personal Interview. 28 Mar. 1999. Easy Access to FBI Arrest Statistics: 1991-1995. United States. Dept, of Justice. Natl. Inst, o f Justice, <http://www.ncjrs.org/ojjhome.htm>. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 Eissa, N. ‘Taxation and Labor Supply o f Married Women: The Tax Reform Act of 1986 As A Natural Experiment.” Working Paper, Harvard University Dept, of Economics, 1993. Elsea, Kelly Keimig. “The Juvenile Crime Debate: Rehabilitation, Punishment, or Prevention.” KS Journal of Law & Public Policy 5-Fall (1995): 135-44. Fee, Alison. “Forbidding States From Providing Essential Social Services to Illegal Immigrants: The Constitutionality o f Recent Federal Action.” Boston University Public Interest Law Journal 7 (1998): 93-115. Fritsch, Eric J., and Craig Hemmens. “An Assessment of Legislative Approaches to the Problem of Serious Juvenile Crime: A Case Study of Texas 1973-1995.” American Journal of Criminal Law 23 (1996): 563-609. Furtado, George E. “Juvenile Hearing Boards: Communities Respond to Juvenile Crime.” RI Bar Journal 44-Mav (1996): 17-20. Gaertner, Susan. “Three Strikes Against Juvenile Crime” Prevention, Intervention and Detention.” Prosecutor 30-Dec. (1996): 18-20. Garcia, Dan. Personal Interview. 15 Mar. 1999. Harris, Michael D. “Lawyer Charged Over Doctored UCLA Transcript.” Los Angeles Daily Journal 12 Feb. 1999: 1+. Harris, Michael D. “UCLA Student Pleads to Ruse to Get Hired.” Los Angeles Daily Journal 9 Feb. 1999: 1. Hurst, Hunter. “A Closer Look at Juvenile Violence?” Juvenile & Family Justice Today 44:1 (1993): 1-10. In re Abdirahman S. 68 Cal. App. 4th 963. CA App. Ct. 1997. John O. v. Superior Court. 169 Cal. App. 3d 823. CA App. Ct. 1985. Johnson, Marcia. “Juvenile Justice.” Whittier Law Review 17 (1996): 713-852. “Kid’s Truancy Puts Mother in County Jail.” Montebello News 5 Mar. 1998: 1. Le Blanc, Marc, and Rolf Loeber. “Developmental Criminology Updated.” Crime and Justice 23 (1998): 115-98. Lester, Richard A. “Shortcomings or Marginal Analysis for Wage-Employment Problems.” American Economic Review 36 (1946): 63-82. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Levitt, Steven D. “Juvenile Crime and Punishment.” Journal of Political Economy 106 (1998): 1156-85. Lichtenbaum, Jill A. “Juvenile Curfews: Protection or Regulation?” N Y Law School Journal of Human Rights 14(1998): 677-720. Los Angeles County Probation Department Home Page. Los Angeles County. Probation Dept 10 Mar. 2000 <http://probation.co.la.us>. Martin, Earl F., and Marsha Kline Pruett. “The Juvenile Sex Offender and the Juvenile Justice System.” American Criminal Law Review 35 (1998): 279-332. McCovery, Greg. Telephone Interview. 20 Mar. 1999. Menacker, Julius. “Legal Policy Affecting School Truancy.” Education Law Reporter 119(1997): 763-80. Meyer, Bruce D. “Natural and Quasi-Experiments in Economics.” Journal of Business and Economic Statistics 13 (1995): 151-61. Mitchell, Alison. “Clinton Likes Boston’s Gains Against Youth Crime.” Seattle Post- Intelligencer 20 Feb. 1997: A3. Molina, Gloria. “Law on Parental Responsibility.” Los Angeles Times 11 July 1989, Metro: 2. Montebello Unified School Dist. Finger Tip Facts. 1997-1998. Montebello: MUSD, 1998. Montebello Unified School District. Educational Measurement Office. Bell Gardens Intermediate 1992-1998. Montebello: MUSD, 1998. Montebello Unified School District. Educational Measurement Office. Montebello Unified School District 1992-1998. Montebello: MUSD, 1998. Mozingo, Joe. “D.A.’s Program Tries to Head Off Truancy Early.” Los Angeles Times 30 Aug. 1998: B1+. Panzer, Cheri. “Reducing Juvenile Recidivism Through Pre-Trial Diversion Programs: A Community’s Involvement.” Journal of Juvenile Law 19 (1997): 186-207. Parker, Jim, and Lucely Escamilla. The Condition of Public Education in Los Angeles County 1996. Los Angeles County. Office of Education, Informational Analysis Unit, 1997. Paul D. v. Superior Court. 158 Cal. App. 3d 838. CA App. Ct. 1984. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 Payne, James W. “Our Children’s Destiny.” Trial 35-Jan. (1999): 83-85. Pell, Charles Edward. “The Effects in California of the Legislative Cuts in Adult Education on High School Student Achievement and Satisfaction.” Thesis. Whittier Coll., 1996. People ex rel. Kottmeier v. Superior Court. 194 Cal. App. 3d 1536. CA App. Ct. 1987. People v. Adam R. 57 Cal. App. 4th 348. CA App. Ct. 1997. Quintero, Jaime. Personal Interview. 22 Mar. 1999. Rasmussen, Cecelia. “Community Profile: Bell Gardens.” Los Angeles Times 3 Jan. 1997, Metro: 1. Robins, Lee N. “What Determines Rates o f Homicide Over Time and Place: Can We Find Out?” University of CO Law Review 69 (1998): 1009-27. Santana, Melodie. Telephone Interview. 6 Jan. 1999. Sauerwein, Kristina, and Karima A Haynes. “A Dispatch From the Front Lines.” Los Angeles Times 15 Oct. 1999: B2. Scarola, Tami. “Creating Problems Rather Than Solving Them: Why Criminal Parental Responsibility Laws Do Not Fit Within Our Understanding of Justice.” Fordham Law Review 66 (1997): 1029-74. “Schools and Staffing Survey.” The Condition of Education 1996. Washington: GPO, 1996. Shah, Roshan. Personal Interview. 29 Nov. 1998. Shuit, Douglas P. “D.A. Steps Up Pressure on Parents o f Truants.” Los Angeles Times 5 Dec. 1997: Bl. Sickmund, Melissa. The Juvenile Delinquency Probation Caseload. 1985-1994. US Dept, of Justice. Office of Juvenile Just. & Delinquency Prevention. Washington: GPO, 1997. Sickmund, Melissa, Howard N. Snyder, and Eileen Poe-Yamagata, eds. Juvenile Offenders and Victims: 1997 Update o f Violence: Statistics Summary. US Dept, of Justice. Office of Juvenile Just. & Delinquency Prevention. Washington: GPO, 1997. Singer, Simon I. “Criminal and Teen Courts as Loosely Coupled Systems o f Juvenile Justice.” Wake Forest Law Review 33 (1998): 509-32. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 Snyder, Howard N. “Juvenile Arrests 1996.” Prevention Bulletin. US Dept, of Justice. Natl. Inst, of Justice. Washington: GPO, 1997. So, Michael. Personal Interview. 20 Mar. 1999. Springer, Larry. Vision 2000. Los Angeles County Office o f Education. Los Angeles: LA Co. Division of Community Schools, 1998. Stahl, Anne L., ed. Drug Offense Cases in Juvenile Court. 1986-1995. US Dept, of Justice. Fact Sheet 81. Washington: GPO, 1998. Steinruck, Stephonie. Personal Interview. 21 Oct. 1998. Tapia, Edward. Personal Interview. 5 Jan. 1999. Thompson, Ginger. “Gang Member’s Mother Denies Failure Charge.” Los Angeles Times 20 May 1989, Metro: 1+. Tonry, Michael. “Crime and Punishment in America.” Handbook of Crime & Punishment. Ed. Michael Tonry. New York: Oxford UP, 1998.3-27. Torbet, Patricia McFall. “Juvenile Probation: The Workhorse o f the Juvenile Justice System.” Juvenile Justice Bulletin. Mar. 1996. US Dept, of Justice. Torbet, Patricia McFall, ed. State Responses to Serious and Violent Juvenile Crime. US Dept of Justice. Washington: GPO, 1996. United States. Dept, of Justice. Manual to Combat Truancy. Washington: GPO, 1996. Velasquez, Eddie. CEIP Handbook. Montebello: MUSD, 1995. Velasquez, Eddie. Personal Interview. 2 Nov. 1998. Velasquez, Eddie. Telephone Interview. 10 Jan. 1999. Watson, Jane. “Crime and Juvenile Delinquency Prevention Policy: Time for Early Childhood Intervention.” Georgetown Journal on Fighting Poverty 2 (19951: 245- 58. Weinstein, Toni. “Visiting the Sins of the Child on the Parent: The Legality of Criminal Parental Liability Statutes.” Southern California Law Review 64 (1991): 859-901. Welkos, Robert W. “Mother Seized Under Gang law Cleared.” Los Angeles Times 10 June 1989, Metro: 1+. Williams v. Garcetti. 853 P.2d 507. CA Supreme Ct. 1993. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wirt, John, et al. The Condition of Education 1998. Washington: GPO, 1998. Yelowitz, A. S. “The Medicaid Notch, Labor Supply and Welfare Participation: Evidence from Eligibility Expansions.” Working Paper. MA Institute of Technology, Department of Economics, 1994. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A study concerning the relevance of using quality measures of education in economic growth models of sub-Saharan African nations

PDF

Articulated thoughts about intentions to commit anti-gay hate crimes

PDF

Effects of threat and self-focus on consensual bias in majorities

PDF

Effects of personal resource sufficiency on perceived difficulty and desirability of earthquake preparedness

PDF

An enquiry into the effectiveness of export assistance organizations serving southern California SMEs

PDF

Belize pegged exchange rate: Heading towards devaluation or maintaining credibility

PDF

A model of student performance in principles of macroeconomics

PDF

Depressed children and the social and behavioral attributes of their best friend

PDF

Debt reduction by way of inflation: The case of Lebanon

PDF

CalWorks in Los Angeles County: Two case studies of time limits and a discussion of marriage policy

PDF

Domestic violence in the South Asian immigrant community

PDF

Effects of a corneal anesthetic on the extinction of the classically conditioned response in the rabbit

PDF

Japan's modernization and troubled identity: Grappling with the West and other foreigners

PDF

Family functioning, intergenerational conflict, and psychological symptomology of Asian American college students

PDF

Content analysis of articulated thoughts of chronic worriers

PDF

Explanations, the availability heuristic, and biases of subjective likelihood in economics

PDF

Applying the theory of reasoned action to condom use: The effect of immediate consequences on intention-behavior consistency

PDF

Effects of bilateral stimulation and stimulus redundancy on performance in processing nonword letter trigrams

PDF

Ethnic identity, acculturation, self-esteem and perceived discrimination: A comparison study of Asian American adolescents

PDF

A typology of maritally violent men and correlates of violence

Asset Metadata

Creator Pell, Charles Edward (author)

Core Title Does intense probation monitoring of truants work? An empirical econometric analysis of its effect on school and individual attendance, grades, and behavior

School Graduate School

Degree Master of Arts

Degree Program Economics

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

Tag Economics, General,Education, Guidance and Counseling,OAI-PMH Harvest,sociology, criminology and penology

Language English

Contributor Digitized by ProQuest (provenance)

Advisor [illegible] (committee chair), [illegible] (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-34095

Unique identifier UC11341331

Identifier 1405220.pdf (filename),usctheses-c16-34095 (legacy record id)

Legacy Identifier 1405220.pdf

Dmrecord 34095

Document Type Thesis

Rights Pell, Charles Edward

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 au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA