University of Southern California Dissertations and Theses

A survey of the public: preference for old and new buildings, attitudes about historic preservation, and preservation-related engagement

(USC Thesis Other)

A survey of the public: preference for old and new buildings, attitudes about historic preservation, and preservation-related engagement

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

A SURVEY OF THE PUBLIC:
PREFERENCE FOR OLD AND NEW BUILDINGS, ATTITUDES ABOUT HISTORIC
PRESERVATION, AND PRESERVATION-RELATED ENGAGEMENT

by

Sandra Shannon

A Thesis Presented to the
FACULTY OF THE USC SCHOOL OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF HERITAGE CONSERVATION

December 2014

Copyright 2014 Sandra Shannon

ii

DEDICATION

For H.O., who is my mantra
and who would have been proud

iii

ACKNOWLEDGEMENTS

Although there is only one name on the title page, there is a village behind every thesis.
This study would never have occurred without the support of my chair, Trudi Sandmeier, who, in
addition to providing ever excellent feedback, was brave enough to endorse something outside
the box and trust that I could pull it off. I am further grateful to my committee members Jeremy
Wells, Ph.D. and Detlof von Winterfeldt Ph.D., who provided key guidance throughout the
process. In particular, Jeremy, who through his research experience and understanding of the
relevant literature, served as a beacon of thoughtfulness and rigor. And Detlof, whose statistical
expertise is bar none, made me confident in my methodology, data analysis, and interpretation of
findings.
In addition to the contributions from my committee, I am grateful for support I received
from several other scholars and practitioners, who, through their willingness to help, made this
study possible. Thank you to Thomas Herzog and Daniel Levi for providing advice and fielding
endless questions and concerns regarding the direction of the study and its design; to Katie
Horak, Gail Peter Borden, and Alice Valania, for serving as my expert panel during the image
selection process; and to Steven Kim for his statistical guidance, patience, and enthusiasm.
Thank you, also, to the photographers who agreed to the use of their images in the survey.
Lastly, I want to acknowledge the emotional pillars of my village. Thank you to my
family and friends, who have encouraged me along the way and kept me from scholastic
burnout, and to my mom and Luis, in particular, for their love and support always and during the
last two years.

iv

TABLE OF CONTENTS

DEDICATION ................................................................................................................................ ii
ACKNOWLEDGEMENTS ........................................................................................................... iii
LIST OF TABLES ......................................................................................................................... vi
LIST OF FIGURES ..................................................................................................................... viii
ABSTRACT ................................................................................................................................... ix
CHAPTER
1. INTRODUCTION ....................................................................................................................1
1.1 Impetus for the Study .........................................................................................................1
1.2 Research Questions ............................................................................................................3
1.3 Hypotheses .........................................................................................................................4
2. LITERATURE REVIEW .........................................................................................................5
2.1 Overview of Visual Preference Research ...........................................................................5
2.2 Research Related to Preference for Old versus New Buildings .........................................7
2.3 Research on Attitudes about Historic Preservation ..........................................................19
2.4 Research on Engagement in Historic Preservation ..........................................................23
2.5 Demographic Trends ........................................................................................................24
2.6 Associations among Preferences for Old Buildings, Preservation Attitudes, and
Engagement in Preservation ...................................................................................................28
2.7 Predictors of Preference in Visual Preference Research ..................................................29
2.8 Conducting Research using Amazon Mechanical Turk Participant Samples ..................31
2.9 Conclusion ........................................................................................................................35
3. METHODOLOGY .................................................................................................................36
3.1 The Survey Instrument .....................................................................................................36
3.1.1 Description of the Survey .........................................................................................36
3.1.2 Compiling a Pool of Stimuli .....................................................................................40
3.1.3 Selecting a Final Sample of Stimuli..........................................................................44
3.2 Survey Administration Methods.......................................................................................49
3.3 Statistical Methods ...........................................................................................................52
v

3.3.1 Primary Analysis .......................................................................................................53
3.3.1.1 Preference for Old versus New Buildings ......................................................53
3.3.1.2 Preservation Attitudes ....................................................................................53
3.3.1.3 Preservation Engagement ...............................................................................54
3.3.1.4 Demographic Trends ......................................................................................54
3.3.1.5 Pairwise Associations among Preference for Old Buildings, Preservation
Attitudes, and Engagement in Preservation ................................................................55
3.3.2 Secondary Analysis: Predictors of the Average Preference Rating of a Building ....56
4. FINDINGS .............................................................................................................................58
4.1 Descriptive Statistics ........................................................................................................58
4.2 Statistical Analysis ...........................................................................................................61
4.2.1 Findings for the Primary Research Questions ..........................................................61
4.2.1.1 Preference for Old versus New Buildings ......................................................61
4.2.1.2 Preservation Attitudes ....................................................................................70
4.2.1.3 Preservation Engagement ...............................................................................75
4.2.1.4 Demographic Trends ......................................................................................78
4.2.1.5 Pairwise Associations among Preference for Old Buildings, Preservation
Attitudes, and Engagement in Preservation ..............................................................92
4.2.2 Secondary Analysis: Predictors of the Average Preference Rating of a Building ....94
5. DISCUSSION ........................................................................................................................98
5.1 Major Findings .................................................................................................................98
5.2 Implications and Recommendations for the Field of Historic Preservation...................101
5.3 Recommendations for Future Research..........................................................................104
5.4 Final Conclusions ...........................................................................................................106
BIBLIOGRAPHY ....................................................................................................................107
APPENDIX A: Results of AIA America’s Favorite Architecture Survey ...............................112
APPENDIX B: Survey Instrument ...........................................................................................117
APPENDIX C: Mean Preference Ratings for Old and New Buildings, Ranked
from 1 to 50 ..............................................................................................................................180

vi

LIST OF TABLES

2.1 Summary of Dori Beth Frewald’s building dimensions and results ..................................11
2.2 Demographics of local preservationists and preservation leaders .....................................28
3.1 Criteria for the selection of photographic stimuli ..............................................................41
3.2 Number of stimuli with high / low ratings on the primary predictor variables .................47
3.3 Distribution of old and new building stimuli across the secondary variables ...................48
4.1 Descriptive statistics of participant demographics ............................................................59
4.2 Mean preference rating of pre-World War II, post-World War II, and new buildings
(7-point scale) ....................................................................................................................70
4.3 Mean ratings of community services (7-point scale) .........................................................71
4.4 Ranking of the importance of community service items ...................................................72
4.5 Mean agreement ratings for the nine preservation attitude items (7-point scale) ..............73
4.6 Percent of respondents who participated in each preservation activity in the last year .....76
4.7 Obstacles to getting involved in historic preservation efforts ...........................................78
4.8 Association between demographics and preference for old buildings ...............................80
4.9 Association between demographics and preference for new buildings .............................82
4.10 Association between demographics and relative preference (old – new) ..........................85
4.11 Unadjusted association between demographics and preservation attitudes .......................87
4.12 Adjusted association between demographics and preservation attitudes ..........................88
4.13 Association between demographics and preservation activities (method 1: linear
regression) ..........................................................................................................................90
4.14 Association between demographics preservation activities (method 2: logistic
regression) ..........................................................................................................................91
4.15 Association between preference for old buildings and preservation attitudes ...................92
4.16 Association between preference for old buildings and historic preservation activities
(method 1: linear regression) .............................................................................................93
4.17 Association between preservation attitudes and historic preservation activities (method 1:
linear regression) ................................................................................................................94
4.18 Unadjusted associations between primary / secondary variables and preference
(dependent variable: average preference rating) ................................................................95
vii

4.19 Multiple linear regression for adjusted association between primary / secondary variables
and preference (dependent variable: average preference rating) .......................................96
4.20 Association between preference and building age, modified by complexity ....................97

viii

LIST OF FIGURES

2.1 Percent of “America’s Favorite Architecture” by resource age .............................15
2.2 Number of old, mid-age, and new buildings in the top, middle, and lower 50 .....15
4.1 Mean preference ratings for old and new buildings...............................................61
4.2 Percent of participants who prefer old versus new buildings ................................62
4.3 Number of old and new buildings in the top 25 and bottom 25 ............................63
4.4 Number of old and new buildings in the top 10 and bottom 10 ............................63
4.5-4.10 Top 10 most preferred buildings ............................................................................64
4.11-4.14 Top 10 most preferred buildings ............................................................................65
4.15-4.20 Bottom 10 most preferred buildings .....................................................................66
4.21-4.24 Bottom 10 most preferred buildings .....................................................................67
4.25 Number of buildings per mean preference range ...................................................68
4.26 Boxplot of mean preference ratings: new buildings ..............................................69
4.27 Boxplot of mean preference ratings: old buildings ................................................69
4.28 Agreement with preservation attitude statements 1-3 ............................................74
4.29 Agreement with preservation attitude statements 4-6 ............................................74
4.30 Agreement with preservation attitude statements 7-9 ............................................75
4.31 Number of participants who engaged in 0, 1, or more preservation activities in the
last year .................................................................................................................77
4.32 Number of participants who engaged in 0, 1, or more preservation activities in the
last year (excluding visiting an historic site) .........................................................77
4.33 Participants’ rating for old and new buildings .......................................................84
4.34 Preference for old buildings and preservation-related attitudes ............................92
4.35 Distribution of complexity among old and new building sets ...............................97

ix

ABSTRACT

Very few quantitative studies exist to support and guide the practice of historic
preservation in the United States. Even the most basic concepts have yet to have been addressed
by scholarly researchers, such as if, how, and why people value the preservation of historic
resources. The purpose of this study is to begin to answer some initial research questions that
may help inform our practice, including: 1) Do people prefer old buildings or new? 2) What are
people’s attitudes about historic preservation? 3) To what extent do people engage in
preservation-related activities? 4) Are there demographic trends in preference, attitudes, and
engagement? The results of a survey of 200 U.S. residents demonstrate that, on average, people
prefer old buildings over new and they generally have favorable attitudes about historic
preservation. Although most people engage in at least one type of preservation activity a year,
preservation-minded preferences and attitudes do not frequently translate into more significant
participation in activities associated with historic preservation. Moderate demographic trends
were further observed in preference, attitudes, and engagement, with the demographic predictors
varying across the three areas of assessment. These results provide a better understanding of the
context within which preservationists work and can be used to guide advocacy, policy, and
practice. The findings further serve as a starting point for more nuanced research.
1

1. INTRODUCTION

1.1 Impetus for the Study
More than 60 years after the formal development of the field of historic preservation in
the United States, preservationists are still answering the question, “Why preserve?”
1
Over the
years, innumerable theorists have contributed answers to this question; most recently, in 2013
and 2014, the National Trust for Historic Preservation published a seven part series called, “Why
Do Old Places Matter?” where author Tom Mayes cited ideas ranging from the architectural
value of old places to the way in which the characteristics of old places can inspire creativity.
2

Although our theories and justifications are robust and varied, rather than support our cause with
empirical evidence, we often turn to anecdotal information. Though preservationists’
observations and experiences are certainly valid, if we could respond to the questions, “Why
preserve?” and, “Why do old places matter?” with scholarly evidence that shows that people do
value historic places and preservation policies – as preservationists suspect they do – that would
give our arguments much more weight and validity.
Unfortunately, empirical research about historic preservation is infrequently conducted
by the scientific community and we know little quantitatively about concepts that are at the heart
of our discipline, including why preservation matters. In the Getty Conservation Institute
publication, Values and Heritage Conservation, authors Erica C. Avrami, Randall Mason, and
Marta De la Torre, acknowledged this gap and mused that although the conservation of historic
resources plays an important role in society, “the benefits of cultural heritage have been taken as
a matter of faith.”
3
They argued that research that identifies “why cultural heritage is important to

1. “A Brief History of the National Trust,” National Trust for Historic Preservation, accessed August 17, 2014,
http://www.preservationnation.org/who-we-are/history.html.
2. “Why Do Old Places Matter: An Introduction,” Preservation Leadership Forum Blog, last modified November
13, 2013, http://blog.preservationleadershipforum.org/2013/11/13/old-places-introduction/#.U_RC-Xck3Nk; “Why
Do Old Places Matter: Architecture,” Preservation Leadership Forum Blog, last modified May 23, 2014,
http://blog.preservationleadershipforum.org/2014/05/23/old-places-matter-architecture/#.U_Q9Hnck3Nk; “Why Do
Old Places Matter: Creativity,” Preservation Leadership Forum Blog, last modified August 13, 2014,
http://blog.preservationleadershipforum.org/2014/08/13/old-places-creativity/#.U_Q9KXck3Nk.
3. Erica C. Avrami, Randall Mason, and Marta De la Torre, Values and Heritage Conservation: Research Report
(Los Angeles, CA: Getty Conservation Institute, 2000), 10.
2

human and social development and why conservation is seemingly a vital function in civil
society” would help us develop a framework to more effectively serve society.
4

This call to action was nearly 15 years ago, and unfortunately, the field has yet to respond
with significant scholarly investigation to answer the question of why we should preserve.
Beyond understanding the value of preservation, many additional research questions exist in the
field. For example: can preservation values be socialized? How broad is the public’s
understanding of the term “historic?” How do people feel about preservation-related policies?
There is great potential to use the results of questions like these to understand the social context
within which historic preservationists work, to inform policy and practice, and potentially,
support the cause.
Empirical data would also be useful as a preservation planning tool. The field of historic
preservation has often been criticized for being top down and reliant on the expertise of a select
number of trained individuals for decisions that affect communities. Open community forums are
part of the preservation planning process; however, participation is often minimal and the actions
taken as a result of these community conversations can be driven by the interests of a select
number of engaged citizens rather than the general population. The National Park Service has
taken a step in the right direction with the implementation of a public input requirement for
statewide historic preservation plans, and as will be discussed in the literature review, several
State Historic Preservation Offices have responded by conducting extensive surveys of their
citizens.
5
This trend toward being data informed is a positive one and if preservationists at the
local, state, and national level adopt the use of empirical evidence in their work, our practice
would be more valid, less hierarchical, and better able to respond to the needs of the
communities we serve.
Not only would preservation-related community research provide a progression in the
field, but also, it seems preservationists are ripe to embrace it. Although very little public opinion

4. A note regarding the use of “historic preservation” and “heritage conservation” in this report: whereas “historic
preservation” is the term used in the United States, “heritage conservation” is used in most other societies and is
generally considered to more broadly reflect the nature of the field. Although in the U.S., heritage professionals are
beginning to shift from the use of the words “historic preservation” to the more global, “heritage conservation,” the
term “historic preservation” was used in the survey conducted by the author as it would be more familiar to the
American public. For consistency then, “historic preservation” is the terminology used throughout this report except
where “heritage conservation” is quoted in text. Ibid.
5. National Park Service, Historic Preservation Fund Grants Manual (Washington D.C.: National Park Service,
2007).
3

research related to historic preservation exists, in recent years, studies regarding the
sustainability and economic value of preservation have been conducted and the preservation
community has widely accepted and promoted this data.
6
Preservationists have begun to augment
their theoretical arguments about why we should preserve with quantitative evidence from the
sustainability and economic movements. This speaks to the potential for additional empirical
information to enter the discourse and be supported and used by the community. It appears that
the time is right to begin a data driven dialogue.

1.2 Research Questions
Given the following: we (preservationists) are still defending our cause, we rely on the
community’s input in our work, and we lack basic empirical information as a field, but are
trending toward its use, the purpose of this thesis is to investigate fundamental concepts about
historic preservation that have yet to be sufficiently quantified. Specifically, the primary research
questions of this study are:

1) Do people prefer old buildings or new?
2) What are people’s attitudes about historic preservation?
3) To what extent do people engage in preservation-related activities?

Additional research questions include:

1) Are there demographic trends in preference, attitudes, and engagement?
2) Are there correlations between preference, attitudes, and engagement?
3) What are the predictors of preference?

6. See, for example: Preservation Green Lab, The Greenest Building: Quantifying the Environmental Value of
Building Reuse (Washington, D.C.: National Trust for Historic Preservation, 2011); Donovan Rypkema, Caroline
Cheong, and Randall Mason, Measuring Economic Impacts of Historic Preservation: A Report to the Advisory
Council on Historic Preservation (Washington, D.C.: PlaceEconomics, 2011); Preservation Green Lab, Older,
Smaller, Better: Measuring How the Character of Buildings and Blocks Influences Urban Vitality (Washington,
D.C.: National Trust for Historic Preservation, 2014).
4

These questions, which will be answered via a quantitative survey of the general public, are
intentionally broad and high level due to the limited amount of existing literature and it is my
hope that the results of the study can serve as a starting point for more nuanced research to come.

1.3 Hypotheses
Based on the existing literature, which will be presented in the next chapter, my
hypotheses are that, on average, people will prefer old buildings over new, and that most people
will express support for historic preservation. I anticipate that despite this, the number of people
who engage in historic preservation-related activities will be low. I expect to find some
demographic trends in preference, attitudes, and engagement; specifically, that participant age
will be a predictor of preference, political orientation a predictor of attitudes, and education and
urbanicity a predictor of engagement. Lastly, due to the extensive literature on the topics, I
hypothesize that building complexity, building condition, and the presence of landscape will
predict preference for old and new buildings, and that other secondary variables may be
moderately associated with preference, as well.

5

2. LITERATURE REVIEW

The following chapter will begin with an overview of visual preference research,
followed by a review of the existing literature of relevance to this study’s primary research
questions. These questions, again, are: 1) Do people prefer old buildings or new? 2) What are
people’s attitudes about historic preservation? 3) To what extent do people participate in historic
preservation-related activities? 4) Are there demographic trends in preference, attitudes, and
engagement? 5) Are there associations between preference, attitudes, and engagement? A review
of the literature relevant to the secondary research question – what are the predictors of
preference? – will also be provided. Lastly, since the study uses a relatively new sampling
methodology (Amazon Mechanical Turk), a review of the research on the reliability and validity
of conducting research on Mechanical Turk is provided.

2.1 Overview of Visual Preference Research
Visual preference research is an appraisal method in which subjective reactions to
environmental scenes are measured quantitatively in order to identify the types of environments
people most and least prefer, and investigate why people prefer certain environments over
others.
1
Preference research has been readily practiced since the 1970s by researchers in a variety
of disciplines, including environmental psychology, urban planning, urban design, architecture,
and landscape architecture, in order to determine the types of landscapes, street scenes,
buildings, interiors, etc. that people most and least prefer, and why. Since the majority of the
studies that have investigated preference for old and new buildings involve visual preference
methods and because the current study utilizes the methodology, an overview of the common
visual preference research practices are presented below.

Review of Common Methodological Practices
Although variations exist, the basic concept of visual preference research is to show
research participants a series of photographs of environmental scenes and have them indicate
their preference for each scene. Sometimes illustrations of scenes are provided as stimuli or in

1. Henry Sanoff, Visual Research Methods in Design (New York, NY: Van Nostrand Reinhold, 1991), 2.
6

situ preference evaluations are obtained, however, photographic representations are the most
commonly employed stimuli. A meta-analysis of studies comparing preference ratings for
photographic stimuli with ratings obtained in situ revealed a high degree of correlation,
indicating that photographs can be used in place of onsite evaluations (r=0.86, p<.001, 0.5
confidence interval of 0.8 < p < 0.9).
2

As Henry Sanoff elaborated upon in his overview of visual research methods, the process
by which stimuli are selected for visual preference studies varies significantly.
3
Some
researchers, for example, have selected images from internet sources or popular and professional
publications.
4
Others have taken their own photographs.
5
Still others utilize a more complex
selection procedure by which a pool of images is collected and a process occurs by which a final
subset of images is selected for the stimuli. This could, for example, be a random selection
process, a selection informed by the researcher or experts, a controlled selection, or a
combination of multiple methods.
6
As an example of an informed selection involving the
researcher and an expert, in her study, “The Socialization of Architecture Preference,” Margaret
A. Wilson, with the help of an architectural educator, selected images from the educator’s
personal architectural photograph collection that they felt represented a wide variety of recent
architectural trends.
7
As an example of an informed random selection, in “Visual Preferences in
Urban Street Scenes,” Jack L. Nasar randomly selected photographs from a pool of stimuli;
however, scenes that Nasar determined to be too visually similar to a previously selected scene
were replaced with another random selection.
8

As will be discussed in section 2.7, researchers have identified variables that predict
preference for scenes and as such, these factors are often controlled for in subsequent visual
preference research during the selection process. The method by which this occurs varies and

2. Arthur E. Stamps III, “Use of Photographs to Simulate Environments: A Meta Analysis,” Perceptual and Motor
Skills 71 (1990): 907-913.
3. Henry Sanoff, Visual Research Methods in Design, 2-4.
4. See, for example, Thomas Henschen and David B. Hershenson, “Values, Interests, and Architectural
Preferences,” Man-Environment Systems 5.4 (1975): 239-244.
5. See, for example, Thomas R. Herzog, “A Cognitive Analysis of Preference for Urban Spaces,” Journal of
Environmental Psychology 12 (1992): 237-248.
6. For an example of a random selection process see Jack L. Nasar, “Adult Viewers’ Preferences in Residential
Scenes: A Study of the Relationship of Environmental Attributes to Preference,” Environment and Behavior 15.5
(1983): 589-614.
7. Margaret A. Wilson, “The Socialization of Architectural Preference,” Journal of Environmental Psychology 16
(1996): 33-44.
8. Jack L. Nasar, “Visual Preferences in Urban Street Scenes: A Cross-Cultural Comparison between Japan and
the United States,” Journal of Cross-Cultural Psychology 15.1 (1984): 79-93.
7

there does not appear to be a standard protocol that researchers follow. To illustrate, in their
study, “The Prediction of Preference for Unfamiliar Urban Places,” Thomas R. Herzog, Stephen
Kaplan, and Rachel Kaplan had a group of participants rate 140 photographs for four variables
that have been found to predict preference. Using the calculated means, the researchers then
selected 70 scenes representing the full range of ratings across the four predictor variables. This
final set of stimuli was then shown to a second group of participants who indicated their
preference for each image.
9
In contrast, in his study, “Does History Matter?” Daniel Levi had a
panel of judges rate photographs for predictor variables (lighting, vegetation, cars, and people) to
ensure his groups of stimuli contained no significant differences in terms of these factors.

Levi
further included only historic buildings in his study that were well-maintained.
10
Although his
publication does not specify who made these maintenance determinations, it is presumed to be
Levi, indicating a bifurcated author-expert selection process.
Some researchers indicate in their methodological overview that they also control for
photographic variables during their image selection process. This could, for example, involve
standardizing the image orientation, camera angle, composition, position of the sun in
comparison to the scene, the time of day and year in which photographs are taken, etc.
11
There
does not appear to be consensus among the visual preference research community regarding
which conditions should be controlled and how they should be controlled. Further, although
controlling for photographic variables would likely be considered good practice by many, visual
preference researchers quite frequently do not provide details of this nature in their methodology
or do not control for photographic variables.

2.2 Research Related to Preference for Old versus New Buildings
Age-based preference research is somewhat limited, and in the author’s experience, since
it is often conducted by those outside the field of historic preservation and published in forums
that are not typically followed by preservationists, preservation professionals are frequently

9. Thomas R. Herzog, Stephen Kaplan, and Rachel Kaplan, “The Prediction of Preference for Unfamiliar Urban
Places,” Population and Environment 5.1 (1982): 43-59.
10. Daniel J. Levi, “Does History Matter? Perceptions and Attitudes Toward Fake Historic Architecture and
Historic Preservation,” Journal of Architectural and Planning Research 22:2 (2005): 148-159.
11. See for example, Thomas R. Herzog and Ronda L. Shier, “Complexity, Age, and Building Preference,”
Environment and Behavior 32.4 (2000): 557-575; Nasar, “Adult Viewers’ Preferences in Residential Scenes,” 589-
614.
8

unaware of the body of literature that exists. As such, a comprehensive review of the existing
research and findings of significance are presented in this section.
Thomas R. Herzog, Stephen Kaplan, and Rachel Kaplan were first to explore a
correlation between building age and preference. In their 1976 study, “The Prediction of
Preference for Familiar Urban Places,” they grouped photographs of buildings into content
domains, including a contemporary dimension, which included all new buildings; a commercial
dimension, which included 10 buildings, nine of which were older; and an entertainment
dimension, which the authors indicated included “mostly” older buildings / scenes such as movie
theaters, a drive-in theater, a nightlife establishment, etc.
12
The authors found that participants –
who had been screened for familiarity with the stimuli – preferred the contemporary dimension
(mean: 3.2) over the dimensions containing mostly older buildings (commercial mean: 2.3;
entertainment mean: 2.5) (F = 23.24, df = 4,92, p < .001).
13

A major limitation with the design of this study, of course, is that the authors were not
strictly comparing old and new buildings, but rather groupings of buildings that sometimes
contained stimuli from the opposite age dimension. Herzog, Kaplan, and Kaplan followed up this
study with one in which the content domains of older buildings and contemporary buildings
contained only old and new architecture respectively. Another difference with their follow-up
study was that whereas the participants in the previous study were familiar with most of the
stimuli, in this study the participants were instead unfamiliar with most of the stimuli.
Furthermore, and perhaps most importantly, whereas Herzog, Kaplan, and Kaplan did not
indicate that older buildings were well maintained in their seminal study, in the follow-up, the
older buildings were all in good condition. Under these circumstances, Herzog, Kaplan, and
Kaplan found that older buildings were preferred to newer buildings (older mean: 3.17;
contemporary mean: 2.74).
14

In their next study, “A Cognitive Analysis of Preference for Urban Nature,” Herzog,
Kaplan, and Kaplan again found a negative relationship between building age and preference.
Although the authors once more categorized stimuli by domains, including an older building

12. The authors did not indicate how many stimuli in the entertainment dimension showed older buildings and
how many showed newer. The authors also did not indicate how they defined new and old buildings. Thomas R.
Herzog, Stephen Kaplan, and Rachel Kaplan, "The Prediction of Preference for Familiar Urban Places,"
Environment and Behavior 8 (1976): 627-645.
13. Ibid.
14. Herzog, Kaplan, and Kaplan, “The Prediction of Preference for Unfamiliar Urban Places,” 43-59.
9

domain and a contemporary domain this time, condition was not controlled for and the authors
noted that many of the older buildings and their grounds were poorly maintained. Herzog,
Kaplan, and Kaplan found that although participants preferred the contemporary building domain
over the older building domain (means of 2.71 and 1.69 respectively), age was not a predictor of
preference. Instead, age had a strong negative correlation with the coherence, which was a
primary predictor of preference. The authors concluded then that “as far as preference is
concerned, the culprit is not age but upkeep and orderliness.”
15

In a follow-up study called, “A Cognitive Analysis of Preference for Urban Spaces,”
Herzog again found a negative relationship between age and preference. Here again, condition
was not controlled for.
16
However, in studies by Arthur E. Stamps III and Ronald Eugene
Widmar, in which similar methodologies were employed, a positive relationship between age
and preference was identified.
17
These studies, and those described above, indicated to
researchers that there was a relationship between age and preference, but the nature of that
relationship was unclear.
There are several limitations to these early studies that should be noted. To start, the
authors did not provide definitions for “old” and “new” buildings so it is unknown how old the
buildings were in each study, or whether the new buildings contained historicized elements or
represented contemporary architecture of their time. Furthermore, aside from the last Herzog
study, the number of stimuli in each of the old and new building domains in each survey was
rather low (ranging from one to 18 scenes). The variety of building types that were presented
within each domain, however, was generally high. These two factors make it difficult to draw
definitive conclusions about the data. Lastly, in “A Cognitive Analysis of Preference for Urban
Nature” and “A Cognitive Analysis of Preference for Urban Scenes,” the stimuli depicted scenes
in which the old or new building was not the focal point, but rather a part of a setting. In these
studies, participants may have been responding to extraneous information aside from the age of
the buildings, such as the composition of the photograph or the appearance of a parking lot

15. Thomas R. Herzog, "A Cognitive Analysis of Preference for Urban Nature," Journal of Environmental
Psychology 9 (1989): 27-43.
16. Thomas R. Herzog, "A Cognitive Analysis of Preference for Urban Spaces," Journal of Environmental
Psychology 12 (1992): 237-248.
17. Arthur E. Stamps III, "Formal and Nonformal Stimulus Factors in Environmental Preferences," Perceptual
and Motor Skills 79 (1994): 3-9; Ronald Eugene Widmar, "Preferences for Multiple-Family Housing in Small City
Neighborhoods" (Ph.D. diss., The University of Michigan, 1983).
10

shown in the foreground of the scene, for example. Despite these limitations, however, these
studies were an important foundation upon which future research would build.
In her study, “Preferences for Older Buildings: A Psychological Approach to
Architectural Design,” Dori Beth Frewald introduced a more considered methodological
approach which further developed the understanding of the correlation between building age and
preference. Frewald gathered preference ratings of 52 photographs of old and new university,
commercial, business, and professional buildings. All the buildings were in good condition and
Frewald further controlled the images for factors such as lighting and the presence of people,
signs, and garbage. The stimuli were categorized by dimension (described below in Table 2.1)
according to the age and features of each image’s focal building. The dimensions of old
buildings included classic textured, classic smooth, and modest old, and the dimensions of new
buildings included modern sleek, modern brick, and modest modern. As can be seen in Table
2.1, Frewald found that the old buildings were preferred overall to the new buildings. The
exception was with the modest old category, which was less preferred than modern sleek and
modern brick, but more preferred than modest modern. Frewald additionally found, however,
that the old buildings that were preferred were also the ones most likely to have visual richness
and ornamentation.
18
This leads to the question of whether participants were responding to the
age of buildings or to building features.

18. Dori Beth Frewald, “Preferences for Older Buildings: A Psychological Approach to Architectural Design”
(Ph.D. diss., University of Michigan-Ann Arbor, 1989).
11

Table 2.1: Summary of Dori Beth Frewald’s building dimensions and results
Dimension
Comprised of
Old or New
Buildings
Description
Mean preference
rating
(5 point scale)
Classic textured (n=6) Old Richardsonian Romanesque in style, with some
Renaissance influences; stone with rustic / textured
surfaces
4.04
Classic smooth (n=3) Old

Renaissance and classical styles; smooth facades
instead of textured
3.77
Modern sleek (n=3) New Walls of windows; light, airy 3.57
Modern brick (n=15) New Brick cladding; stark, sparse, little ornament 2.45
Modest old (n=5) Old Vernacular; largely unadorned 2.11
Modest modern (n=5) New Plain, drab, no ornamentation 1.40
From Dori Beth Frewald, “Preferences for Older Buildings: A Psychological Approach to Architectural Design” (Ph.D.
diss., University of Michigan-Ann Arbor, 1989), 290.

Following Frewald’s finding that when building maintenance is controlled and stimuli
carefully selected, there is a positive relationship between building age and preference, Thomas
R. Herzog then conducted two follow-up studies with Theresa A. Gale and Ronda L. Shier,
respectively, to determine if this finding could be replicated. In both studies, the stimuli, which
were now sizeable in number (n=60 and 64 respectively), were also more rigorously selected.
Factors such as image orientation, weather, seasonal cues, and the presence of people and signs
were controlled for, and building types appear to be more limited, at least in one study.
19
In both
studies, the researchers found a positive relationship between building age and preference when
building condition was controlled and a negative relationship when condition was not
controlled.
20
In their study, Herzog and Gale concluded, “Methodologically, this implies that
researchers investigating building age need to control building care, directly or indirectly, to
obtain valid results.”
21
The researchers also posited that the reason why people favor older
buildings is because of the visual richness often associated with older architecture.
22

If people prefer older buildings because they are more visually complex as Herzog and
Gale suggested, then how do new buildings that emulate the complexity and richness of old
buildings perform in comparison? In other words, as researcher Daniel J. Levi asks, “Does
history matter?” Levi found that, in fact, it does. He presented participants with three categories

19. In Herzog and Gale, building types included commercial, educational, industrial, cultural, governmental, and a
few residential structures, and in Herzog and Shier, the buildings were described simply as “urban.”
20. Thomas R. Herzog and Theresa A. Gale, “Preferences for Urban Buildings as a Function of Age and Nature
Context,” Environment and Behavior 28.1 (1996): 44-72; Herzog and Shier, “Complexity, Age, and Building
Preference,” 557-575.
21. Herzog and Gale, “Preferences for Urban Buildings as a Function of Age and Nature Context,” 66.
22. Ibid.
12

of stimuli: historic, fake historic, and contemporary. Historic buildings were in fact historic, fake
historic buildings were new buildings designed to look old or historic, and contemporary
buildings included Modern, Post Modern, and contemporary commercial styles. Building type
varied from hotels and offices to theaters and residences. Building condition was controlled for,
as was photographic composition and lighting, and the presence of vegetation, cars, and people.
Participants rated the beauty of each building (as opposed to how much they liked it, as in
previous research), and also indicated how historical or not historical they thought each building
was. Here again, with building maintenance controlled, participants favored historic buildings
the most (mean: 82.2.), followed by fake historic (74.3) and then contemporary (68.0). Not only
did participants prefer the real historic buildings over the fake historic buildings, but they were
also able to tell the difference between the two. The mean historic rating assigned to historic
buildings was 94.1 whereas the mean historic rating assigned to fake historic buildings was only
66.8.
23

The aforementioned studies by Herzog et al, Frewald, Stamps, and Levi best explain the
development of the relevant literature. Early studies identified a relationship between building
and preference but produced mixed results in terms of the direction of the relationship. Many of
these studies, however, had significant methodological limitations. Later studies, in which
methodological design was more rigorous and building condition controlled (in addition to other
predictive variables such as complexity and landscape), resulted in a positive relationship
between building age and preference. As will be discussed next, additional studies have been
conducted in which the methodological approach has varied, resulting in a more robust
understanding of the age-preference relationship.

Additional Studies of Preference for Old and New Buildings
Several relevant studies have deviated from the traditional visual preference methodology
of showing participants a series of stimuli one at a time and gathering a preference rating for
each image. Two such studies employed a technique in which residents were asked to identify
the areas of their city (Knoxville and Chattanooga, Tennessee in one study and Málaga, Spain in

23. Levi, “Does History Matter?” 148-159.
13

the other) they thought were the most and least visually attractive.
24
In both studies, researchers
determined that preference was found to be associated with places containing historical
significance. Additional preference attributes included places for leisure or walking, and places
with naturalness, panoramic views, upkeep, openness, and order.
25
In another study, by Richard
W. Berman, participants were shown 36 photographs of scenes depicting the city in which they
were in and asked to select the six photographs that depicted areas of the city they liked the most
or thought were most interesting. Participants showed a strong preference for intact older
buildings over International style buildings, historicized new buildings, and historic structures
that had been altered for a new use. Interestingly, when Berman repeated the study in Japan with
Japanese participants and scenes, the trend was reversed in favor of the International style
buildings and adaptively re-used buildings, indicating cultural differences in preference for
buildings based on building age.
26

Several more studies have deviated from the typical methodological approach used by
researchers studying preference for new and old buildings by more narrowly controlling for
building type. The studies described in this literature review have typically involved the
evaluation of a mixture of non-residential resources. That is, commercial, office, institutional,
entertainment, civic buildings, etc, are often presented to participants in the same study and
within content domains. Some researchers have instead taken a more focused approach in which
age-based preference is assessed for specific types of buildings. This has an advantage in that
controlling for building type lessens the possibility that participants respond to information about
the building use in their assessments. The results of these studies have been mixed. In his study
on preferences for new and old high rise buildings, Arthur E. Stamps III found that participants
preferred new high rise buildings with complex designs over pre-World War II brick high rises,

24. Specifically, in Jack L. Nasar’s Tennessee study, participants were asked to indicate the areas of the city they
visually liked and disliked. In M. P. Galindo and M. C. Hidalgo’s Spain study, participants were asked to identify
the most attractive and unattractive places in the city.
25. Jack L. Nasar, “The Evaluative Image of the City,” Journal of the American Planning Association 56.1
(1990): 41-54; M. P. Galindo and M. C. Hidalgo, “Aesthetic Preferences and the Attribution of Meaning:
Environmental Categorization Processes in the Evaluation of Urban Scenes,” International Journal of Psychology
40.1 (2005): 19-26.
26. Richard W. Berman, “Assessing Urban Design: Historical Ambience on the Waterfront” (Ph.D. diss.,
University of Pennsylvania, 1999).
14

but that they preferred pre-WWII high rises to new high rises that were plain in design.
27
In a
similar study, led again by Thomas R. Herzog, participants preferred older churches to newer
churches, including both those contemporary churches with striking architectural features and
those without. It is worth noting that a primary difference between the two studies was that
whereas building condition was not controlled for in the Stamps study, Herzog et al. did control
for this predictor variable.
28

Related to these studies by Stamps and Herzog et al. is a study by the American Institute
of Architects (AIA), conducted in conjunction with a professional research firm. Although the
study did not involve a specific building type in the common architectural or urban planning
definition of the word (such as “institutional buildings” or “residential buildings”), the stimuli
were all relatively well known works of architecture. In this regard, the study offers a typological
approach similar to Stamps and Herzog et al. “America’s Favorite Architecture Survey,” as the
AIA study was called, was conducted in 2006, and began with a random sample of 2,448 AIA
members who were asked to provide a list of 20 of their favorite built resources in the United
States. Works receiving six or more individual votes (N=247) were then presented to members of
the public to vote on via an online random sample.
29
Based on the results, the AIA compiled a
list of the top 150 resources, ranked in order from one to 150, which they deemed “America’s
Favorite Architecture.” Because the AIA removed the study’s website since the inception of this
research, a reproduced list of American’s Favorite Architecture is provided in Appendix A for
the reference of future researchers. As can be seen in the appendix, the resources included
buildings, as well as objects and structures such as the Vietnam Memorial and Brooklyn
Bridge.
30

Although the purpose of the AIA study was not to investigate a relationship between
preference and resource age, the data was available for this kind of analysis. A simple analysis

27. It does not appear, however, that building condition was controlled for in this study. Arthur E. Stamps III,
"Public Preferences for High Rise Buildings: Stylistic and Demographic Effects," Perceptual and Motor Skills 72
(1992): 839-844.
28. Thomas R. Herzog et al., “Preference and Tranquility for Houses of Worship,” Environment and Behavior
45.4 (2011): 504-525.
29. I was unable to obtain from the AIA additional information regarding the survey instructions and question
content that was presented to members of the public. Additionally, the AIA was unable to provide a list of the 247
resources for the purposes of this analysis.
30. “America’s Favorite Architecture,” The American Institute of Architects, accessed November 12, 2013,
http://www.favoritearchitecture.org/afa150.php; Phil Simon (Vice President, Strategic Communications and
Marketing, The American Institute of Architects), in discussion with the author, October 2013.
15

(conducted by the author) is provided below. The results suggest that in terms of well known
works of architecture, people prefer older resources. Fifty-nine percent of the top 150 resources
were old (more than 50 years old at the time of the study), whereas only 17% were new (less
than 15 years old at the time of the study). The remaining 25% were mid-age: that is, they were
constructed between 1990 and 1957. (Figure 2.1) Further analysis of the results by tercile
indicates that among the top 50 resources, only two were new works. The majority (86%) were
old. (Figure 2.2)

Figure 2.1: Percent of “America’s Favorite Architecture” by
resource age.

Figure 2.2: Number of old, mid-age, and new buildings in the top,
middle, and lower 50.

59%
25%
17%
Old (Pre 1956)
Mid-age (1957-1990)
New (1991-2006)
43
24
21
5
15
17
2
11
12
0
5
10
15
20
25
30
35
40
45
50
Top 50 51 to 100 101 to 150
Old (pre 1956)
Mid age (1957 - 1990)
New (1991 - 2006)
16

With the exception of the high rise study conducted by Stamps, each of the studies
described in this section found a positive relationship between preference and building age. This
is true for studies that offer general methodological designs other than traditional visual
preference methodologies, as well as for studies that investigate preference for specific types of
buildings. These findings, coupled with the findings in the post-Frewald literature provide a
strong case for the argument that people prefer older buildings over new. However, due to a
number of limitations, these studies leave room for further investigation.

Limitations
One of the primary limitations of the 16 studies described above is that 10 of them used
undergraduate student samples. Although Arthur E. Stamps III conducted a meta-analysis of
preference research and concluded that there is a strong correlation between the environmental
preferences of student participants and the general public, we do not know whether this holds
true when considering preferences for old and new buildings, specifically.
31
This is coupled with
the limitation that eight of the 10 student samples were from one of two universities in
Michigan’s lower peninsula. Furthermore, nearly all of the studies conducted with Michigan
college students utilized photographs of buildings in Michigan. Many times the buildings were
located in the same city in which the students attended college, making the stimuli not just
familiar to participants in terms of regional architectural characteristics, but personally familiar
to them as well. Based on these sampling conditions, we know less about what people in general
think about old and new buildings, but rather what undergraduate students from the lower
peninsula of Michigan think about old and new buildings located in the lower peninsula of
Michigan for which they may have pre-existing feelings or experiences. While these findings are
certainly useful, as will be described in section 2.5, demographic trends in architectural
preference do exist, and furthermore, the effect of familiarity on preference for old and new
buildings is inconclusive; as such, a more diverse body of literature would better address the
question of whether people prefer old buildings or new.
32

31. The Stamps analysis will be described in more detail in section 2.5. Arthur E. Stamps III, “Demographic
Effects in Environmental Aesthetics: A Meta-Analysis,” Journal of Planning Literature 14.2 (1999): 155-175.
32. For research related to familiarity and preference see: Herzog, Kaplan, and Kaplan, “The Prediction of
Preference for Familiar Urban Places,” 627-645; D. Canter and R. Thorne, “Attitudes to Housing: A Cross-cultural
Comparison,” Environment and Behavior 4 (1972): 3-23; Henschen and Hershenson, “Values, Interests, and
Architectural Preferences,” 239-244.
17

In addition to college students and the state of Michigan being over-represented in
samples and stimuli, many of the other studies simply do not indicate where the research
participants or buildings are from or what their participants’ demographics are. Many researchers
also provide very limited information about their stimuli selection process, as well as the
characteristics of the buildings that were assessed. Very few researchers, in particular, provide
definitions for what age ranges constitute new and old buildings. Given this lack of context, it is
difficult, therefore, to know what exactly was studied and how, which affects our ability to draw
conclusions.
As previously discussed, in most studies related to preference and building age,
researchers often have not controlled for building type and instead present many different types
of buildings, from schools and hospitals to office buildings and theaters. In many instances the
range of building types is too broad. A more conservative approach in which building type is
limited would allow us to be more certain in our findings because building type would be
removed as a possible predictor variable. This approach would require that numerous studies of
different types of buildings be conducted, but over time, the collection of data would tell the
story regarding preference for buildings in general and not just preference for old and new
churches, old and new high rises, etc.
A final limitation of the studies described above is that most were conducted in the 1980s
and 1990s. Since buildings are constantly aging and architectural trends changing, this affects
our ability to apply decades old research in practice. How can a preservationist today argue that,
based on the empirical evidence, people prefer older buildings over new, if the “new” buildings
to which they are referring were built in the 1970s and 1980s? It is possible that people always
prefer what is older in comparison to what is new at the time; however, it is also possible that
people of today prefer new buildings of today more than people of the 1980s, for example,
preferred the new buildings of their time. Either way, contemporary research is needed.

Attachment to Place
Whereas some researchers study preference for places, others seek to understand and
unravel the idea of how, why, and to what extent people feel an attachment to places. Although
not wholly synonymous, the two ideas share conceptual similarities and place attachment may
have the possibility for some of being indicative of preference. The results of relevant attachment
18

research can provide perspective for preference studies and vice versa. Although this section is
not intended to be an exhaustive review of the literature, one study is worth noting for its
assessment of attachment to old and new places, specifically.
In his study, “Attachment to the Physical Age of Urban Residential Neighborhoods,”
Jeremy Wells asked residents of two communities – one historic and one new urbanist – to
photograph places in their respective communities that were meaningful to them.
33
After
conducting interviews with the subjects about their photographs and what they found to be
meaningful, Wells determined that residents of the historic place had a greater attachment to
their community than residents of the new development. This attachment, he concluded, was due
not to the value participants placed on the historic buildings themselves, as one might expect
(although this was a small factor), but rather to the less tangible cognitive processes that occur
when experiencing an historic place; specifically, the unfolding of an old places’ mystery,
intrigue, and layers of age, and the imaginative thought that occurs about the past. When subjects
did speak to the architectural qualities of their communities, they often focused on specific
building elements such as doors, windows, balconies, shutters, etc., indicating an appreciation for
the detail and complexity of buildings. Wells’ findings were then replicated in a quantitative
study of residents of the same communities.
34

This study adds to the body of literature that supports the idea that old and historic places
are perceived and experienced by people differently than new places, and reminds us of the
narrowed lens with which researchers have thus far approached the topic. Preference is just one
aspect of feelings for a place and also just one aspect of historic preservation. One might prefer
the appearance of a proposed new building, but have a strong attachment to the one that already
exists and therefore potential replacement is resisted not for preference, but for attachment.
Although a simple example, it illustrates the importance of quantifying the various ways in
which people interact with old and new environments, but unfortunately, researchers thus far
have contributed very little outside of the age-based preference studies. Although the current
study admittedly stops short of a more holistic approach, the focus on preference will hopefully

33. For context, the new urbanist community was a new development (less than 15 years old) based on traditional
architecture and urban design principles. It was built to look and feel old in architecture and plan.
34. Jeremy Wells, “Attachment to the Physical Age of Urban Residential Neighborhoods: A Comparative Case
Study of Historic Charleston and I’on” (Ph.D. diss., Clemson University, 2009).
19

address some of the outstanding issues in the existing literature so that future research may move
forward into new and less investigated areas.

2.3 Research on Attitudes about Historic Preservation
Very little empirical evidence exists regarding what the general public thinks about
historic preservation or the extent to which they are supportive of it. Generally, this is not a topic
the academic research community has explored in the United States, and in fact, there is only one
scholarly study of relevance.
35
Most of the research that exists has been non-academic,
frequently having been conducted by preservation agencies in collaboration with market research
firms. Both academic and non-academic sources will be discussed below.
The academic work of relevance to the current study is “Does History Matter?” by Daniel
Levi. In this study, Levi found that participants were generally supportive of historic
preservation. When asked to indicate the importance of a series of eight ancillary city services,
historic preservation was rated as the second most important service behind street trees and
public landscaping and ahead of economic development, cultural events, recreational programs,
architectural review for the aesthetics of new construction, public art, and tourism promotion.
Eighty-eight percent of participants indicated it was an important to very important service.
Further, Levi found that slightly more than half of his participants (52%) believed that historic
buildings should be preserved, regardless of how beautiful they are. Lastly, participants were
generally in disagreement with the idea that historic preservation should be secondary to
economic development. Only 12% agreed that expanding economic development is more
important than preserving historic buildings and 23% agreed that preserving historic buildings
should not get in the way of economic development.
36

A limitation of Levi’s study is the use of undergraduate students as subjects. It is not
known if age and education have an associative relationship with preservation attitudes. As such,
a sample of the general public would be more appropriate. A further limitation of the research is
that the study involved residents of San Luis Obispo, a community with a lively and intact

35. Academic research of relevance has been conducted outside the United States, however, since preservation
attitudes are likely to be influenced by culture I have chosen to focus only on studies of U.S. participants.
Recommended readings (from the Netherlands and England) include: J. F. Coeterier, “Lay People’s Evaluation of
Historic Sites,” Landscape and Urban Planning 59 (2002): 111-123; R. Bishop, “The Perception and Importance of
Time in Architecture,” (Ph.D. diss., University of Surrey, England, 1982).
36. Levi, “Does History Matter?” 148-159.
20

historic downtown. It is possible that because participants lived in a charming and bustling
historic place, they would be more likely to appreciate historic preservation and to see it as
something that does not interfere with economic vitality. Possible evidence of this is the finding
that participants who had lived in the community longer were less likely to agree with the
statement, “Preserving historic buildings should not get in the way of economic development.”
Although Levi’s study provides important information regarding the extent to which people
value historic preservation, given these two limitations, an iteration of the study, with
participants from the general public who live in a variety of locations, would broaden our
understanding of preservation-related attitudes in the U.S.
Regarding non-academic sources of preservation attitudes, because the National Park
Service requires that public input be gathered as part of the development of state historic
preservation plans, a number of quantitative and qualitative studies have been conducted by State
Historic Preservation Offices (SHPOs).
37
Most of these studies have used samples of people who
were already engaged in preservation, and they are, therefore, irrelevant to the current study.
Two studies, by the Hawai’i and Arizona SHPOs, however, were comprehensive attitudinal
surveys of the general public. In both studies, support for historic preservation was generally
strong.
The study from the Hawai’i SHPO, which was administered in 2012 by a market
research firm, involved a random sample of 812 adult participants, most of whom (84%) were
uninvolved in preservation (either as opponents or proponents).
38
The survey touched upon a
number of topics of relevance to the current study including: general support for preservation,
impression of community commitment to preservation, preference for historic buildings,
preservation and economic development, preservation regulation, and preservation funding.
Nearly all the participants indicated that preservation of the state’s history, culture, and
architecture was important to them (98% agreed to strongly agreed) and 81% felt that historic
places were valuable contributions to a desirable community.
39
When asked if they would rather
work in an historic building or a new building, twice as many participants indicated that they
would prefer an historic building over a new building (42% versus 19%). A series of items

37. National Park Service, Historic Preservation Fund Grants Manual.
38. Responses were collected online (N=712) and via telephone (N=100).
39. Statistics presented for the Hawai’i and Arizona studies reflect the percent of participants who agreed to
strongly agreed (or vice versa where relevant) with each survey item.
21

related to preservation and development revealed that participants generally felt that preservation
did not impede development, that preservation was generally more important than development,
and that it actually contributed to the economic well-being of communities. Further, participants
strongly believed that with proper planning, both development and preservation goals could be
achieved (89% agreed to strongly agreed). In terms of property rights and regulation, a majority
of respondents (60%) felt that owners of historic resources should not be able to do whatever
they please with their property. Lastly, most participants (67%) supported the idea that as a
public benefit, the government should provide financial support for preservation. Despite this
strong evidence of support for historic preservation in Hawai’i, only 44% of participants felt
their community was committed to preservation, indicating incongruity between the values of the
community and what may be happening in practice.
40

The Arizona study was conducted by the SHPO in collaboration with researchers from
Arizona State University and a market research firm. The landline telephone survey involved a
random sample of 600 adult residents of Arizona, stratified by gender and region. Several of the
same topics that were investigated in the Hawai’i study were additionally investigated in this
study, including: general support for preservation, preservation and economic development,
regulation, and preservation funding. Additionally, and of relevance to the current study, the
Arizona study investigated the idea that preservation is environmentally friendly.
In general, it appears that although the Arizona participants saw value in historic
preservation, their attitudes were not nearly as positive or encompassing as were the attitudes of
the Hawai’i participants. Seventy-one percent of Arizona’s participants believed that historic
preservation helps make a better future. In both the Hawai’i and Arizona studies, approximately
16 to 18% of participants felt that preservation gets in the way of growth and development.
Nearly a quarter of Arizona participants further indicated that preservation actually prevents
change. Although 65% of Arizona respondents indicated that the government should be involved
in historic preservation, only 55% agreed that the government should provide financial
incentives to owners of historic resources. Though phrased somewhat differently, this is
considerably lower than the 67% of Hawai’i residents who supported government funding of

40. State of Hawai’i Historic Preservation Division, Hawai’i State Historic Preservation Plan: October 2012 to
October 2017 (Hilo, HI: State Historic Preservation Division, 2012).
22

historic preservation. Lastly, just over half of participants (55%) saw compatibility between
preservation and recycling / sustainability.
41

A limitation of both SHPO studies was that the participant samples skewed significantly
older than the general population, and in Hawai’i, the sample also skewed female. As previously
mentioned, given the lack of research regarding an association between demographic
characteristics and preservation attitudes, a stratified sample that better represents the general
population would offer more conclusive results.
Although not as comprehensive as the Hawai’i and Arizona studies, the State Historic
Preservation Office of Ohio, in collaboration with the University of Cincinnati Institute for
Policy Research conducted a simple survey of Ohio residents. The study was administered in
1999 and again in 2001 to random samples of just over 800 participants.
42
In both years of the
administration, 68% of participants indicated that, “Most of the historic buildings in my
community are an asset and should be preserved whenever possible.” Only 4% felt historic
buildings were obstructions to progress that should not be preserved and 28% had no opinion.
Those who reported not having historic buildings in their community or not knowing if there
were historic buildings in their neighborhood were more likely to indicate they did not have an
opinion.
When the results of this item are compared to results of similar items from the Hawai’i
and Arizona studies (“Historic places are valuable contributions to a desirable community,” and
“Historic preservation helps make a better future”), it appears that residents of Ohio are perhaps
less supportive of preservation than residents of Hawai’i and just about as supportive as residents
of Arizona. Although support varies somewhat from state to state, the majority of respondents to
the SHPO studies appear to see a general community value in historic preservation.
Whereas the results of the community asset item seem to reflect general support for
preservation by residents of Ohio, or at worst, indifference to preservation, when asked, “Would
you like to live in an historic home or does that sound like too much trouble?” the result was
quite different. Here, just over half of the participants (58% in 1999, 55% in 2001) indicated that
they would not want to live in an historic residence because it sounds like too much trouble. Just

41. Megha Budruk, Kathleen Andereck, and Gautam Prateek, 2013-2014 Arizona State Historic Preservation
Study: Final Technical Report (Draft) (Phoenix, AZ: Arizona State Parks Board, 2014).
42. Methodological conditions beyond this are unknown.
23

over a third would want to live somewhere historic (37% in 1999, 39% in 2001).
43
Although the
question text, response options, and scale differ between the two items, it is nonetheless
interesting to compare the results of this question with the work preference item from the
Hawai’i study. The percent of people who would want to live and work in an historic building is
quite similar; 37% to 39% of Ohio respondents would want to live in an historic building and
42% of Hawai’i participants would want to work in an historic building. These results suggest
that although people may generally value historic places, when preservation is put in the context
of daily interaction with historic buildings, there is perhaps less interest.
A final non-academic study of note was conducted for the National Trust for Historic
Preservation in 2011. Aside from indicating that the study was administered online to 3,734
people, no additional methodological information was provided. (I have chosen to report this
study because of the minimal amount of relevant research available; however, because the report
lacks methodological information, the findings should be considered with reservations.) The
purpose of the study was to learn who local preservationists are, what they do, and how they
become actively engaged in preservation. One of the primary findings was that roughly 29% of
U.S. adults are sympathetic to the work of historic preservation. No information was provided
regarding how this figure was quantified, making it difficult to fully understand the data, but
regardless, this figure appears to be significantly lower than the overall levels of support
observed in the Hawai’i and Arizona studies.
44

2.4 Research on Engagement in Historic Preservation
The National Trust study appears to be the only research that provides insight into the
public’s engagement in historic preservation. The study looked at “local preservationists,” a
group of people who are involved in traditional historic preservation activities on a somewhat
casual level. The Trust identified 7% of the U.S. population as local preservationists. Again,
since the report did not contain methodological information regarding how “local
preservationist” was defined quantitatively, the applicability of the data is somewhat limited;

43. Eric W. Rademacher, The Ohio Poll – Project Report for: Ohio Historic Preservation Office, Institute for
Policy Research, University of Cincinnati (Columbus, OH: Ohio Historic Preservation Office and University of
Cincinnati, n.d.).
44. National Trust for Historic Preservation, Field Guide to Local Preservationists: Learn Who They Are, What
They Do, and How to Win Their Support (Washington D.C.: National Trust for Historic Preservation, 2011).
24

however, the idea that less than 10% of the population is engaged locally is useful to keep in
mind.
As part of their research, the Trust asked the local preservationists about preservation-
related activities in which they would be most likely to participate in the future. Respondents
were most interested in visiting historic places or seeing interesting architecture (74-79% would
likely engage in these activities). Further, nearly three quarters (73%) would attend a lecture on
important historic sites and 62% said they would be likely to promote an historic place on social
media. Just over half would be interested in fundraising or actively volunteering in some way, by
for example, repairing an historic building, participating in an archeological dig, surveying the
city, etc. Although this data provides useful information regarding what a subset of the
population would be likely to do in the future, little is known regarding the extent to which the
population actually engages in activities like these. If local preservationists’ future interests are,
however, indicative of how people actually engage in preservation, we can expect visiting
historic places to be a top activity among the general population, with traditional preservation
activities like fundraising and volunteering being less popular.
45

2.5 Demographic Trends
Demographics and Preference
In 1999, Arthur E. Stamps III, conducted a meta-analysis to summarize the body of
literature that has investigated demographic effects in environmental preference. Forty studies
were included in Stamps’ analysis, which aggregated 5,301 participants’ evaluations of 1,001
scenes. Since the current study is a study of architecture, it is important to note that the stimuli
included in the meta-analysis were both natural and architectural scenes, with the majority being
natural. Despite this caveat, the outcome of the meta-analysis can still provide a general sense of
direction for future studies of the built environment.
Among his findings, Stamps reported high preference correlations across gender,
cultures, and among students and the general population (r = .84, .85, and .83 respectively). He
also found a strong correlation among people of different ethnicities and political affiliations (r =
.87 and .85), but he noted that because there were not a lot of studies contributing to these
findings, these findings were weakly established and additional research was needed for more

45. Ibid.
25

conclusive results. He further found a low correlation (.63) between the preferences of children
(<13 years old) and adults, indicating a need to better understand the relationship between
preference and participant age.
Lastly, Stamps reported a low preference correlation (.56) between the general public and
special interest groups, which according to Stamps could include neighborhood activists or those
with business or conservation interests, for example. It is unknown if Stamps was referring to
heritage conservation or perhaps environmental conservation, but this is an interesting point to
consider for preservationists who may assume that their environmental preferences are
representative of the general publics’.
46

Based on Stamps’ analysis, we can conclude that there are some demographic trends in
environmental preference. To what extent, though, are there trends in preference for old versus
new buildings, specifically? Less is known about this. Of the studies described in section 2.2 that
investigated preference for old and new buildings, four also assessed demographic effects,
including age, gender, ethnicity, income, political affiliation, education, and familiarity.
In Berman’s study, a strong and generally positive correlation was identified between
increasing participant age and increasing preference for older buildings that had been restored.
Specifically, preference for old, restored buildings increased as participant age increased up to
the age range of 45-64. Preference then dropped among those ages 65 and up. Berman further
found that the youngest participants (those ages 15-24) most liked the modern, International style
buildings and those ages 45-64 liked them the least.
47
In contrast, in the Stamps high-rise study,
(which, it should be noted, was included in the Stamps meta-analysis), there were no preference
differences among participants of different age groups. Stamps further found no evidence of
trends in preference based on differences in ethnicity or income. There was evidence, though, for
differences in preference among politically liberal and conservative people. Specifically, in
comparison to other participants, liberal participants had a statistically significant preference for
older buildings over complex new buildings. Further, conservative participants preferred plain
new buildings more than older buildings, whereas the opposite was true among other
participants.
48

46. Stamps III, “Demographic Effects in Environmental Aesthetics: A Meta-Analysis,” 155-175.
47. The findings reported in this section include only Berman’s U.S. participants; Berman, “Assessing Urban
Design.”
48. Stamps III, "Public Preferences for High Rise Buildings," 839-844.
26

Three studies investigated an association between education and preference for old and
new buildings. Frewald, whose participants were undergraduate students, found no differences in
preference based on the number of years of college completed.
49
In contrast, Stamps and
Berman, whose samples represented the general public, identified differences. Stamps found that
people with a high school education or less liked old buildings and plain new buildings equally,
whereas other participants preferred old over plain new.
50
Berman found a strong and positive
relationship between increasing education and increasing preference for restored old buildings,
as well as a negative correlation between education and modern-International style buildings.
51

Related to level of education, both Frewald and Levi compared the preferences of
participants with architectural knowledge and those without. Whereas Frewald found no
relationship between the preferences of architecture and non-architecture students, Levi found
that those participants with more historic architectural knowledge were more likely to prefer old
buildings and less likely to prefer new.
52

The final demographic variable investigated by researchers who have studied preferences
for new and old buildings is familiarity. Berman found that participants who both lived and
worked in the area from which his stimuli were selected were more likely to appreciate modern-
International style buildings. Visitors to the city, as well as participants who worked in or lived
in the area (but not both) had lower preference scores for the modern-International style
buildings.
53

Based on this research, it appears that there is still much to be learned regarding
demographic differences in preference for old and new buildings, both because the amount of
data is minimal and because there are, in some instances, contrasting findings. As such, it is
important to consider demographic variables in the design and analysis of future studies that
explore preference for old and new buildings.

Demographics and Preservation Attitudes
Of the four studies described in section 2.3 on attitudes related to historic preservation,
only one, the study by Daniel Levi, reported findings based on participants’ demographics. For

49. Frewald, “Preferences for Older Buildings.”
50. Stamps III, "Public Preferences for High Rise Buildings," 839-844.
51. Berman, “Assessing Urban Design.”
52. Frewald, “Preferences for Older Buildings”; Levi, “Does History Matter?” 148-159.
53. Berman, “Assessing Urban Design.”
27

two of his attitudinal statements, he identified statistically significant differences between
architecture students and non-architecture students. Architecture students were less likely to
agree with the statement, “Historic buildings should only be preserved if they are of historic
importance to the community,” and more likely to disagree with the statement, “Historic
buildings should be preserved, regardless of how beautiful they are.”
54

Demographics and Engagement
In their study, the National Trust provided demographic information for those they had
identified as local preservationists and preservation leaders. (Table 2.2)
55
In general, the local
preservationists were young, mostly male, and mostly white. Most had at least some college
education, but they were not particularly wealthy. Urban and suburban people were equally
likely to be local preservationists. Those who were very involved in preservation – i.e., the
preservation leaders – were older, and more female, white, and educated than the local
preservationists. They were, further, wealthier and more likely to be city dwellers.
56
Aside from
this study, no additional information appears to exist regarding the demographics of people who
are engaged in historic preservation.

54. Levi, “Does History Matter?” 148-159.
55. As with those who were identified as local preservationists, no methodological information was provided
regarding how local leaders were categorized.
56. National Trust for Historic Preservation, Field Guide to Local Preservationists.
28

Table 2.2: Demographics of local preservationists and preservation leaders
Demographic Variable
Local
Preservationist
Preservation
Leader
Age Average age 35 51
Gender Male 61% 35%
Female 39% 65%
Race /
Ethnicity
White 67% 93%
African American 16% 2%
Hispanic 9% 1%
American Indian or Alaska Native 1% <1%
Asian / Pacific Islander 6% 2%
Other 1% 2%
Education Some college 69% 96%
Bachelors 40% 94%
Masters 13% 71%
Doctorate / Law 1% 11%
Income Income <$50,000 50% 18%
Income >$100,000 12% 32%
Urbanicity Urban 42% 62%
Suburban 41% 18%
Rural 16% 18%
Reproduced from National Trust for Historic Preservation, Field Guide to Local Preservationists:
Learn Who They Are, What They Do, and How to Win Their Support (Washington D.C.: National
Trust for Historic Preservation, 2011)

2.6 Associations among Preference for Old Buildings, Preservation Attitudes, and
Engagement in Preservation
To the author’s knowledge, the only research in which an association between preference
for old buildings and preservation attitudes has been investigated is the aforementioned study by
Daniel Levi. Levi did not find a correlation between participants’ preference ratings and their
attitudes about historic preservation.
57
However, in a study about cultural landscapes (rather than
old buildings, specifically), contradictory findings were identified. Amanda J. Walker and Robert
L. Ryan found that support for landscape preservation had a strong and positive correlation with
attachment to the landscape.
58
Although attachment to place is a somewhat different concept than
preference, as was discussed in section 2.2, and although the study involved landscapes rather
than buildings, the finding is worth noting given the lack of more germane research. Although
there is minimal research available regarding preference and preservation attitudes, there does
not appear to be any existing research related to associations between preference for old
buildings and engagement or preservation attitudes and engagement, indicating a need for
research in these areas.

57. Levi, “Does History Matter?” 148-159.
58. Amanda J. Walker and Robert L. Ryan, “Place Attachment and Landscape Preservation in Rural New
England: A Maine Case Study,” Landscape and Urban Planning 86.2 (2008): 141-152.
29

2.7 Predictors of Preference in Visual Preference Research
In addition to investigating which environments people prefer most, preference
researchers have long sought to understand why people prefer certain environments over others.
The topic is widely researched, complex, and often the subject of entire theses and dissertations.
The purpose of this section is to present a brief overview of the literature in terms of what is
most researched and most relevant to age-based preference studies.
In 1989, Rachel and Stephen Kaplan proposed a framework of human’s environmental
perceptions that involved four measures of preference: coherence, legibility, complexity, and
mystery.
59
Researchers have since fervently tested this model with mixed results. In 2004, Arthur
E. Stamps III conducted a meta-analysis of the literature investigating correlations between
preference and coherence, legibility, complexity, and mystery. In his review of 28 studies, 6,288
participants and 1,820 scenes (depicting both natural and built environments), Stamps’ goal was
to determine the collective correlations among preference and the four variables; he found,
however, that the range of results produced by these studies was too varied to calculate a
collective correlation for any of the four variables. For example, whereas one study identified a
correlation of -0.11 between preference and complexity, another reported 0.97. Others,
meanwhile, found no correlation. Stamps concluded that although there is likely an association
between the variables and preference, the strength and direction is unknown.
60

Although the Kaplans’ model is frequently studied, three of the four dimensions –
coherence, legibility, and mystery – appear to be more relevant to studies of nature scenes and
street scenes, rather than studies of scenes of individual buildings. The fourth dimension of
complexity appears to be the dimension that is most frequently investigated by contemporary
environmental researchers who study the relationship between building age and preference. In
age-based research, researchers have generally found a strong and positive relationship between
complexity and preference.
61
This relationship appears to be multifaceted; when studying
preference for old and new places of worship, Herzog and Shier identified that complexity was a

59. Rachel Kaplan and Stephen Kaplan, The Experience of Nature (Cambridge: Cambridge University Press,
1989).
60. Arthur E. Stamps III, “Mystery, Complexity, Legibility, and Coherence: A Meta-Analysis,” Journal of
Environmental Psychology 24 (2004): 1-16.
61. See, for example: Herzog, Kaplan, and Kaplan, "The Prediction of Preference for Familiar Urban Places,"
627-645; Frewald, “Preferences for Older Buildings”; Herzog and Shier, “Complexity, Age, and Building
Preference,” 557-575; Herzog et al., “Preference and Tranquility for Houses of Worship,” 504-525; Wells,
“Attachment to the Physical Age of Urban Residential Neighborhoods.
30

moderator of the association between building age and preference; meaning that, complexity was
a predictor of preference for both old and new buildings, however, the effect was stronger for old
buildings.
62

In addition to the role of complexity in age-based preference studies, as was discussed in
section 2.2 of the literature review, researchers have come to better understand the relationship
between maintenance and preference for old and new buildings. That is, again, it appears that
when building maintenance is not controlled, new buildings are preferred and the opposite is true
when maintenance is controlled. As such, researchers of age-based preference studies typically
control for building condition in order to investigate preference.
Visual preference researchers have further identified nature as a moderate predictor of
preference in environmental studies.
63
Scenes with more foliage and greenery are, on average,
more preferred. For example, Herzog, Kaplan, and Kaplan found that 71% of participants’ most
preferred scenes (the top quartile) had substantial foliage compared with 44% in the 2
nd
quartile,
11% in the 3
rd
quartile, and 0% in the 4
th
quartile.
64
Similarly, W.C. Sullivan, when investigating
preference for housing, found that scenes with greater numbers of mature trees were preferred
over those with fewer mature trees.
65
This relationship is well established in visual preference
studies in general, but it appears to be evident in age-based studies, as well.
Many additional variables have been investigated by researchers as possible predictors of
environmental preference, including: compositional elements, such as the angle of the
photograph; architectural elements, such as massing, type of entrance, number of windows; and
elements of the scene, such as the presence of people, vehicles, signs, garbage, or wires.
66
In
general, each of these variables have been studied with less frequency than the aforementioned
variables of complexity, maintenance, and nature, and, are generally given less consideration by
researchers. It appears to be good practice, however, to control for elements that the researcher

62. Herzog and Shier, “Complexity, Age, and Building Preference,” 557-575.
63. For example: Nasar, “Visual Preferences in Urban Street Scenes,” 79-93; W. C. Sullivan, “Perceptions of the
Rural-Urban Fringe: Citizen Preferences for Natural and Developed Settings,” Landscape and Urban Planning 29
(1994): 85-101; Nasar, “Adult Viewers’ Preferences in Residential Scenes,” 589-614; Herzog, Kaplan, and Kaplan,
“The Prediction of Preference for Unfamiliar Urban Places,” 43-59; Herzog and Gale, “Preferences for Urban
Buildings as a Function of Age and Nature Context,” 44-72.
64. Herzog, Kaplan, and Kaplan, “The Prediction of Preference for Unfamiliar Urban Places,” 43-59.
65. W. C. Sullivan, “Perceptions of the Rural-Urban Fringe: Citizen Preferences for Natural and Developed
Settings,” Landscape and Urban Planning 29 (1994): 85-101.
66. For example: Nasar, “Adult Viewers’ Preferences in Residential Scenes,” 589-614; Nasar, “Visual Preferences
in Urban Street Scenes,” 79-93.
31

thinks may relate to preference. For example, in “Preferences for Urban Buildings as a Function
of Age and Nature Context,” Herzog and Gale, in their selection of stimuli, controlled for
building maintenance and landscape, signs and people were avoided, and all photographs were
oriented horizontally and were taken the same time of year with no fall colors or weather visible.

2.8 Conducting Research using Amazon Mechanical Turk Participant Samples
As will be described in section 3.2, the participant sample was drawn from workers on
Amazon Mechanical Turk (MTurk). Since MTurk is somewhat new in academic research and
few, if any, visual preference studies have been conducted using MTurk samples, an overview of
the platform is provided below. Also provided is a brief review of the academic research
regarding the validity and reliability of conducting surveys on MTurk.
Amazon Mechanical Turk is a virtual marketplace in which employers outsource labor
intensive projects to workers who individually complete small components of the work – called
“Human Intelligence Tasks,” or “HITs” for short – in exchange for a nominal payment.
67
The
website launched in 2005, and although it has traditionally been used by businesses to complete
such tasks as transcription and coding, it has in recent years become a popular source of subjects
for academic research, in part because of the large pool of potential participants, low costs, and
immediacy of data collection.
68

As a new method of conducting research, questions arise regarding the motivations of
people who participant in studies for minimal pay, representativeness of MTurk workers in
comparison to the general population, and the validity of the data collected through the website.
69

Researchers from various disciplines have investigated these issues and have generally found
that despite some limitations – particularly for certain kinds of research – Mechanical Turk is an
appropriate tool for conducting academic research.
The motivations of MTurk workers – who may be paid less than a dollar for participating
in a survey – can be described as intrinsic and extrinsic. Most workers report that they find

67. See: “Amazon Mechanical Turk,” accessed May 1, 2014, www.mturk.com.
68. Panagiotis G. Ipeirotis, “Analyzing the Amazon Mechanical Turk Marketplace,” 16-21; Joseph K. Goodman,
Cynthia E. Cryder, and Amar Cheema, “Data Collection in a Flat World: The Strengths and Weaknesses of
Mechanical Turk Samples,” Journal of Behavioral Decision Making 26 (2013): 213-224.
69. It should be noted that although MTurk workers live worldwide, since the study conducted for this thesis
involves U.S. participants only, findings regarding the representativeness and validity of American MTurk workers
is most relevant. Unless otherwise noted, all research described involves only American participants.
32

MTurk to be a fruitful way to spend their free time and a way to earn spare money. Some
additionally indicate that they find the tasks to be fun and enjoyable. And although people may
assume that MTurk workers tend to be unemployed, only a small minority rely on MTurk as
their primary source of income, and, furthermore, the percent of MTurk workers employed in
certain sectors mirrors a traditional survey population.
70

With regard to the primary demographic variables of age, gender, education, income, and
race / ethnicity, researchers have consistently found that the population of American MTurk
workers is slightly younger, more female, and more educated than the general U.S. population,
and although the MTurk income distribution is similar to the general population, MTurk workers
are slightly poorer on average.
71
Additionally, the population of Asian workers on MTurk is
higher than the general population and the population of Black and Hispanic workers is lower.
72

Despite these differences, the MTurk sample is generally more representative of the U.S.
population than the traditional college student participant pool, and at least as representative as
traditional Internet samples and convenience samples.
73

However, significant differences have been identified between MTurk samples and
Internet and convenience samples in regard to political affiliation and personality. One study
concluded that an MTurk sample was more politically liberal than a convenience sample (and
also the general population), which has implications for studies involving politically oriented
attitudinal statements.
74
Another study found personality differences between MTurk participants
and both student and convenience samples. Specifically, MTurk workers were more likely to
score lower than convenience sample and student sample participants on the traditional
personality dimensions of extraversion and emotional stability, and they were also less open to

70. Gabriele Paolacci, Jesse Chandler, and Panagiotis G. Ipeirotis, “Running Experiments on Amazon Mechanical
Turk,” Judgment and Decision Making 5.5 (2010), 411-419; Panagiotis G. Ipeirotis, “Demographics of Mechanical
Turk,” (white paper, Leonard N. Stern School of Business, New York University, New York City, NY, 2010).
Connor Huff and Dustin Tingley, “’Who are these People?’: Evaluating the Demographic Characteristics and
Political Preferences of MTurk Survey Respondents,” (white paper, Harvard University, Cambridge, MA, 2014).
71. Ipeirotis, “Analyzing the Amazon Mechanical Turk Marketplace,” 16-21; Ipeirotis, “Demographics of
Mechanical Turk,”; Adam J. Berinsky, Gregory A. Huber, Gabriel S. Lenz, “Evaluating Online Labor Markets for
Experimental Research: Amazon.com’s Mechanical Turk,” Political Analysis 20 (2012): 351-368; Paolacci,
Chandler, and Ipeirotis, “Running Experiments on Amazon Mechanical Turk,” 411-419; Goodman, Cryder, and
Cheema, “Data Collection in a Flat World,” 213-224.
72. Berinsky, Huber, and Lenz, “Evaluating Online Labor Markets for Experimental Research,” 351-368.
73. Ibid.
74. Ibid.
33

experiences than convenience sample participants and less conscientious than students.
75

Additional research, however, has demonstrated that there are no significant relationships
between personality dimensions and architectural preferences.
76
For the purposes of this study,
the personality differences of MTurk workers may have implications for the attitudinal and
preservation participation components of the research, but the variable is not expected to affect
the preference component.
With regard to the reliability of data obtained via MTurk, researchers generally find it to
be as reliable as data collected via traditional research samples. For example, researchers have
generally not found statistical differences in how MTurk participants and traditional participants
respond to survey items, and additionally, researchers have been able to replicate the findings of
classic experiments on Mechanical Turk.
77
Furthermore, research suggests that MTurk workers
are, for the most part, honest and conscientious research participants. They report their
demographic profile with consistency over time and appear to only take surveys once.
78
(Rand
2012; Mason & Suri 2012). Lastly, when tested to see if they pay full attention to survey prompts
(such as being asked, “While watching the television, have you ever had a fatal heart attack?”),
MTurk samples perform better than Internet research samples and just as well as convenience
samples.
79
MTurk workers were, however, more likely to fail the attention check than
participants from a student research pool.
80
Researchers have found that the inclusion of the data
of MTurk participants who fail attention checks does not affect overall findings, but does reduce
statistical power and increase the potential for Type II error.
81
As such, many researchers who

75. Goodman, Cryder, and Cheema, “Data Collection in a Flat World,” 213-224.
76. Kristyn Clayton, “Personality and Architectural Preferences: A Search for Patterns” (master’s thesis,
Washington State University, 2007).
77. Paolacci, Chandler, and Ipeirotis, “Running Experiments on Amazon Mechanical Turk,” 411-419; Berinsky,
Huber, and Lenz, “Evaluating Online Labor Markets for Experimental Research,” 351-368; Michael Buhrmester,
Tracy Kwang, and Samuel D. Gosling, “Amazon’s Mechanical Turk: A New Source of Inexpensive, Yet High-
Quality, Data?” Perspectives on Psychological Science 6.1 (2011): 3-5.
78. Berinsky, Huber, and Lenz, “Evaluating Online Labor Markets for Experimental Research,” 351-368; Danielle
N. Shapiro, Jesse Chandler, and Pam A. Mueller, “Using Mechanical Turk to Study Clinical Populations,” Clinical
Psychological Science 1.2 (2013): 213-220.
79. Berinsky, Huber, and Lenz, “Evaluating Online Labor Markets for Experimental Research,” 351-368;
Paolacci, Chandler, and Ipeirotis, “Running Experiments on Amazon Mechanical Turk,” 411-419.
80. Goodman, Cryder, and Cheema, “Data Collection in a Flat World,” 213-224; Berinsky, Huber, and Lenz,
“Evaluating Online Labor Markets for Experimental Research,” 351-368; Paolacci, Chandler, and Ipeirotis,
“Running Experiments on Amazon Mechanical Turk,” 411-419.
81. For reference, Type II error is a failure to identify the real outcome, such as an analysis returning a false
negative or a false positive result. Goodman, Cryder, and Cheema, “Data Collection in a Flat World,” 213-224.
34

use MTurk samples include an attention check and exclude the data of those participants who fail
the test.
Inter-worker communication poses another concern regarding the validity of data
collected via MTurk, particularly when the researcher screens participants for eligibility prior to
participation in the study or when knowledge regarding the study’s content or purpose may
influence results. Twenty-six percent of MTurk workers reported personally knowing another
worker and 28% indicated that they read online worker forums and blogs, such as those on
www.mturkforum.com, www.turkernation.com, Facebook, or Reddit.
82
One study reported,
however, that only a small minority of workers (2.4%), submitted HITs from the same IP address
as another worker, indicating the unlikelihood that participants share information about a study
with another worker in the same household.
83
Furthermore, workers report that the primary
reasons they discuss HITs on online forums is to share information about payment, completion
time, and a requester’s reputation. They are less likely to have a conversation about the purpose
of a HIT, and only 13% report having ever seen an online discussion regarding the content of a
research study.
84
Although inter-worker communication that would affect the validity of data
appears to be uncommon, the researcher can do things to minimize the threat, such as monitoring
popular online discussion forums for content related to their study.
Though additional research is necessary to better understand the advantages and
limitations of MTurk samples, thus far, researchers have generally come to a favorable
conclusion regarding Mechanical Turk.
85
That is, although the demographic characteristics of
MTurk samples are somewhat different than the U.S. population, they are more diverse than
traditional research samples, and the data obtained from workers is generally as reliable as data
obtained via conventional survey methods. This, coupled with the affordability and ease of
conducting studies on MTurk, makes it an attractive – though not perfect – survey research
method.

82. Jesse Chandler, Pam Mueller, and Gabriele Paolacci, “Nonnaïveté among Amazon Mechanical Turk Workers:
Consequences and Solutions for Behavioral Researchers,” Behavioral Research Methods 46.1 (2014): 112-130.
83. Berinsky, Huber, and Lenz, “Evaluating Online Labor Markets for Experimental Research,” 351-368.
84. Chandler, Mueller, and Paolacci, “Nonnaïveté among Amazon Mechanical Turk Workers,” 112-130.
85. For example: Berinsky, Huber, and Lenz, “Evaluating Online Labor Markets for Experimental Research,”
351-368; Goodman, Cryder, and Cheema, “Data Collection in a Flat World,” 213-224; Paolacci, Chandler, and
Ipeirotis, “Running Experiments on Amazon Mechanical Turk,” 411-419.
35

2.9 Conclusion
A review of the relevant research reveals that generally, people appear to prefer old
buildings over new and they have mostly positive attitudes about historic preservation; however,
the limitations of the existing research and the small number of contributing scholarly studies
warrants further investigation. Additionally, very little is known about the public’s engagement
in preservation activities; the extent to which demographics are associated with preference for
old and new buildings, attitudes about preservation, and preservation-related engagement; and
whether there is an association between preference, attitudes, and engagement. This study aims
to address these gaps in the literature and also provide more evidence regarding environmental
predictors of preference. The use of Amazon Mechanical Turk was reviewed for the purposes of
this investigation and determined to be a suitable sampling methodology.
36

3. METHODOLOGY

The following chapter begins with an overview of the survey instrument, followed by a
description of the methodological development of the visual preference component of the survey.
The survey administration methods are then summarized. The chapter concludes with a section
in which the data analysis methods are described.

3.1 The Survey Instrument
3.1.1 Description of the Survey
The survey, which is available in Appendix B consisted of four sections: 1) visual
preference items, 2) attitudinal items, 3) engagement items, and 4) demographic items. The
visual preference section followed the standard visual preference survey format in which stimuli
were presented to participants one at a time with a preference scale. A total of 50 stimuli were
rated by participants in two sets: an initial set of 30 stimuli to begin the survey, followed by a
later set of 20 stimuli. To provide participants with a mental break from the preference task, a set
of demographic items were completed between the two sets of stimuli. Within each preference
section, the stimuli were presented in a random order with equal numbers of old and new
buildings (definitions of which will be provided in section 3.1.2). Participants were given the
following instructions for the preference items:

You will be presented with a series of images of buildings and will be asked to rate how
much you like the appearance of each building. Consider only the appearance of each
building and not other factors such as the neighborhood, lighting, or landscaping.

A version of these instructions was repeated after the first 10 images as a reminder, and again
before the second set of images. The preference question and scale, which was coded from one
(not at all) to seven (like a lot), is provided below.

Not at all -- Somewhat -- Like -- Like a lot
How much do you like
the appearance of this
building?
ο ο ο ο ο ο ο

37

The attitudinal items were presented after the second set of preference items. The first
attitudinal question was the community service item, which was based on the previously
discussed item from Daniel Levi’s study, “Does History Matter?” It was designed to assess how
important people feel preservation is in general and in relation to other community services.
Participants were given the following instructions for the item:

Below is a list of services a city / town might provide its citizens. Not considering your
community specifically, but communities in general, how important is each service to
you?
1

The list of services, which were repeated directly from Daniel Levi’s study, was randomized and
included: historic preservation, architectural review for aesthetics of new construction, tourism
promotion, economic development, recreational programs, public art, cultural events, and street
trees and public landscaping. Participants were provided with the definitions listed below.

Historic places: districts, sites, buildings, structures, or objects significant to history,
architecture, engineering, archeology, or culture at the national, state, or local level

Historic preservation: the process of maintaining or managing appropriate change to
historic places
2

The language for the definition of historic places was modified from the National Register of
Historic Places Criteria for Evaluation.
3
The definition of historic preservation was intended to
be simple, yet broad so that people would interpret it as more than just preservation, but also
things like adaptive re-use (rehabilitation), and beyond simply historic buildings, but also other
kinds of historic places. The scale for the importance items was seven points and is provided
below.
4

Not at all
important --
Slightly
important --
Moderately
important --
Very
important
5

ο ο ο ο ο ο ο

1. These instructions differed slightly from Daniel Levi’s study because his participants all lived in a city which
offered each of these services.
2. Daniel Levi does not indicate that he provided definitions of terms in his study.
3. “National Register Bulletin: How to Apply the National Register Criteria for Evaluation,” National Park
Service, accessed May 5, 2014, http://www.nps.gov/nr/publications/bulletins/nrb15/nrb15_2.htm.
4. Whereas a seven-point scale was used in the current study, Levi used a four-point scale.
5. The scaled was coded from one (not at all important) to seven (very important).
38

The next set of attitudinal items was a series of nine statements designed to better
understand how participants view historic preservation and the extent to which they support it.
Content included: general beliefs about the value of historic preservation and the extent to which
we preserve in the U.S.; beliefs about preservation-related regulation and funding; the
relationship between preservation and economic development, as well as preservation and
environmental sustainability; and the extent to which people have been negatively affected by
the loss of a building. The items (listed below) were randomized and presented with a seven-
point agreement scale, which was coded from one (strongly disagree) to seven (strongly agree).
Two of the statements were reversed so that a low score would represent pro-preservation
sentiments.

Attitudinal Items

Historic places should be treated as community assets

We do too little to protect historic places in the U.S.

I have been saddened by the demolition of an historic place

Local governments should have the ability to prevent the demolition of historic places

The preservation or re-use of historic places is environmentally friendly

Owners of historic buildings should be able to demolish their building without any kind of legal
restrictions

Tax payer dollars should be used to help preserve privately owned historic places because this
activity benefits the public

Historic preservation gets in the way of economic development

Preserving historic outdoor spaces is as important as preserving historic buildings

Strongly
disagree Disagree
Somewhat
disagree
Neither agree
nor disagree
Somewhat
agree Agree
Strongly
agree
ο ο ο ο ο ο ο

The next section of the survey was the preservation engagement section. Participants
were first asked to indicate which, if any, typical preservation-related activities they had
participated in during the last year. The nine activities (listed below) ranged from things that
require little effort or resources (e.g., following an historic preservation organization on social
media) to those that require more effort / resources (e.g., nominating a place for historic
designation). The activities were presented in random order to each participant, participants
39

could check all the activities that applied, and the terms historic places and historic preservation
were again defined.

Engagement Activities

Visited a place because it was historic

"Liked" or "shared" a post on social media related to the preservation of an historic place

Followed an historic preservation organization on social media (E.g., National Trust for Historic
Preservation, Docomomo, Cultural Landscape Foundation, Preservation Texas, Los Angeles
Conservancy, etc.)

Signed a petition for the preservation of an historic place

Contacted a public official to support the preservation of an historic place

Advocated for the preservation of an historic place at a public hearing

Paid for a membership to an historic preservation organization

Volunteered for an historic preservation organization

Nominated an historic place for local, state, or national historic designation

In addition to asking about their engagement in preservation in the last year, participants
were asked about obstacles to getting involved in historic preservation. The purpose of the
preservation obstacle items (listed below with the response scale) was to determine the extent to
which people would participate in historic preservation efforts, but either do not know how to or
do not have the time to get involved.

Preservation Obstacle Items

I would like to support historic preservation efforts in my community, but I do not have the time.

I would like to support historic preservation efforts in my community, but I do not know how to get involved.

Strongly
disagree Disagree
Somewhat
disagree
Neither
agree nor
disagree
Somewhat
agree Agree
Strongly
agree
N/A – I am
already
involved
6

ο ο ο ο ο ο ο ο

The final section of the survey was the demographic section. The demographic
information collected included age, gender, ethnicity, education, political orientation, state, and
urbanicity. Participants were additionally asked to indicate if they had higher education or

6. The scale was coded from one (strongly disagree) to eight (N/A – I am already involved).
40

professional work experience in architecture, urban planning, or historic preservation. Since
previous studies about historic preservation have often gathered the opinions of those already
involved in preservation, this item was included to ensure that participants in the current study
better represented the general public.
The survey was piloted to a group of eight participants prior to being launched. Pilot
participants’ ages ranged from 18 to 65 and no participant had a professional architecture, urban
planning, or historic preservation background. To test for technical issues with Qualtrics, pilot
respondents were asked to report the type of electronic device they used to take the study (laptop,
desktop, or tablet), technical glitches they encountered, as well as the relative speed with which
images loaded. They were also asked to share feedback regarding instructions, items, language,
and formatting that was confusing or unclear. Lastly, their total survey time was recorded and
they were asked for feedback regarding the length of the study. A small number of content and
formatting changes were made as a result of the pilot administration.

3.1.2 Compiling a Pool of Stimuli
As stressed by Henry Sanoff in Visual Research Methods in Design, the selection of
environmental stimuli is a critical component of visual preference research.
7
Despite this, there is
little consensus among researchers regarding appropriate methodological frameworks for
selecting stimuli.
8
Given these two factors, a strict set of sampling procedures were utilized in
the current study, with many of the procedures adhering to the precedent of previous research.
A pool of 116 photographs of buildings was collected for the study which included
photographs found on the Internet, as well as photographs taken by the author. Images obtained
online were primarily located on chapter websites for the American Institute of Architects or
architecture groups on Flickr.
9
In order to ensure consistency across stimuli, the photographs that
comprised the image pool met a number of conditions outlined in Table 3.1. A rigid set of
controls was important in selecting photographs as the more similar the images were in terms of
building, scene, and photographic conditions, the less likely participants would respond to
extraneous factors in assigning preference ratings.

7. Henry Sanoff, Visual Research Methods in Design, 3-5.
8. Ibid.
9. The study was limited to images taken by photographers who agreed to their use. Of the photographers who
responded to email requests to utilize their images, all but one responded affirmatively.
41

Table 3.1: Criteria for the selection of photographic stimuli

Definition of old and new Old Constructed at least 50 years ago (1964 or
earlier)
New Constructed within the last 15 years (1999
or later)
Building conditions Type Non-descript commercial / office
buildings
Style Predominate styles of the time, ranging
from vernacular to high styles, but
excluding historicized new buildings and
buildings void of style
Integrity All or most of the original character
defining features retained
Height Five stories or less
Massing Relatively cubic in shape
Size Any
Materials Any
Color Any
Signage None to limited, avoiding large corporate
signs
Location U.S. and Canada
Scene conditions Setting Little to no setback in a densely built
environment
Visibility No large trees, foliage, vehicles, etc.
obstructing the façade
People / cars None to limited
Photographic conditions Composition Close crop of building, showing all or
nearly all of the building and little
surrounding context
Lighting Front lit during the day time
Weather No rain, snow, fall leaves
Sky Gray skies avoided
Image quality In focus, not grainy, not washed out or too
dark; no photographic filters applied, not
overly Photoshopped, etc.
Image orientation Horizontal
Perspective Three-quarter view or façade view
42

The reason or purpose for each of these parameters varies. Regarding the definition of
“old” buildings, this is in line with the federal eligibility standard established by the National
Park Service, which limits historic designation to those resources constructed more than 50 years
ago.
10
Hence all “old” buildings in the study were constructed on or before 1964. Buildings
defined as new would have ideally included only those constructed in the last 10 years to allow
for more of a gap between old and new; however, because construction of new buildings stalled
during the late 2000s financial crisis, a sufficient number of new buildings less than 10 years old
that fit the study parameters was difficult to locate. As such, the definition of new was extended
to 15 years.
11
Where building age could not be confirmed with assessor records or online
research, visual estimates were utilized.
Commercial and office buildings were chosen as the type of building for the study in lieu
of other building types because of the abundance of the type and the consistency in massing and
setting that would be available. Further, from a preservation perspective, old, low-rise “Main
Street” type buildings are often the types of buildings facing demolition threats in favor of newer
development, so an understanding of the extent to which people prefer them would be useful
from a practical standpoint. The study was restricted to just commercial and office buildings,
(rather than a broader range of building types, as has been done in previous research), to allow
for a more fair comparison by eliminating potential confounders.
Since Herzog, Kaplan, and Kaplan established that there is a relationship between
familiarity and preference, the commercial and office buildings that were selected for the study
were relatively nondescript.
12
Though some have historic resource designation and others have
received awards by the American Institute of Architects and other organizations, the buildings
would be mostly unknown to the general population.
Although the building type was held constant, a wide variety of styles were selected in
order to ensure the study addressed the question of old versus new rather than stylistic
preference. All the buildings had stylistic elements; there were, for example, no plain stucco
boxes. Further, in order to capture the range of architecture that exists, the buildings varied from

10. A property that is less than 50 years old may be considered for national historic designation if it has
“exceptional importance,” but generally, historic resources are those that are 50 years old or older. “National
Register of Historic Places Program: National Register Federal Program Regulations,” National Park Service,
accessed June 22, 2014, http://www.nps.gov/nr/regulations.htm#604.
11. Authors of future studies of this nature may want to consider a more narrow definition of new buildings.
12. Herzog, Kaplan, and Kaplan, “The Prediction of Preference for Familiar Urban Places,” 627-645.
43

vernacular to high styles. A relatively equal number of award winning new buildings and old
buildings with architecturally related historic designation were included, as these buildings could
arguably represent the “best” of each category given their established architectural recognition.
All of the selected buildings represent common styles of the time, and not obscure or rare styles,
which may, due to their foreignness to the general population, influence preference. Likewise,
only buildings from the U.S. and Canada were used since there is some evidence that novelty
may affect preference.
13

Despite Daniel Levi’s findings that people can discern between old buildings and
historicized new buildings, new buildings with historicized features were excluded from the
study in order to provide a clear distinction between old and new buildings.
14
To further
distinguish old and new buildings, the buildings had to retain all or nearly all of their integrity.
This criterion primarily applied to old buildings, which over the years have been more
susceptible to alterations. Buildings were assessed for modifications and in addition to those
buildings that were unaltered, those that had sustained some degree of alteration, but overall
retained a high level of integrity, were included in the pool of photographs.
15

As outlined in the previous table, additional parameters were established to guide the
selection of stimuli. These included massing, height, size, material, color, setting, and visibility
parameters, as well as parameters regarding the presence of corporate signs, people, and cars in
the photographs. Although the buildings could be any size, color, or material, they were all low-
rise and relatively cubic in shape. Each contained little to no set back from the street and
visibility was mostly unobstructed. Large, corporate signs were avoided in the instance that
feelings associated with widely known businesses would influence the preference of the
buildings that housed these businesses. Since the presence of people and cars has been found to
affect preference, these variables were avoided when possible.
16
Finally, photographic
conditions, such as image quality and composition, have the potential to affect ratings in visual

13. Herzog, Kaplan, and Kaplan, “The Prediction of Preference for Unfamiliar Urban Places,” 43-59; Herzog,
Kaplan, and Kaplan, "The Prediction of Preference for Familiar Urban Places,"627-645; A. Oostendorp and D.E.
Berlyne, "Dimensions of the Perception of Architecture: III. Multidimensional Preference Scaling," Scandinavian
Journal of Psychology 19.1 (1978): 145-150.
14. Levi, “Does History Matter,” 148-159.
15. These integrity considerations were made by the author and the author’s faculty advisor, a professor of
heritage conservation in the School of Architecture.
16. Nasar, “Adult Viewers’ Preferences in Residential Scenes,” 589-614; Nasar, “Visual Preferences in Urban
Street Scenes,” 79-93; Jack L. Nasar, "Environmental Correlates of Evaluative Appraisals of Central Business
District Scenes," Landscape and Urban Planning 17 (1987): 117-130.
44

preference research and therefore a variety of photographic variables were controlled for in the
selection of images.
17
All the images were presented horizontally, and the focal building was
centered within the frame with surrounding context cropped out of the scene consistently. Each
building was photographed straight on (showing just the façade) or from an angle, so that both
the façade and a secondary elevation were visible. (I refer to the former as “façade view” and the
later as “three quarter view.”) All buildings were front lit and the images lacked weather related
elements, such as the presence of snow, rain, or fall leaves. Further, although not always
possible, attempts were made to use images with blue skies. Image quality was high and
significantly altered photographs, such as those in which photographic filters had been applied,
were excluded.

3.1.3 Selecting a Final Sample of Stimuli
As described in the literature review, several variables have consistently been found by
researchers to be predictive of building preference. These are: building condition, building
complexity, and the presence of nature or landscape in a photograph. As will be described in
more detail below, these primary variables were controlled for in the current study in order to
ensure that they would be equally present in the old and new building samples. In addition to the
primary variables, there are other factors that may predict the preference of a photographed
building. I refer to these as “secondary variables” because the body of evidence supporting a
predictive relationship is not as strong as that of the primary variables, or a correlation has not
been tested; however, it is plausible to believe that these factors might influence preference. As
such, both primary and secondary variables were controlled for in the current study. The factors
selected as secondary variables include: height, size, cladding type, and location of the building;
architectural recognition and designation; presence of cars and people; image orientation; and the
grayness of the sky. Additionally, the age of the building (within the respective old and new time
spans) was considered.
As was discussed in the literature review, there are several methods by which visual
preference researchers control for predictive factors, including rating images themselves for
these factors, using ratings from a participant population, or using ratings from a team of experts.

17. Herzog and Shier. “Complexity, Age, and Building Preference,” 557-575; Stamps III, "Formal and Nonformal
Stimulus Factors in Environmental Preferences," 3-9.
45

In this study, a panel of three experts rated the pool of 116 photographs (67 old buildings and 49
new buildings) for the primary factors of building condition, building complexity, and landscape.
The secondary variable ratings were assigned by the author. In order to ensure the expertise of
the panelists was diverse and well-balanced, the panel included a contemporary architect, a
preservation architect, and a preservation planner. It should be noted that the use of expert and
author ratings in lieu of ratings from a sample of participants may be a limitation of the study as
those with architectural training and experience may analyze and interpret environmental stimuli
differently than those outside the profession.
18
An expert panel was used in the current study due
to resource limitations; however, authors of future studies of this nature may consider using a
participant sample to control for predictive variables.
The experts were provided with the following prompts and seven-point scales for the
rating task.

Building Condition
To what extent is the building well-maintained? Note: When evaluating each image,
consider only the focal building and not its surrounds.

Not at all
maintained
Somewhat
maintained
Well-
maintained
Very well-
maintained
1 2 3 4 5 6 7

Visual Complexity
To what extent does the building have visual complexity? (Visual complexity refers to
the degree to which the building is intricate, rather than simple.) Note: When evaluating
each image, consider only the focal building and not its surrounds.

Not at all
complex
Somewhat
complex

Complex
Very
complex
1 2 3 4 5 6 7

18. See for example, Stamps III, “Demographic Effects in Environmental Aesthetics,” 155-175; Arthur E. Stamps
III and Jack L. Nasar, “Design Review and Public Preferences: Effects of Geographical Location, Public Consensus,
Sensation Seeking, and Architectural Styles," Journal of Environmental Psychology 17 (1997): 11-32; Robert
Gifford, Donald W. Hine, Werner Muller-Clemm, and Kelly T. Shaw, "Why Architects and Laypersons Judge
Buildings Differently: Cognitive Properties and Physical Bases," Journal of Architectural and Planning Research
19.2 (2002): 131-148; Margaret A. Wilson, "The Socialization of Architectural Preference," Journal of
Environmental Psychology 16 (1996): 33-44.
46

Presence of Landscape
To what extent are plants present in the image? (Plants include trees, shrubs, flowers,
grass, etc.) Note: consider the presence of plants throughout the entire image.

Not at all
present
Somewhat
present

Present
Very
present
1 2 3 4 5 6 7

The experts’ ratings for each of the primary variables were averaged. Per the precedent of
previous research, only buildings in good condition were included in the survey; here, those
buildings with poor maintenance scores (means of 4.0 or less) were removed from the list of
potential stimuli.
19
Thirteen of the 67 old buildings had low condition scores and were removed,
leaving 54 old buildings from which to select the final set of 25 stimuli. There were no new
buildings with poor maintenance scores.
The principal goal in the selection of the 50 stimuli was to ensure that the old and new
building sets were balanced in terms of the primary variables. To achieve this, the median
condition, complexity, and landscape scores were determined for the remaining 103 buildings,
and the images coded based on whether they fell above or below each of the three medians,
which were 5.7 for condition, 3.3 for complexity, and 2.7 for landscape. From here, 25
photographs of old buildings and 25 photographs of new buildings were selected, ensuring that
the old and new building sets contained relatively equal numbers of old and new buildings with
high and low complexity scores, high and low condition scores, and high and low landscape
scores. In addition to selecting equal numbers across the high-low conditions, a complete
factorial design was utilized to ensure that each of the eight possible high-low combinations of
condition, complexity, and landscape was represented in the old and new building sets.
20
That is,
there were old and new buildings with high condition, high complexity, and high landscape, as
well as old and new buildings with low condition, high complexity, and high landscape, and so
on.
In addition to considering the primary variables in the selection of the images, the
secondary variables were considered as well; however, whereas the primary variables drove the

19. Levi, “Does History Matter?” 148-159.
20. See factorial cell design precedent in Herzog and Shier, “Complexity, Age, and Building Preference,” 557-
575.
47

selection of the images and care was taken to adequately balance the proportions between the old
and new buildings, the secondary variables, given their lesser importance, were only more or less
balanced among the two sets. As previously mentioned, these variables were assigned ratings by
the author. There are a few things to note regarding the secondary variables. First, Internet
research was conducted to determine which buildings have received architectural recognition or
designation; however, it is possible that additional buildings have been recognized, but this
information was either not available online or not located online.
21
Second, it should be noted
that the presence of cars and people variables were assigned binary ratings (yes or no) rather than
continuous scores to reflect the extent to which cars or people were present. This was done for
simplicity; however, authors of future studies may wish to assign continuous ratings. Third,
regarding age and cladding: rather than balancing these variables among the old and new
building sets, the purpose of monitoring these two variables was to ensure that there was variety
of age and cladding types represented across the stimuli rather than a balance. A smaller number
of post-WWII buildings were selected within the old building set because the span of time this
period represents is significantly smaller than the span of time prior to World War II. With
cladding, due to changes in architecture and construction, new and old buildings are frequently
clad in different materials. Since cladding could not be balanced between the old and new sets,
variety within the old and new building sets was important to control its confounding effect on
the ratings of buildings.
Based on the described selection methods, a set of 50 stimuli were chosen for the study.
The images are presented in Appendix C and the composition of primary and secondary
predictor variables within the old and new building sets is provided below in Tables 3.2 and 3.3.

Table 3.2: Number of stimuli with high / low ratings
on the primary predictor variables
Primary variables
# Low # High Total
Building
Condition
Old 12 13 25
New 12 13 25
Building
Complexity
Old 10 15 25
New 10 15 25
Presence of
Landscape
Old 11 14 25
New 10 15 25

21. Designation includes buildings listed for their architectural significance on local, state, or national registers of
historic places either individually or as a contributor to a district.
48

Tables 3.3: Distribution of old and new building stimuli across the secondary variables
Height of building
1 story 2 stories 3 stories 4 stories 5 stories Total
Old 5 8 8 1 3 25
New 3 5 13 4 0 25

Size of building
Small Medium Large Total
Old 10 13 2 25
New 6 16 3 25

Building location
22

Canada
U.S.:
Northeast
U.S.:
South
U.S.:
Midwest
U.S.:
West
U.S.:
Pacific Total
Old 0 4 4 8 8 1 25
New 0 0 3 3 19 0 25

Architectural recognition /
designation
Yes No Total
Old 9 16 25
New 7 18 25

Cars present
Yes No Total
Old 13 12 25
New 14 11 25

People present
Yes No Total
Old 23 2 25
New 21 4 25

Photo angle
Façade view Three quarter view Total
Old 17 8 25
New 16 9 25

Quality of sky
Gray / washed out sky Blue sky Total
Old 4 21 25
New 1 24 25

22. U.S. locations categorized according to the U.S. Census Regions. “Census Regions and Divisions of the
United States,” United States Census Bureau, accessed June 19, 2014, https://www.census.gov/geo/maps-
data/maps/pdfs/reference/us_regdiv.pdf.
49

Tables 3.3 (Cont’d): Distribution of old and new building stimuli across the secondary variables
Age of old buildings
Built pre-WWII Built post-WWII Total
21 4 25

Age of new buildings
Built 1999-2006 Built 2007-2014 Total
10 15 25

Primary cladding
Painted
brick
Red
brick
Brick
(other)
Stone Stucco Glass Other Total
Old 5 10 4 3 2 0 1 25
New 0 7 1 0 3 7 7 25
Note: For the variables of age of building and cladding type, the intention was to ensure variety rather
than a balance between old and new buildings.

3.2 Survey Administration Methods
Amazon Mechanical Turk workers were recruited to participate in the study. Workers,
who must certify they are at least 18 years old when registering for an MTurk account, were
offered $1.00 for completing the survey, which was in line with the current market rate for
research study participation on MTurk. The study was anticipated to take approximately 15 to 20
minutes and in order to ensure participants completed the study in one sitting, a one hour time
cap was established. Participants who accepted the MTurk HIT were redirected to the consent
form and survey on Qualtrics.
In order to mask the focus of the study as one about historic preservation, the recruitment
content and consent information were vague, indicating only that the survey was about
architecture (and not specifically historic preservation). Furthermore, the preservation-related
items were located at the end of the study, after the preference items. This eliminated the
possibility that participants, after realizing the study was about preservation, would respond more
favorably to the older buildings in an effort to please the researcher. Despite these controls, since
the study was advertised as one about architecture, it is possible that people who generally like or
are interested in architecture self-selected as participants. This potential limitation, however, was
not determined to be significant enough to warrant the use of deception in the study description.
Only MTurk workers who were United States residents were able to view and access the
study. There are instances, however, in which non-U.S. workers obtain a U.S. MTurk account in
order to qualify for more HITs. Although MTurk has strict policies and enforcement against this,
it does happen, and researchers commonly conduct additional residency checks as a result. In this
50

study, the survey included a residency item in the demographic section and participants who
indicated they lived outside the U.S. were not included in the data analysis (n=1). As an
additional check (and again, one often utilized by researchers), the participants’ IP addresses
were reviewed to ensure their computer networks were located within the U.S.
23
Eight
participants returned foreign IP addresses. It is possible, of course, that these participants were
U.S. residents participating in the study while traveling abroad, however, to be conservative,
these participants’ data were not included in the data analysis. In order to decrease the likelihood
that non-U.S. residents would participate in the study, the survey was only posted on MTurk
during waking hours in the United States.
As described in the literature review, there are a number of advantages and limitations to
conducting studies on Amazon Mechanical Turk. In order to lessen the limitations related to data
integrity, researchers often follow a set of methodological practices in administering surveys on
MTurk. Many of these practices were utilized in the current study. For instance, in order to
increase the likelihood that participants would read each prompt and item carefully in an
unsupervised, anonymous research setting, only those workers with a track record of successfully
completing MTurk HITs (a 95% lifetime approval rating) were able to participate in the study.
Further, in an effort to encourage conscientious participation, the academic nature of the study
was emphasized, as was the importance of participants reading all the survey prompts and items.
As another data integrity measure, researchers using Mechanical Turk samples often insert
unique attention checks into their surveys. Attention checks are items that are simple to answer
appropriately if they are read, but if a participant is just selecting options randomly or not
reading items carefully, they will likely answer the question inappropriately. The data from those
participants who fail the attention check is then removed from the data analysis. In the current
study, the attention check, which was displayed approximately mid-way through the survey with
a group of demographic items, was:

23. The service used to check IP addresses was IP Location: “Hostname / Reverse IP Lookup,” IP Location,
accessed June 2014, http://www.iplocation.net/index.php. Since the accuracy of geolocation services can vary, this
website was used because it displays the results of four independent geolocation providers to identify the
geographical location of participants based on their IP addresses. Although the four providers sometimes returned
different cities or states, the country of origin was always consistent.
51

The following are award winning architects: Shigaru Ban, Frank Gehry, Thom Mayne,
Jean Nouvel, James Stirling, and Kenzo Tange. Based on this information, which of the
following is not an award winning architect?
a. James Stirling
b. Kenzo Tange
c. Geoffrey Poss
d. Frank Gehry

Ten respondents failed the attention check and their data was excluded from the analysis.
As previously discussed, the sample was stratified by age and gender; however, the
study’s demographic needs were not advertised in the recruitment messaging because, given the
anonymous nature of the research, workers could have misrepresented their age and gender in
order to qualify for the study. Many researchers who use Mechanical Turk, instead of advertising
their participation criteria, utilize screening questions. In the current study, after completing the
consent form and prior to beginning the survey, workers were asked to provide their MTurk ID
number, age, and gender.
24
Participants who met the criteria of a demographic pool that had
already been filled were redirected to a thank you page where they were informed that they were
ineligible for the study. Although workers were instructed they could only attempt the HIT once
and subsequent attempts would be denied, a review of the survey records indicated that 38
workers, after being told they were ineligible for the study, re-took the qualifying questions and
provided alternate demographic information in, what is presumed, an attempt to qualify. Six
were successful in providing the correct demographic information needed to proceed, and took
the survey. The data from these participants was excluded from the analysis since they
misrepresented their demographics.
Similarly, as an additional integrity check, the data was reviewed for duplicate
participants. One worker took the study twice despite the instructions that it could only be
completed once and that subsequent submissions would be denied. The second set of data from
this participant was removed from the data analysis.
As a final measure of integrity, popular online forums used by Mechanical Turk workers
were monitored to ensure that information that would bias or inform future participants was not
being shared.
25
In one forum, some participants described being asked to evaluate old and new
buildings; however, importantly, they did not share that later questions were specifically about

24. It should be noted that Amazon Mechanical Turk has very strict policies regarding workers having only one
MTurk account. When this policy is violated, it is grounds for removal from MTurk.
25. The forums that were monitored were MTurk Forum, TurkerNation, and Reddit - MTurk.
52

historic preservation, so future participants would be unaware of the specific research goals of
the study. Some participants additionally shared that there had been an attention check; however,
they did not indicate what it was or where it was in the survey. Future participants who had read
this information would know to look out for a check, but they would still have to read each
prompt and item with care to avoid missing it. Lastly, the forums were reviewed for
communication regarding the demographic criteria that were necessary to qualify for the study.
No such content was observed.
Two final notes related to the survey administration and the demographic quotas: first,
although the sample was stratified by age and gender to reflect the 2010 U.S. Census figures, due
to the low population of older adults on Amazon Mechanical Turk, it was anticipated that it
would not be possible to obtain an adequate proportion of adults over the age of 65.
26
As such,
respondents over the age of 55 were grouped together with the expectation that this group would
skew younger than the real U.S. population. Lastly, it should be noted that data from participants
who completed surveys after a demographic quota had been met was not analyzed. Two hundred
nineteen people completed the study (excluding those whose data was removed for the reasons
stated above); however, in order to duplicate the U.S. Census age and gender proportions, only
the data of the first 200 participants was analyzed.

3.3 Statistical Methods
As described in the introduction, the primary research questions for the study are: Do
people prefer older buildings or new? What are people’s attitudes about historic preservation? To
what extent do people engage in historic preservation-related activities? Are there demographic
trends in preference, attitude, and engagement? And, is there an association among preference,
attitudes, and engagement? The secondary research questions relate to the association between
the preference rating of buildings (both new and old) and the variables listed in Tables 3.2 and
3.3 (e.g., visual complexity, building condition, presence of people, number of stories, etc.). In
the following sections, the statistical methods used to address these questions are described. The
significance level of 0.05 was used throughout the analysis to determine statistical significance
where appropriate.

26. “American Fact Finder,” United States Census Bureau, accessed April 8, 2014,
http://factfinder2.census.gov/faces/nav/jsf/pages/community_facts.xhtml.
53

3.3.1 Primary Analysis
The analysis of the primary research questions began with a compilation of descriptive
statistics, including the means and frequencies of each of the survey items.

3.3.1.1 Preference for Old versus New Buildings
To answer the question, “Do people prefer old buildings or new?” the old and new
building scores were averaged across all participants to determine a mean old building score and
a mean new building score. A paired t-test was then conducted to determine if there was a
statistically significant difference between the two means.
As an alternative measure of preference for old or new buildings, a mean old and new
building score was calculated for each participant. A binary value was then assigned to each
participant to indicate the higher of their two building scores.
27
To determine if more participants
preferred old buildings or new based on the binary outcome, I tested whether the proportion of
participants who preferred old buildings was 0.5 or not. That is, if there was no difference in
preference for old and new buildings, the binary outcome would have a probability of 0.5. A
95% confidence interval (CI) for the proportion was calculated to account for the variability of
the estimated proportion and to determine whether 0.5 was a reasonable value for the proportion.
For further investigation, the buildings were ranked from one to 50 according to their
mean scores, and the proportion of old and new buildings in the top and bottom of the list was
observed. Lastly, the variability of the preference ratings for old and new buildings was assessed
by charting the distribution of the mean preference scores.
A sub-question within the preference section was: are there differences in preference for
old buildings from our recent past (i.e., post-World War II buildings)? To answer this question,
the 25 old buildings were subdivided as pre-war and post-war buildings and the difference in
their cumulative mean ratings was tested using the p-value of a t-test and a 95% CI.

3.3.1.2 Preservation Attitudes
The preservation attitude section of the study included community service items and
attitudinal statements. Regarding the former, in order to assess the importance of the nine

27. Four participants had equal old and new building scores. For the purposes of the binary analysis, and to be
conservative, these participants were included in the subset of participants whose average new building rating was
higher than their old building rating.
54

community services evaluated by participants, mean importance scores were calculated, as well
as the frequency with which participants indicated each service was somewhat to very important.
To determine if the findings replicated those of Daniel Levi, a t-test was used to test whether
preservation ranked as the second most important service following street trees and public
landscaping. Next, the estimated proportion of participants in Levi’s study who rated
preservation as important to very important (.88) was re-evaluated with 95% CIs.
Analysis of the attitudinal statements began with the calculation of means and
frequencies for each item.
28
Additionally, in order to account for uncertainties in the observed
proportions, 95% CIs were calculated for the proportion of participants who somewhat to
strongly agreed with each statement. For the two reverse scaled items, the regular scale was
matched by reversing the ratings (e.g. a one was reversed to a seven, a seven to a one, and so on).

3.3.1.3 Preservation Engagement
To evaluate the degree to which respondents engaged in historic preservation-related
activities, the frequency with which they reported participating in each of the nine preservation
activities was calculated and the types of activities that were most and least reported were
observed. The number of activities each respondent participated in and the number of
participants who participated in zero, one, two, or more activities were also calculated.
To understand the extent to which participants wanted to get involved in historic
preservation but faced obstacles in participating, the observed proportions of people who
somewhat to strongly agreed with the preservation obstacle items were reported with 95% CIs.

3.3.1.4 Demographic Trends
Associations between Demographic Variables and Preference for Buildings
To study the association between demographic variables and the mean score of the 25 old
buildings, simple linear regressions were used to quantify unadjusted associations between each
of the covariates and preference for old buildings.
29
Two multiple regression models were then
used to quantify adjusted associations. The first model included all the demographic covariates
of age, gender, ethnicity, education, urbanicity, political orientation, and region. In the second

28. The mean attitude rating was adjusted in this analysis and future analyses for the two reverse coded items.
29. Simple linear regressions were used throughout the study as a precursor to multiple regressions for their
potential ability to inform the identification of adjustment variables (i.e., confounders).
55

model, a number of covariates were excluded for parsimony in order to heighten the
interpretability of the data. The determination regarding which variables to include in the second
multiple regression was based on the results of the single regressions, as well as intuition and
knowledge about the existing literature. The same procedure was then performed with the mean
score for the 25 new buildings as the dependent variable.
A multiple linear regression was used to test for the association between the covariates
and the relative preference, which was defined as the difference in participants’ mean ratings for
old buildings and new buildings. As opposed to modeling the ratings of old buildings and the
ratings of new buildings, which can be thought of as absolute preference, this model tested for
the relative preference of old buildings versus new buildings. Each of the three models, thus,
addressed different questions of interest.

Associations between Demographic Variables and Preservation Attitudes
Simple linear regressions and a multiple linear regression were again used to test for
unadjusted and adjusted associations between the demographic variables and preservation
attitudes, where the dependent variable was the mean score for the nine attitudinal statements.

Associations between Demographic Variables and Engagement in Preservation
Two regression models were considered to test for an association between participants’
demographics and participation in the nine historic preservation activities. The first model was a
linear model, where engagement in preservation was defined as the number of activities
participated in during the last year. The number of activities was treated as a continuous
dependent variable. The second model was a logistic regression where a value of one was
assigned to respondents who participated in two or more activities in the last year and a zero
assigned to all others.

3.3.1.5 Pairwise Associations among Preference for Old Buildings, Preservation Attitudes, and
Engagement in Preservation
Association between Preference for Old Buildings and Preservation Attitudes
In order to assess a potential association between preference for old buildings and
preservation attitudes, a simple linear regression model was used with the average rating for old
56

buildings as the independent variable and the mean score for the nine attitudinal items as the
dependent variable. In this model, when there is a one point difference in the average rating of
old buildings, the associated average difference in the mean score for preservation attitude is
quantified.

Association between Preference for Old Buildings and Engagement in Preservation
A simple linear regression and a simple logistic regression were used to test for a
potential association between preference for old buildings and engagement in preservation-
related activities. Here, preference was considered the independent variable and engagement the
dependent variable. As previously described, engagement was quantified in two different ways:
continuously (as the number of activities selected) and binarily (whether a respondent
participated in two or more activities or not). In the logistic regression, an estimated odds ratio
was calculated for a one unit difference in preference to determine if participants who prefer old
buildings tend to have engaged in two or more preservation-related activities in the last year. An
odds ratio of one would indicate no association between the two variables, whereas a ratio of
greater than one would indicate a positive association and a ratio of less than one would indicate
a negative association.

Association between Preservation Attitudes and Engagement in Preservation
Testing for a possible association between preservation attitudes and engagement in
historic preservation activities followed the same modeling. In this analysis, the average score
for preservation attitudes was the independent variable, and the engagement in preservation the
dependent variable.

3.3.2 Secondary Analysis: Predictors of the Average Preference Rating of a Building
In order to test for an association between the average preference ratings of a building
and the age of a building (old or new), the primary and secondary predictors discussed in section
3.1.3 were adjusted in a multiple regression. After controlling for potential confounders, the
average difference in the preference ratings for old and new buildings was estimated. In addition,
an association modified by building complexity was also investigated to determine if, for
example, preference for a simple new building versus a simple old building is different than
57

preference for a complex new building versus a complex old building. Here, an interaction model
was used to test whether the average difference in the rating of old and new buildings depends
upon building complexity.

58

4. FINDINGS

4.1 Descriptive Statistics
The survey was administered on Amazon Mechanical Turk between June 16 and June 28,
2014. Two hundred U.S. adults participated with an average survey completion time of
approximately 12 minutes. The subjects’ demographic data is outlined in Table 4.1. Per the U.S.
Census figures, the number of male and female participants was roughly balanced. As previously
mentioned, although the sample was stratified by age to reflect the 2010 U.S. Census figures, due
to the low population of adults over the age of 65 on Mechanical Turk, those ages 55 and older
were grouped together into one age category. As can be seen in the table, most of the participants
over the age of 55 reported being between 55 and 64 years old. Only seven were over the age of
65. The majority of the participants were white – 83% – which is higher than the population of
white citizens in the U.S. (72%). Geographic diversity was good, with participants residing in a
variety of community types (urban, suburban, and rural) across 40 states in the U.S.: 40% were
from the South, 26% from the West, and 17% from both the Northeast and Midwest.
1
All the
participants had at least a high school education or equivalent and more than half had a
Bachelor’s degree or higher. More participants were politically liberal than moderate or
conservative. Lastly, approximately 6% of participants had professional work experience or
higher education in historic preservation, architecture, or urban design (n=8), or they were
involved in historic preservation activities within the community (n=3).

1. According to the 2010 U.S. Census, 19% of people reside in rural areas and the rest urban / suburban areas;
“Frequently Asked Questions,” United States Census Bureau, accessed June 27, 2014,
https://ask.census.gov/faq.php?id=5000&faqId=5971; Geographic regions in the study correspond to U.S. Census
regions: “Census Regions and Divisions of the United States,” United States Census Bureau, accessed June 19,
2014, https://www.census.gov/geo/maps-data/maps/pdfs/reference/us_regdiv.pdf.
59

Table 4.1: Descriptive statistics of participant demographics
Demographic N Category # %
Gender 200
Male 99 50%
Female 101 51%
Age 200
18-24 25 13%
25-34 36 18%
35-44 35 18%
45-54 38 19%
55-64 66 34%
65 or older 7 4%
Ethnicity 198
African American 11 6%
American Indian or Alaska Native 0 0%
Asian / Pacific Islander 8 4%
Hispanic / Latino 6 3%
White / Caucasian 164 83%
Other 1 1%
Biracial 8 4%
Education 200
Elementary / middle school 0 0%
Some high school 0 0%
High school graduate or GED 21 11%
Some college 44 22%
Trade / technical / vocational training 2 1%
Associate degree 25 13%
Bachelor’s degree 74 37%
Graduate degree 34 17%
Political orientation 200
Liberal 92 46%
Moderate 57 29%
Conservative 45 23%
Other 6 3%
Architecture, urban planning,
or historic preservation
background
198
No 186 94%
Yes 8 4%
Not sure 4 2%

60

Table 4.1 (Cont’d): Descriptive statistics of participant demographics
Demographic N Category # % Category # %
State 200
Alabama 1 1% Montana 1 1%
Alaska 0 0% Nebraska 1 1%
Arizona 5 3% Nevada 0 0%
Arkansas 4 2% New Hampshire 1 1%
California 21 11% New Jersey 6 3%
Colorado 7 4% New Mexico 1 1%
Connecticut 1 1% New York 10 5%
Delaware 0 0% North Carolina 9 5%
District of Columbia 0 0% North Dakota 1 1%
Florida 13 7% Ohio 5 3%
Georgia 8 4% Oklahoma 0 0%
Hawaii 0 0% Oregon 6 3%
Idaho 1 1% Pennsylvania 12 6%
Illinois 4 2% Puerto Rico 0 0%
Indiana 2 1% Rhode Island 1 1%
Iowa 3 2% South Carolina 8 4%
Kansas 0 0% South Dakota 0 0%
Kentucky 2 1% Tennessee 5 3%
Louisiana 2 1% Texas 15 8%
Maine 0 0% Utah 1 1%
Maryland 5 3% Vermont 0 0%
Massachusetts 3 2% Virginia 4 2%
Michigan 7 4% Washington 9 5%
Minnesota 2 1% West Virginia 1 1%
Mississippi 3 2% Wisconsin 6 3%
Missouri 3 2% Wyoming 0 0%
Demographic N Category # %
U.S. Region
(based on U.S.
Census regions)
200
Northeast 34 17%
Midwest 34 17%
South 80 40%
West 52 26%
Pacific (Hawaii /
Alaska) 0 0%

61

4.2 Statistical Analysis
4.2.1 Findings for the Primary Research Questions
4.2.1.1 Preference for Old versus New Buildings
Analysis of the mean old and new building preference ratings reveals that, on average,
participants preferred old buildings over new buildings. The mean rating participants assigned to
old buildings was 3.92, whereas the mean rating for new buildings was 3.14. (Figure 4.1) The
results of a paired t-test indicate that there is a statistically significant difference between these
two means at the 0.05 significance level (T = -3.12, p-value = 0.0047).

Figure 4.1: Mean preference ratings for old and new buildings.

To better understand the extent to which people prefer older building over new, each
participant’s new and old building means were calculated to determine the proportion of
participants who preferred new versus the number who preferred old. As illustrated in Figure 4.2,
participants were three times as likely to prefer old buildings. Seventy-five percent of
respondents preferred the old buildings; their mean old building score was higher than their mean
new building score. In contrast, 24% of participants preferred the new buildings and 1% liked
both old and new equally. A one-sample test for proportion of 0.5 indicates the findings are
statistically significant at the 0.05 level (z = -7.1, p-value = 1.5 x 10
-12
) and with 95% CI (0.69,
0.81).

3.14
3.92
1.00 2.00 3.00 4.00 5.00 6.00 7.00
New
Old
Preference (Mean)
62

Figure 4.2: Percent of participants who prefer old versus new
buildings.

As an additional way to explore preference, the buildings were ranked from one to 50 in
terms of their mean preference score (Appendix C). Figure 4.3 illustrates that 19 of the top 25
buildings (75%) were old buildings, whereas 75% of the bottom 25 buildings were new. When
considering the top 10 – i.e., the very most preferred buildings – nine out of the 10 were old.
Regarding the new building in the top 10, although the study was designed to exclude new
buildings that were designed to look old, this building arguably, and more so than any other new
building, has some neotraditional influences, which could explain its higher preference rating.
Whereas the top 10 most preferred buildings were predominantly old buildings, the bottom 10
was more mixed with six buildings being new and four being old. (Figure 4.4)

24.0%
75.0%
1.0%
Participants who
prefer new
Participants who
prefer old
Participants who
equally like new and
old
63

Figure 4.3: Number of old and new buildings in the top 25 and
bottom 25.

Figure 4.4: Number of old and new buildings in the top 10 and
bottom 10.

1
6
9
4
0
1
2
3
4
5
6
7
8
9
10
Top 10 Bottom 10
New
Old
6
19 19
6
0
2
4
6
8
10
12
14
16
18
20
Top 25 Bottom 25
New
Old
64

1. Old building. Mean: 5.35

2 (tied). New building. Mean: 5.07

2 (tied). Old building. Mean: 5.07

4. Old building. Mean: 5.02

5. Old building. Mean: 4.93

6 (tied). Old building. Mean: 4.77
Figures 4.5-4.10: Top 10 most preferred buildings. Photo credits from top left: Adam Smith, Brandon
Bartoszek, Brandon Bartoszek, Adam Smith, Marc Belanger, Paige Miller.
65

6 (tied). Old building. Mean: 4.77

8. Old building. Mean: 4.76

9. Old building. Mean: 4.61

10. Old building. Mean: 4.58
Figures 4.11-4.14: Top 10 most preferred buildings. Photo credits from top left: Flickr – OzinOH, Phil
Squattrito, Phil Squattrito, John Chambers, Jr.

66

50. Old building. Mean: 1.90

49. Old building. Mean: 2.05

48. Old building. Mean: 2.16

47. New building. Mean: 2.31

46. New building. Mean: 2.43

45. New building. Mean: 2.57

Figures 4.15-4.20: Bottom 10 preferred buildings. Photo credits from top left: Sandra Shannon, Sandra
Shannon, Sandra Shannon, Robert Benson / 4240 Architecture, Sandra Shannon, Yichen Lin.
67

44. New building. Mean: 2.58

43. New building. Mean: 2.62

42. Old building. Mean: 2.79

41. New building. Mean: 2.84
Figures 4.21-4.24: Bottom 10 preferred buildings. Photo credits from top left: Sandra Shannon, Sandra
Shannon, Brandon Bartoszek, Sandra Shannon.

68

0
10
14
0
1
0
1
3
9 9
3
0
0
2
4
6
8
10
12
14
16
1 - 1.99 2 - 2.99 3 - 3.99 4 - 4.99 5 - 5.99 6 - 7
# of Images
Preference Range (Means)
New
Old
Although on average people preferred the old buildings more than the new buildings,
there was higher variability in the ratings of the old buildings. Figure 4.25 illustrates this. In this
histogram, the seven point scale used to assess preference was broken down into six roughly one-
point “buckets.” The buckets tell us the number of buildings with a mean preference score of
1.00 to 1.99, 2.00 to 2.99, 3.00 to 3.99, and so on. The histogram illustrates that whereas there is
variability in the distribution of means within the old building set, there is more consistency
among the new buildings. With the exception of just one building, all the new buildings have a
mean between 2.00 and 3.99, indicating agreement among people regarding how they felt about
new buildings in general. In other words, they liked each of the new buildings about the same
amount, give or take one point. In contrast, whereas most of the old buildings had a mean
between 3.00 and 4.99, seven of the old buildings elicited a stronger positive or negative
response. The box plots in Figures 4.26 and 4.27 provide further illustration of the variability of
this data.

Figure 4.25: Number of buildings per mean preference range.

69

Figure 4.26: Boxplot of mean preference ratings: new buildings.

Figure 4.27: Boxplot of mean preference ratings: old buildings.

Post-War Buildings
As illustrated in Table 4.2, when the old buildings were separated into two categories –
pre-WWII and post-WWII resources – the results indicate that post-war buildings were not only
considerably less preferred than pre-war buildings, but they were also less preferred than new
buildings.
2
These differences showed very strong statistical significance.
3
Further, when the

2. Post-war buildings include old buildings #7, 13, 18, and 24.
1 2 3 4 5 6 7
Preference (Mean)
1 2 3 4 5 6 7
Preference (Mean)
70

images were ranked from one to 50 in terms of their mean rating, three of the four post-war
buildings were, in fact, in the bottom 10 with means between just 2.0 and 2.8. So, although
participants preferred old buildings over new in general, there was a considerable difference in
their preferences for pre-war and post-war buildings.

Table 4.2: Mean preference rating of pre-World War II,
post-World War II, and new buildings (7-point scale)
Stimuli category Mean N
Pre-war old buildings 4.18 21
New buildings 3.15 25
Post-war old buildings 2.60 4

4.2.1.2 Preservation Attitudes
This section of the survey included two sets of items: the importance of community
services item and the preservation attitude items. In general, as will be discussed in more detail
below, the participants showed strong support for historic preservation.

Importance of Community Services
The mean importance rating for historic preservation was 5.68 (on a seven point scale),
indicating that people generally find the service to be rather important. When put in the
perspective of the importance of other community services, preservation was ranked as the third
most important service behind street trees / public landscaping, followed by economic
development. Based on the point estimate, economic development was rated higher than historic
preservation by 0.07; however, this difference was not statistically significant, meaning that we
cannot conclude with certainty whether economic development or historic preservation was
ranked higher. (Table 4.3)
4

3. Pre-war vs. post-war (z=1.6, p-value=1.3 x 10-39) with 95% CI (1.4, 1.7); pre-war vs. new (z=1.0, p-value =
1.38 x 10-23) with 95% CI (0.8, 1.2); post-war vs. new (z=0.5, p-value=2.2 x 10-20) with 95% CI (0.4, 0.6).
4. The p-value for testing the mean difference between economic development and historic preservation was 0.61.
71

Table 4.3: Mean ratings of community services (7-point scale)
Community Services Mean
Street trees and public landscaping 5.85
Economic development 5.75
Historic preservation 5.68
Cultural events 5.13
Recreation programs 5.13
Architectural review for aesthetics of new construction 5.07
Public art 4.83
Tourism promotion 4.27

When the proportion of participants who rated each service as somewhat to very
important was examined, these findings, as illustrated in Table 4.4, seemed to be consistent with
the results of Daniel Levi’s study.
5
Eighty-three percent of people identified historic preservation
as somewhat to very important in the current study whereas 88% rated preservation as important
to very important in Levi’s study. A 95% confidence interval for the proportion of participants in
the current study who rated historic preservation as somewhat to very important was (0.77, 0.88),
indicating consistency with the previous findings. When the services were ranked based on the
observed proportions (rather than means), historic preservation was still ranked third, but now
economic development was first and street trees / public landscaping was second by a difference
of 1%. In Levi’s study, in contrast, historic preservation was second behind street trees / public
landscaping and economic development was third. Although it appears that street trees and
public landscaping, economic development, and historic preservation are the most important
community services to people, the order of importance is still unclear.
6

5. As previously discussed in section 3.1.1, although a slightly different importance scale was used in Levi’s study
(4-point versus 7-point), general conclusions can still be drawn regarding the consistency of the data.
6. Levi, “Does History Matter?” 148-159.
72

Table 4.4: Ranking of the importance of community service items
Current study Daniel Levi’s study
Community Service Item
% of ratings for
“somewhat
important” to
“very important” Community Service Item
% of ratings for
“important” to
“very important”
Economic development 87% Street trees and public landscaping 92%
Street trees and public landscaping 86% Historic preservation 88%
Historic preservation 83% Economic development 81%
Recreation programs 76% Cultural events 80%
Architectural review for aesthetics of
new construction
74% Recreation programs 76%
Cultural events 72% Architectural review for aesthetics
of new construction
70%
Public art 66% Public art 61%
Tourism promotion 53% Tourism promotion 49%
Daniel J. Levi, “Does History Matter? Perceptions and Attitudes Toward Fake Historic Architecture and Historic Preservation,”
Journal of Architectural and Planning Research 22:2 (2005): 148-159.

Attitudinal Statements
Mean and frequency analysis of the attitudinal statements revealed that respondents were
generally supportive of historic preservation. (Tables 4.5 and Figures 4.28-4.30) Eighty-eight
percent somewhat to strongly agreed that historic places should be treated as community assets
and a majority (64%) believe that too little is done to protect historic places in the U.S.
7
Further,
most participants have personally been affected by the loss of an historic resource: nearly three
quarters indicated that they have been saddened by the demolition of an historic place.
8

Participants additionally believe that preserving historic outdoor spaces is as important as
preserving historic buildings (87% somewhat to strongly agreed), and they think that the
preservation or re-use of historic places is environmentally friendly (82% somewhat to strongly
agreed).
9
Lastly, participants generally do not see historic preservation as an activity that
interferes with economic development. Seventy-three percent somewhat to strongly disagreed
with the statement: historic preservation gets in the way of economic development.
10

Although respondents were generally supportive of historic preservation, there was
inconsistency in terms of how they felt about preservation-related regulation. Whereas a large

7. The estimated proportion was .884 with 95% CI of (.839, .928). Sixty-four is the percent of participants who
somewhat to strongly agree that we do too little to protect historic places in the U.S.; 95% CI (.576, .710).
8. Seventy-three is the percent of participants who somewhat to strongly agree with the statement: I have been
saddened by the demolition of an historic place; 95% CI (.667, .792).
9. The estimated proportions were .873 and .817 with 95% CIs (.827, .920) and (.763, .871), respectively.
10. The estimated proportion was .732 with 95% CI (.671, .794).
73

majority of respondents (84%) somewhat to strongly agreed that local governments should have
the ability to prevent the demolition of historic places, when regulation was framed in the context
of personal property rights, a smaller majority were supportive.
11
Only 67% of participants
somewhat to strongly disagreed with the statement: owners of historic buildings should be able
to demolish their building without any kind of legal restrictions.
12

Lastly, although support for preservation was generally strong and people think of it as a
community asset and an important community service, not everyone wants to pay for it. Only
57% of participants somewhat to strongly agreed that tax payer dollars should be used to help
preserve privately owned historic places because the activity benefits the public.
13

Table 4.5: Mean agreement ratings for the nine preservation attitude items (7-point scale)
Attitudinal Item
Mean Agreement
Rating
Historic places should be treated as community assets 5.7

We do too little to protect historic places in the U.S. 4.9

I have been saddened by the demolition of an historic place 5.3

Preserving historic outdoor spaces is as important as preserving historic buildings

5.7
The preservation or re-use of historic places is environmentally friendly

5.6
Historic preservation gets in the way of economic development

2.8*
Local governments should have the ability to prevent the demolition of historic places

5.5
Owners of historic buildings should be able to demolish their building without any kind
of legal restrictions

2.9*
Tax payer dollars should be used to help preserve privately owned historic places
because this activity benefits the public

4.5
Ratings based on a 7-point scale.
*Because these two statements were reverse coded, lower means indicate greater support for historic preservation.

11. The estimated proportion was .838 with 95% CI (.787, .890).
12. The estimated proportion was .670 with 95% CI (.604, .736).
13. The estimated proportion was .571 with 95% CI (.502, .640).
74

Figure 4.28: Agreement with preservation attitude statements 1-3.

Figure 4.29: Agreement with preservation attitude statements 4-6.

1% 1%
14%
2% 1%
39%
4%
3%
20%
7%
14%
12%
21%
18%
9%
45%
43%
6%
22% 21%
0%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Preserving historic
outdoor spaces is as
important as
preserving historic
buildings
The preservation or
re-use of historic
places is
environmentally
friendly
Historic preservation
gets in the way of
economic
development
Strongly agree
Agree
Somewhat agree
Neither agree nor disagree
Somewhat disagree
Disagree
Strongly disagree
1%
2%
3% 2%
5%
6%
3%
13%
4%
7%
17%
14%
20%
22%
18%
46%
28%
26%
22%
14%
29%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Historic places should
be treated as
community assets
We do too little to
protect historic places
in the U.S.
I have been saddened
by the demolition of
an historic place
Strongly agree
Agree
Somewhat agree
Neither agree nor disagree
Somewhat disagree
Disagree
Strongly disagree
75

Figure 4.30: Agreement with preservation attitude statements 7-9.

4.2.1.3 Preservation Engagement
Seventy-nine percent of respondents participated in one or more preservation-related
activities in the last year. For a large majority (72%), this meant that they had visited a place
because it was historic. Forty-one percent of participants did something preservation related in
the last year other than visiting an historic place. Commonly, this involved tasks that are
relatively simple to do and require few resources, including liking or sharing a post on social
media related to the preservation of historic place (23%), following a preservation group on
social media (12%), or signing a petition in support of the preservation of an historic place
(15%). Only one to 4% of respondents reported engaging in the types of preservation activities
that require more time or resources, including: nominating an historic place for local, state, or
national historic designation (1%); volunteering for a preservation organization (2%); purchasing
a membership to an historic preservation organization (3%); advocating for the preservation of
an historic place at a public hearing (3%); or contacting a public official to support the
preservation of an historic place (4%). Most respondents, further, appear to be casually involved
with historic preservation, with only 13% having participated in three or more preservation
2%
22%
5% 3%
24%
9%
3%
20%
8%
9%
15%
21%
24%
9%
28%
38%
7%
21%
22%
2%
8%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Local governments
should have the
ability to prevent the
demolition of historic
places
Owners of historic
buildings should be
able to demolish their
building without any
kind of legal
restrictions
Tax payer dollars
should be used to
help preserve
privately owned
historic places
because this activity
benefits the public
Strongly agree
Agree
Somewhat agree
Neither agree nor disagree
Somewhat disagree
Disagree
Strongly disagree
76

activities in the last year. The majority (67%) participated in just one or two activities. (Table 4.6
and Figures 4.31-4.32)

Table 4.6: Percent of respondents who participated in each preservation activity in the last year
Activity %
Visited a place because it was historic

72.2%
"Liked" or "shared" a post on social media related to the preservation of an historic place

22.7%
Followed an historic preservation organization on social media (E.g., National Trust for
Historic Preservation, Docomomo, Cultural Landscape Foundation, Preservation Texas, Los
Angeles Conservancy, etc.)

12.1%
Signed a petition for the preservation of an historic place

14.6%
Contacted a public official to support the preservation of an historic place

4.0%
Advocated for the preservation of an historic place at a public hearing

2.5%
Paid for a membership to an historic preservation organization

3.0%
Volunteered for an historic preservation organization

1.5%
Nominated an historic place for local, state, or national historic designation

1.0%

77

Figure 4.31: Number of participants who engaged in 0, 1, or more
preservation activities in the last year.

Figure 4.32: Number of participants who engaged in 0, 1, or more
preservation activities in the last year (excluding visiting an historic
site).

Fifty percent of participants somewhat to strongly agreed with the statement: I would like
to support historic preservation efforts in my community, but I do not have the time. Likewise,
37% indicated that they would like to support preservation in their community, but do not know
how to get involved. (Table 4.7) Although 41% of participants engaged in some sort of
preservation-related activity in the last year (excluding those who only visited an historic site),
41
87
45
19
2
3
1
0 20 40 60 80 100
0
1
2
3
4
5
6
# of respondents
# of Activities
117
53
20
4
3
1
0 20 40 60 80 100 120 140
0
1
2
3
4
5
# of respondents
# of Activities
78

interestingly, when asked about obstacles to getting involved in historic preservation in their
community, only 1.5% of participants indicated that they were already involved.

Table 4.7: Obstacles to getting involved in historic preservation efforts
Survey Item % Somewhat to Strongly Agree
I would like to support historic preservation efforts in my
community, but I do not have the time.

49.5%
I would like to support historic preservation efforts in my
community, but I do not know how to get involved.

37.4%

4.2.1.4 Demographic Trends
Associations between Demographic Variables and Preference for Buildings
When an association was tested between the demographic variables and the mean
preference score of the 25 old buildings, the primary finding was a positive and statistically
significant association between participant age and preference for old buildings. That is, as
participant age increased, preference for old buildings increased. Notable findings are listed
below and presented in Table 4.8.
14

• In the unadjusted model, participants ages 55 and up, on average, had an estimated
mean old building score 0.33 points higher than people ages 18 to 34 with 95% CI
(0.058, 0.608).
• In the unadjusted model, participants ages 45 to 54 and ages 55+, on average, had an
estimated mean old building score roughly half a point higher than people ages 18-
24 (

= .47 with 95% CI (0.072, 0.870) and

= .50 with 95% CI (0.135, 0.862)
respectively).
• In the unadjusted model, when the age group was treated as a continuous variable,
the estimated mean difference in the preference for old buildings was 0.11 with 95%
CI (0.029, 0.182) when any two adjacent age groups were compared (ages 18-24, 25-
34, 35-44, 45-54, and 55+).

14. Several different models were used to investigate the association between participant age and preference. In
model 1, participant age was classified into three categories: young (ages 18-34), middle (ages 35-54), and old (ages
55+). The reference group was young. In model 2, age was classified into five categories: ages 18-24 (the reference
group), 25-34, 35-44, 45-54, and 55+. In model 3, age was continuous. In model 10, which was an adjusted model,
age was again defined as young, middle, and old, with young being the reference group. In model 11, another
adjusted model, age was defined continuously.
79

• In the adjusted model, participants ages 55 and up, on average, had an estimated
mean old building score 0.29 points higher than people ages 18-34 of the same
gender, education, and political orientation with 95% CI (-0.003, 0.582).
With regard to the other demographic variables, although several showed directional tendencies
with preference for old buildings, statistical significance was not achieved (see model 10 in
Table 4.8).

80

Table 4.8: Association between demographics and preference for old buildings. Models 1-9: simple linear
regressions. Models 10-11: multiple linear regressions.
Model
Reference
Group
Variable Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
1
Young
(18-34)
(Intercept) 3.713 0.101 3.514 3.911 36.683 < 0.0001
Age: Middle (35-54) 0.258 0.137 -0.011 0.527 1.880 0.0616
Age: Old (55+) 0.333 0.140 0.058 0.608 2.370 0.0187*
2 18-24
(Intercept) 3.547 0.158 3.237 3.857 22.445 < 0.0001
Ages: 25-34 0.281 0.206 -0.123 0.684 1.364 0.1742
Ages: 35-44 0.372 0.207 -0.034 0.778 1.798 0.0738
Ages: 45-54 0.471 0.203 0.072 0.870 2.314 0.0217*
Ages: 55+ 0.498 0.186 0.135 0.862 2.686 0.0079*
3
(Intercept) 3.450 0.182 3.093 3.806 18.964 < 0.0001
Age: Continuous 0.106 0.039 0.029 0.182 2.695 0.0076*
4 Male
(Intercept) 3.854 0.080 3.696 4.011 48.033 < 0.0001
Gender: Female 0.125 0.113 -0.096 0.346 1.108 0.2692
5 Not college grad
(Intercept) 3.824 0.097 3.632 4.015 39.222 < 0.0001
Education: College grad+ 0.140 0.120 -0.094 0.374 1.172 0.2425
6 Not white
(Intercept) 3.817 0.137 3.548 4.085 27.827 < 0.0001
Race: White 0.128 0.151 -0.167 0.424 0.852 0.3951
7 Urban
(Intercept) 3.912 0.103 3.711 4.113 38.083 < 0.0001
Community: Suburban -0.081 0.132 -0.339 0.177 -0.616 0.5388
Community: Rural 0.219 0.159 -0.093 0.530 1.375 0.1708
Community: Other 0.008 0.471 -0.915 0.931 0.017 0.9863
8 Liberal
(Intercept) 3.815 0.083 3.652 3.978 45.831 < 0.0001
Political: Moderate 0.143 0.135 -0.120 0.407 1.065 0.2881
Political: Conservative 0.240 0.145 -0.045 0.524 1.649 0.1007
Political: Other 0.239 0.336 -0.421 0.898 0.709 0.4791
9 South
(Intercept) 3.895 0.090 3.718 4.071 43.336 < 0.0001
Region: Northeast 0.066 0.165 -0.256 0.389 0.404 0.6867
Region: Midwest 0.085 0.165 -0.237 0.408 0.519 0.6040
Region: West -0.014 0.143 -0.295 0.267 -0.098 0.9223
10
(Intercept) 3.518 0.206 3.114 3.922 17.072 < 0.0001
Young
(18-34)
Age: Middle (35-54) 0.201 0.147 -0.088 0.490 1.366 0.1737
Age: Old (55+) 0.289 0.149 -0.003 0.582 1.937 0.0543
Male Gender: Female 0.069 0.121 -0.168 0.306 0.574 0.5669
Not college grad Education: College grad+ 0.089 0.126 -0.158 0.336 0.706 0.4812
Not-white Race: White -0.001 0.160 -0.315 0.314 -0.004 0.9970
Urban
Community: Suburban -0.102 0.138 -0.372 0.169 -0.736 0.4628
Community: Rural 0.135 0.175 -0.207 0.478 0.775 0.4392
Community: Other -0.201 0.499 -1.180 0.777 -0.403 0.6872
Liberal
Political: Moderate 0.127 0.137 -0.142 0.396 0.927 0.3552
Political: Conservative 0.210 0.153 -0.090 0.511 1.370 0.1724
Political: Other 0.354 0.354 -0.340 1.048 1.000 0.3188
South
Region: Northeast 0.133 0.172 -0.205 0.470 0.772 0.4411
Region: Midwest 0.149 0.171 -0.187 0.485 0.867 0.3873
Region: West 0.097 0.150 -0.197 0.391 0.647 0.5182
11
(Intercept) 3.269 0.206 2.865 3.674 15.835 < 0.0001
Age: Continuous 0.095 0.040 0.016 0.174 2.355 0.0195*
Male Gender: Female 0.125 0.113 -0.096 0.346 1.105 0.2706
Not college grad Education: College grad+ 0.119 0.119 -0.115 0.352 0.996 0.3205
Liberal
Political: Moderate 0.124 0.133 -0.136 0.385 0.934 0.3513
Political: Conservative 0.180 0.147 -0.108 0.468 1.227 0.2213
Political: Other 0.340 0.335 -0.316 0.997 1.015 0.3112

*Associations significant at the 0.05 level

81

When associations between the demographic covariates and preference for new buildings
were investigated, evidence for an association with participant age was again found. (Table 4.9)
Here, however, the relationship was negative instead of positive. That is, as age increased,
preference for new buildings decreased. Main findings include:
• In the unadjusted model, respondents ages 55 and older rated new buildings, on
average, .38 points less than respondents ages 18 to 24 with 95% CI (-0.708, -0.052);
and
• In the unadjusted model, when the age group was treated as a continuous variable, the
estimated mean difference in the preference for new buildings was -.11 with 95% CI
(-0.203, -0.020) between any two adjacent age groups (ages 18-24, 25-34, 35-44, 45-
54, and 55+).
A statistically significant association between gender and preference for new buildings was also
found. When the covariates listed in Table 4.9 were controlled, men were more likely to prefer
new buildings than women. They rated new buildings .28 points higher on average with 95% CI
(-0.564, -0.003) than female participants of the same age, ethnicity, education, etc (see model 10
in Table 4.9). Regarding the other demographic covariates, none of the tested associations
produced statically significant results (see models 10 and 11 in Table 4.9).

82

Table 4.9: Association between demographics and preference for new buildings. Models 1-9: simple linear
regressions. Models 10-11: multiple linear regressions.
Model
Reference
Group
Variable Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
1
Young
(18-34)
(Intercept) 3.348 0.121 3.111 3.584 27.761 < 0.0001
Age: Middle (35-54) -0.217 0.163 -0.537 0.103 -1.330 0.1851
Age: Old (55+) -0.380 0.167 -0.708 -0.052 -2.271 0.0242*
2 18-24
(Intercept) 3.386 0.189 3.015 3.756 17.906 < 0.0001
Ages: 25-34 -0.064 0.246 -0.547 0.418 -0.262 0.7936
Ages: 35-44 -0.178 0.248 -0.663 0.307 -0.719 0.4733
Ages: 45-54 -0.327 0.243 -0.804 0.150 -1.342 0.1812
Ages: 55+ -0.418 0.222 -0.853 0.017 -1.882 0.0613
3
(Intercept) 3.635 0.217 3.210 4.061 16.748 < 0.0001
Age: Continuous -0.111 0.047 -0.203 -0.020 -2.383 0.0181*
4 Male
(Intercept) 3.255 0.095 3.068 3.441 34.261 < 0.0001
Gender: Female -0.221 0.134 -0.483 0.041 -1.654 0.0997
5 Not college grad
(Intercept) 3.204 0.116 2.976 3.431 27.582 < 0.0001
Education: College grad+ -0.091 0.142 -0.370 0.188 -0.641 0.5224
6 Not-white
(Intercept) 3.428 0.162 3.111 3.745 21.169 < 0.0001
Race: White -0.351 0.178 -0.699 -0.002 -1.971 0.0501
7 Urban
(Intercept) 3.175 0.124 2.933 3.417 25.688 < 0.0001
Community: Suburban -0.069 0.159 -0.380 0.242 -0.436 0.6635
Community: Rural 0.019 0.191 -0.356 0.394 0.101 0.9200
Community: Other -0.322 0.566 -1.432 0.788 -0.568 0.5706
8 Liberal
(Intercept) 3.156 0.098 2.963 3.348 32.068 < 0.0001
Political: Moderate 0.160 0.159 -0.152 0.471 1.003 0.3171
Political: Conservative -0.269 0.172 -0.606 0.067 -1.570 0.1181
Political: Other 0.084 0.398 -0.695 0.864 0.212 0.8320
9 South
(Intercept) 3.060 0.107 2.851 3.269 28.688 < 0.0001
Region: Northeast 0.122 0.195 -0.261 0.504 0.623 0.5340
Region: Midwest 0.152 0.195 -0.231 0.535 0.777 0.4381
Region: West 0.140 0.170 -0.193 0.473 0.824 0.4111
10
(Intercept) 3.610 0.244 3.133 4.088 14.824 < 0.0001
Young
(18-34)
Age: Middle (35-54) -0.176 0.174 -0.517 0.166 -1.008 0.3148
Age: Old (55+) -0.284 0.176 -0.630 0.062 -1.610 0.1092
Male Gender: Female -0.283 0.143 -0.564 -0.003 -1.982 0.0490*
Not college grad Education: College grad+ -0.052 0.149 -0.343 0.240 -0.348 0.7284
Not-white Race: White -0.320 0.190 -0.692 0.052 -1.688 0.0931
Urban
Community: Suburban -0.030 0.163 -0.350 0.290 -0.184 0.8546
Community: Rural 0.226 0.206 -0.179 0.630 1.094 0.2755
Community: Other 0.022 0.590 -1.135 1.179 0.037 0.9704
Liberal
Political: Moderate 0.172 0.162 -0.145 0.490 1.063 0.2892
Political: Conservative -0.184 0.181 -0.539 0.171 -1.017 0.3106
Political: Other 0.029 0.419 -0.791 0.850 0.070 0.9439
South
Region: Northeast 0.119 0.204 -0.280 0.517 0.582 0.5610
Region: Midwest 0.124 0.203 -0.273 0.521 0.611 0.5417
Region: West 0.173 0.177 -0.174 0.521 0.978 0.3293
11
(Intercept) 3.720 0.245 3.240 4.200 15.200 < 0.0001
Age: Continuous -0.096 0.048 -0.190 -0.002 -2.009 0.0459*
Male Gender: Female -0.215 0.134 -0.477 0.047 -1.608 0.1094
Not college grad Education: College grad+ -0.069 0.141 -0.346 0.208 -0.486 0.6276
Liberal
Political: Moderate 0.175 0.158 -0.134 0.484 1.111 0.2679
Political: Conservative -0.211 0.174 -0.551 0.130 -1.211 0.2275
Political: Other -0.042 0.397 -0.820 0.737 -0.106 0.9160
*Associations significant at the 0.05 level

83

There are a number of interesting findings from the relative preference model, which
tested for an association between demographics and the difference between average old building
ratings and average new building ratings (Table 4.10). As we saw in the absolute preference
models, older people, on average, preferred old buildings (relative to new buildings) more than
younger people. In this model, however, there was stronger statistical significance. When older
participants (ages 55+) were compared with younger participants (ages 18-34), the estimated
mean difference was 0.71 with a p-value of 0.0001 under the unadjusted model, and 0.57 with a
p-value of 0.0029 under the adjusted model. Likewise, in this model there was increased
statistical significance for an association between gender and preference. That is, women, on
average, preferred old buildings more than men. Here, the estimated mean difference was
approximately 0.35 in both the unadjusted and adjusted models (with p-values of 0.0205 and
0.0229, respectively). (Recall that in the absolute preference model, statistical significance could
not be achieved at the 0.05 level.) An additional finding from the relative preference model is
that, on average, politically conservative people preferred old buildings (relative to new
buildings) more than those who identified as liberals. The estimated mean difference was 0.51
and 0.39 in the unadjusted and adjusted models, respectively (with p-values of 0.0081 and
0.0445).
As seen in Figure 4.33, participants’ average ratings for old buildings and for new
buildings were positively correlated. Because of this positive correlation, when relative
preference is calculated, there is less variability in these values than in the variability of the
separate ratings for old and new buildings. This is one possible reason why there was stronger
statistical significance in the preference model in comparison to the first two models described in
this section.

84

Figure 4.33: Participants’ rating for old and new buildings.

Mean score for old buildings
Mean score for new buildings
85

Table 4.10: Association between demographics and relative preference (old – new). Models 1-9: simple linear
regressions. Models 10-11: multiple linear regressions.
Model
Reference
Group
Variable Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
1
Young
(18-34)
(Intercept) 0.365 0.131 0.108 0.623 2.782 0.0059
Age: Middle (35-54) 0.475 0.178 0.126 0.824 2.671 0.0082*
Age: Old (55+) 0.713 0.182 0.356 1.070 3.913 0.0001*
2 18-24
(Intercept) 0.162 0.205 -0.240 0.563 0.789 0.4309
Ages: 25-34 0.345 0.266 -0.177 0.867 1.295 0.1969
Ages: 35-44 0.550 0.268 0.024 1.075 2.051 0.0416*
Ages: 45-54 0.797 0.264 0.281 1.314 3.025 0.0028*
Ages: 55+ 0.916 0.240 0.445 1.387 3.812 0.0002*
3
(Intercept) -0.185 0.235 -0.647 0.276 -0.788 0.4314
Age: Continuous 0.217 0.051 0.118 0.316 4.282 < 0.0001*
4 Male
(Intercept) 0.599 0.105 0.392 0.806 5.682 < 0.0001
Gender: Female 0.346 0.148 0.055 0.637 2.334 0.0206*
5 Not college grad
(Intercept) 0.620 0.129 0.367 0.873 4.799 < 0.0001
Education: College grad+ 0.231 0.158 -0.079 0.542 1.461 0.1457
6 Not-white
(Intercept) 0.389 0.179 0.037 0.740 2.166 0.0315
Race: White 0.479 0.197 0.093 0.866 2.430 0.0160*
7 Urban
(Intercept) 0.737 0.137 0.467 1.006 5.360 < 0.0001
Community: Suburban -0.012 0.176 -0.358 0.333 -0.068 0.9456
Community: Rural 0.199 0.213 -0.218 0.616 0.937 0.3500
Community: Other 0.330 0.630 -0.904 1.564 0.524 0.6010
8 Liberal
(Intercept) 0.659 0.109 0.445 0.873 6.043 < 0.0001
Political: Moderate -0.016 0.176 -0.362 0.329 -0.092 0.9267
Political: Conservative 0.509 0.190 0.136 0.882 2.674 0.0081*
Political: Other 0.154 0.441 -0.710 1.018 0.349 0.7272
9 South
(Intercept) 0.835 0.119 0.601 1.068 6.997 < 0.0001
Region: Northeast -0.055 0.218 -0.483 0.373 -0.253 0.8007
Region: Midwest -0.066 0.218 -0.494 0.362 -0.304 0.7618
Region: West -0.154 0.190 -0.526 0.218 -0.810 0.4188
10
(Intercept) -0.092 0.262 -0.605 0.421 -0.352 0.7256
Young (18-34)
Age: Middle (35-54) 0.377 0.187 0.010 0.744 2.012 0.0457*
Age: Old (55+) 0.573 0.190 0.201 0.945 3.021 0.0029*
Male Gender: Female 0.353 0.154 0.051 0.654 2.295 0.0229*
Not college grad Education: College grad+ 0.141 0.160 -0.173 0.454 0.879 0.3806
Not-white Race: White 0.319 0.204 -0.080 0.719 1.567 0.1189
Urban
Community: Suburban -0.072 0.175 -0.415 0.272 -0.408 0.6835
Community: Rural -0.090 0.222 -0.525 0.345 -0.407 0.6844
Community: Other -0.223 0.634 -1.467 1.020 -0.352 0.7253
Liberal
Political: Moderate -0.045 0.174 -0.387 0.297 -0.259 0.7957
Political: Conservative 0.394 0.195 0.012 0.776 2.024 0.0445*
Political: Other 0.325 0.450 -0.557 1.207 0.721 0.4717
South
Region: Northeast 0.014 0.219 -0.414 0.443 0.066 0.9475
Region: Midwest 0.025 0.218 -0.402 0.452 0.113 0.9099
Region: West -0.076 0.191 -0.450 0.297 -0.400 0.6895
11
(Intercept) -0.451 0.261 -0.963 0.061 -1.725 0.0861
Age: Continuous 0.191 0.051 0.091 0.291 3.741 0.0002
Male Gender: Female 0.340 0.143 0.060 0.619 2.379 0.0184*
Not college grad Education: College grad+ 0.187 0.151 -0.108 0.483 1.242 0.2159
Liberal
Political: Moderate -0.051 0.168 -0.381 0.279 -0.303 0.7625
Political: Conservative 0.391 0.186 0.026 0.755 2.103 0.0368*
Political: Other 0.382 0.424 -0.449 1.214 0.901 0.3688
*Associations significant at the 0.05 level

86

Association between Demographic Variables and Preservation Attitudes
As can be seen in regression models one through nine in Table 4.11 (the unadjusted
analyses), the only finding of significance was a negative association between preservation
attitudes and political conservativeness. That is, participants who identified themselves as
politically conservative had lower preservation attitude ratings on average (-.37 units) than
participants who indicated they were politically liberal.
In model 10 (Table 4.12), which controlled for covariates, there were several interesting
findings. Whereas participant age and gender were strong predictors of preference for buildings,
in this analysis, they were not found to have an association with preservation attitudes. Also, the
findings indicate that white people, on average, had higher preservation attitude scores than
people of other ethnicities of the same age, gender, education, etc. It is important to note,
however, that 83% of participants were white, therefore the sample does not sufficiently reflect
the preservation attitudes of other ethnicities and this finding should be interpreted with caution.
Although statistical significance has not quite been met, there is evidence of an association
between urbanicity and preservation attitudes. Specifically, participants who resided in urban
areas had higher attitudinal scores on average than people of the same covariates who resided in
rural areas (the estimated mean difference was 0.39 with a p-value of .0551). Lastly, and again,
though statistical significance could not be achieved, politically conservative people, although
they showed higher preference for old buildings, had lower preservation attitude scores on
average when compared to liberals of the same covariates (the estimated mean difference was
0.31 with a p-value of .0806).
In model 12 (Table 4.12), which provided the most compelling information, preference
for old buildings was treated as a confounder between the demographic variables and
preservation attitude. So here, preference for old buildings was held constant in addition to
controlling for the covariates of age, gender, education, ethnicity, urbanicity, political affiliation,
and region. In this model there was increased statistical significance for associations between
preservation attitude and urbanicity, political orientation, and ethnicity. When urban participants
and rural participants with the same preference score, age, ethnicity, education, etc., were
compared, urban participants, on average, had higher attitudinal scores than rural participants
(the estimate mean difference was 0.42 with a p-value of .0336). Likewise, when the covariates
were held constant, political conservatives had lower attitudinal scores that liberals (the estimate
87

mean difference was 0.37 with a p-value of .0389), and white participants had higher attitudinal
scores, on average, than participants of other ethnicities (the estimate mean difference was 0.41
with a p-value of .0233). Again, however, an association between ethnicity and attitude should
be interpreted cautiously given the low proportion of non-white participants in the study.

Table 4.11: Unadjusted association between demographics and preservation attitudes. Models 1-9: simple linear
regressions.
Model
Reference
Group
Variable Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
1
Young
(18-34)
(Intercept) 5.372 0.118 5.141 5.602 45.637 < 0.0001
Age: Middle (35-54) -0.078 0.161 -0.394 0.238 -0.484 0.6289
Age: Old (55+) -0.137 0.165 -0.461 0.187 -0.830 0.4077
2 18-24
(Intercept) 5.484 0.184 5.123 5.846 29.747 < 0.0001
Ages: 25-34 -0.191 0.240 -0.662 0.279 -0.797 0.4265
Ages: 35-44 -0.128 0.243 -0.604 0.348 -0.528 0.5981
Ages: 45-54 -0.250 0.240 -0.720 0.221 -1.041 0.2991
Ages: 55+ -0.250 0.218 -0.677 0.177 -1.147 0.2529
3
(Intercept) 5.516 0.213 5.099 5.934 25.901 < 0.0001
Age: Continuous -0.049 0.046 -0.140 0.041 -1.072 0.2849
4 Male
(Intercept) 5.174 0.092 4.993 5.355 56.004 < 0.0001
Gender: Female 0.250 0.131 -0.006 0.506 1.911 0.0575
5
Not college
grad
(Intercept) 5.142 0.113 4.920 5.364 45.471 < 0.0001
Education: College grad+ 0.236 0.139 -0.036 0.508 1.704 0.0901
6 Not-white
(Intercept) 5.064 0.157 4.755 5.373 32.155 < 0.0001
Race: White 0.300 0.173 -0.039 0.639 1.734 0.0845
7 Urban
(Intercept) 5.490 0.120 5.255 5.725 45.787 < 0.0001
Community: Suburban -0.280 0.153 -0.581 0.020 -1.828 0.0692
Community: Rural -0.295 0.186 -0.661 0.070 -1.585 0.1147
Community: Other 0.361 0.541 -0.698 1.421 0.668 0.5047
8 Liberal
(Intercept) 5.435 0.096 5.246 5.623 56.536 < 0.0001
Political: Moderate -0.175 0.155 -0.479 0.130 -1.125 0.2619
Political: Conservative -0.366 0.170 -0.700 -0.032 -2.146 0.0331*
Political: Other -0.194 0.385 -0.947 0.560 -0.504 0.6148
9 South
(Intercept) 5.234 0.105 5.028 5.440 49.840 < 0.0001
Region: Northeast 0.089 0.196 -0.295 0.473 0.453 0.6510
Region: Midwest 0.158 0.190 -0.213 0.530 0.835 0.4049
Region: West 0.087 0.165 -0.237 0.411 0.524 0.6006
*Associations significant at the 0.05 level

88

Table 4.12: Adjusted association between demographics and preservation attitudes. Models 10-12: multiple linear
regressions.
Model
Reference
Group
Variable Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
10
(Intercept) 5.128 0.234 4.671 5.586 21.959 < 0.0001
Young
(18-34)
Age: Middle (35-54) 0.019 0.168 -0.311 0.349 0.112 0.9110
Age: Old (55+) -0.116 0.169 -0.448 0.216 -0.684 0.4946
Male Gender: Female 0.194 0.138 -0.077 0.465 1.401 0.1629
Not college grad Education: College grad+ 0.130 0.144 -0.152 0.412 0.901 0.3689
Not-white Race: White 0.409 0.184 0.049 0.769 2.225 0.0274*
Urban
Community: Suburban -0.263 0.158 -0.574 0.047 -1.665 0.0978
Community: Rural -0.388 0.201 -0.782 0.006 -1.931 0.0551
Community: Other 0.158 0.564 -0.947 1.262 0.280 0.7800
Liberal
Political: Moderate -0.144 0.157 -0.451 0.163 -0.918 0.3597
Political: Conservative -0.312 0.178 -0.660 0.036 -1.757 0.0806
Political: Other -0.139 0.400 -0.922 0.644 -0.347 0.7287
South
Region: Northeast -0.045 0.202 -0.440 0.350 -0.223 0.8236
Region: Midwest 0.093 0.194 -0.288 0.473 0.479 0.6327
Region: West 0.004 0.170 -0.329 0.337 0.025 0.9803
11
(Intercept) 5.336 0.238 4.870 5.802 22.424 < 0.0001
Age: Continuous -0.038 0.047 -0.130 0.054 -0.811 0.4185
Male Gender: Female 0.213 0.132 -0.046 0.471 1.612 0.1085
Not college grad Education: College grad+ 0.216 0.139 -0.057 0.488 1.548 0.1234
Liberal
Political: Moderate -0.170 0.154 -0.472 0.133 -1.099 0.2732
Political: Conservative -0.315 0.173 -0.654 0.024 -1.820 0.0704
Political: Other -0.093 0.385 -0.848 0.662 -0.241 0.8095
12
(Intercept) 4.334 0.371 3.607 5.061 11.680 < 0.0001
Pref. for new Preference for Old Buildings 0.225 0.083 0.063 0.387 2.724 0.0071*
Young
(18-34)
Age: Middle (35-54) -0.028 0.166 -0.354 0.297 -0.170 0.8650
Age: Old (55+) -0.184 0.168 -0.514 0.145 -1.096 0.2747
Male Gender: Female 0.175 0.136 -0.092 0.441 1.283 0.2013
Not college grad Education: College grad+ 0.109 0.142 -0.169 0.386 0.768 0.4434
Not-white Race: White 0.413 0.181 0.059 0.767 2.288 0.0233*
Urban
Community: Suburban -0.238 0.156 -0.543 0.068 -1.526 0.1287
Community: Rural -0.416 0.198 -0.804 -0.029 -2.107 0.0366*
Community: Other 0.211 0.554 -0.875 1.296 0.380 0.7043
Liberal
Political: Moderate -0.173 0.154 -0.476 0.129 -1.123 0.2630
Political: Conservative -0.365 0.176 -0.710 -0.021 -2.081 0.0389*
Political: Other -0.220 0.394 -0.992 0.551 -0.560 0.5763
South
Region: Northeast -0.080 0.199 -0.469 0.309 -0.403 0.6876
Region: Midwest 0.061 0.191 -0.313 0.436 0.320 0.7497
Region: West -0.016 0.167 -0.343 0.312 -0.094 0.9252
*Associations significant at the 0.05 level

89

Associations between Demographic Variables and Engagement in Preservation
When associations between the demographic variables and engagement in preservation
were investigated in both linear and logistic regressions, the only finding of significance was a
relationship between urbanicity and participation. (Table 4.13) Specifically, under the linear
model, participants who lived in urban areas, in comparison to those of the same covariates who
lived in suburban areas, participated in 0.41 more preservation activities on average (p-value =
0.0314). For the other demographic variables, under both models, no statistically significant
associations were found with engagement in historic preservation. (Tables 4.13 and 4.14)

90

Table 4.13: Association between demographics and preservation activities (method 1: linear regression). Models
1-9: simple linear regressions. Models 10-11: multiple linear regressions.
Model
Reference
Group
Variable Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
1
Young
(18-34)
(Intercept) 1.459 0.139 1.187 1.731 10.513 < 0.0001
Age: Middle (35-54) -0.267 0.188 -0.636 0.101 -1.421 0.1568
Age: Old (55+) -0.111 0.193 -0.488 0.267 -0.574 0.5665
2 18-24
(Intercept) 1.280 0.217 0.854 1.706 5.895 < 0.0001
Ages: 25-34 0.303 0.283 -0.251 0.857 1.073 0.2845
Ages: 35-44 -0.023 0.284 -0.580 0.534 -0.080 0.9360
Ages: 45-54 -0.148 0.280 -0.696 0.399 -0.531 0.5961
Ages: 55+ 0.068 0.255 -0.431 0.568 0.269 0.7885
3
(Intercept) 1.449 0.251 0.957 1.941 5.768 < 0.0001
Age: Continuous -0.028 0.054 -0.134 0.078 -0.518 0.6047
4 Male
(Intercept) 1.343 0.109 1.129 1.557 12.301 < 0.0001
Gender: Female -0.037 0.154 -0.338 0.265 -0.238 0.8125
5 Not college grad
(Intercept) 1.194 0.132 0.935 1.453 9.026 < 0.0001
Education: College grad+ 0.197 0.162 -0.121 0.515 1.214 0.2262
6 Not-white
(Intercept) 1.147 0.186 0.782 1.512 6.167 < 0.0001
Race: White 0.225 0.204 -0.176 0.625 1.100 0.2725
7 Urban
(Intercept) 1.533 0.140 1.259 1.807 10.963 < 0.0001
Community: Suburban -0.361 0.179 -0.713 -0.010 -2.014 0.0454*
Community: Rural -0.161 0.216 -0.586 0.263 -0.745 0.4573
Community: Other -0.200 0.641 -1.456 1.056 -0.312 0.7554
8 Liberal
(Intercept) 1.348 0.114 1.125 1.571 11.852 < 0.0001
Political: Moderate -0.032 0.184 -0.392 0.328 -0.174 0.8619
Political: Conservative -0.014 0.198 -0.403 0.374 -0.073 0.9418
Political: Other -0.348 0.460 -1.249 0.553 -0.757 0.4501
9 South
(Intercept) 1.300 0.122 1.061 1.539 10.655 < 0.0001
Region: Northeast 0.112 0.223 -0.326 0.550 0.500 0.6175
Region: Midwest -0.035 0.223 -0.473 0.403 -0.158 0.8746
Region: West 0.046 0.194 -0.335 0.427 0.237 0.8126
10
(Intercept) 1.420 0.282 0.867 1.974 5.032 < 0.0001
Young
(18-34)
Age: Middle (35-54) -0.323 0.202 -0.718 0.073 -1.598 0.1118
Age: Old (55+) -0.178 0.204 -0.579 0.222 -0.872 0.3842
Male Gender: Female -0.155 0.166 -0.479 0.170 -0.933 0.3522
Not college grad Education: College grad+ 0.187 0.172 -0.151 0.525 1.086 0.2791
Not-white Race: White 0.292 0.220 -0.139 0.723 1.329 0.1856
Urban
Community: Suburban -0.410 0.189 -0.781 -0.039 -2.169 0.0314*
Community: Rural -0.117 0.239 -0.586 0.352 -0.488 0.6263
Community: Other -0.156 0.684 -1.497 1.184 -0.228 0.8196
Liberal
Political: Moderate -0.017 0.188 -0.385 0.352 -0.088 0.9297
Political: Conservative 0.088 0.210 -0.324 0.499 0.418 0.6768
Political: Other -0.319 0.485 -1.269 0.632 -0.657 0.5120
South
Region: Northeast 0.096 0.236 -0.366 0.559 0.408 0.6838
Region: Midwest -0.131 0.235 -0.591 0.329 -0.557 0.5779
Region: West 0.076 0.205 -0.327 0.478 0.368 0.7131
11
(Intercept) 1.398 0.287 0.836 1.960 4.875 < 0.0001
Age: Continuous -0.036 0.056 -0.146 0.074 -0.645 0.5194
Male Gender: Female -0.068 0.157 -0.375 0.239 -0.436 0.6631
Not college grad Education: College grad+ 0.206 0.166 -0.119 0.530 1.242 0.2159
Liberal
Political: Moderate -0.032 0.185 -0.394 0.330 -0.175 0.8615
Political: Conservative 0.020 0.204 -0.379 0.419 0.099 0.9211
Political: Other -0.351 0.465 -1.262 0.561 -0.753 0.4521
*Associations significant at the 0.05 level

91

Table 4.14: Association between demographics preservation activities (method 2: logistic regression). Models 1-
9: simple linear regressions. Models 10-11: multiple linear regressions.
Model
Reference
Group
Variable Estimate SE
Odds
Ratio
Lower
95%
OR
Upper
95%
OR
Z-
statistic
P-
value
1
Young
(18-34)
(Intercept) -0.433 0.262 0.649 0.388 1.084 -1.652 0.0986
Age: Middle (35-54) -0.281 0.362 0.755 0.372 1.534 -0.777 0.4373
Age: Old (55+) -0.193 0.368 0.825 0.401 1.696 -0.524 0.6003
2 18-24
(Intercept) -0.754 0.429 0.471 0.203 1.090 -1.758 0.0787
Ages: 25-34 0.531 0.544 1.700 0.585 4.941 0.975 0.3297
Ages: 35-44 0.228 0.553 1.256 0.424 3.714 0.411 0.6807
Ages: 45-54 -0.144 0.558 0.866 0.290 2.586 -0.258 0.7963
Ages: 55+ 0.128 0.501 1.137 0.426 3.032 0.256 0.7981
3
(Intercept) -0.368 0.479 0.692 0.271 1.772 -0.767 0.4434
Age: Continuous -0.052 0.104 0.949 0.774 1.164 -0.502 0.6159
4 Male
(Intercept) -0.693 0.213 0.500 0.329 0.759 -3.251 0.0011
Gender: Female 0.188 0.296 1.206 0.675 2.155 0.634 0.5263
5 Not college grad
(Intercept) -0.649 0.257 0.523 0.316 0.866 -2.521 0.0117
Education: College grad+ 0.077 0.314 1.080 0.583 2.000 0.246 0.8059
6 Not-white
(Intercept) -0.875 0.376 0.417 0.199 0.871 -2.326 0.0200
Race: White 0.352 0.410 1.421 0.637 3.172 0.858 0.3906
7 Urban
(Intercept) -0.405 0.264 0.667 0.398 1.117 -1.539 0.1239
Community: Suburban -0.488 0.349 0.614 0.310 1.216 -1.400 0.1614
Community: Rural 0.172 0.405 1.188 0.537 2.625 0.425 0.6711
Community: Other -0.288 1.253 0.750 0.064 8.738 -0.230 0.8184
8 Liberal
(Intercept) -0.488 0.215 0.614 0.403 0.935 -2.271 0.0231
Political: Moderate -0.127 0.351 0.880 0.443 1.751 -0.363 0.7164
Political: Conservative -0.107 0.378 0.899 0.428 1.886 -0.283 0.7773
Political: Other -16.078 979.610 0.000 0.000 Inf -0.016 0.9869
9 South
(Intercept) -0.619 0.234 0.538 0.340 0.852 -2.641 0.0083
Region: Northeast 0.139 0.424 1.150 0.501 2.637 0.329 0.7420
Region: Midwest -0.119 0.435 0.888 0.379 2.084 -0.272 0.7853
Region: West 0.067 0.371 1.069 0.516 2.214 0.180 0.8569
10
(Intercept) -0.413 0.549 0.662 0.226 1.940 -0.753 0.4517
Young (18-34)
Age: Middle (35-54) -0.446 0.396 0.640 0.294 1.393 -1.124 0.2608
Age: Old (55+) -0.391 0.397 0.677 0.311 1.473 -0.985 0.3248
Male Gender: Female -0.056 0.320 0.946 0.505 1.771 -0.174 0.8622
Not college grad Education: College grad+ -0.017 0.336 0.983 0.509 1.899 -0.051 0.9597
Not-white Race: White 0.444 0.443 1.559 0.654 3.717 1.001 0.3169
Urban
Community: Suburban -0.530 0.370 0.589 0.285 1.216 -1.431 0.1525
Community: Rural 0.301 0.451 1.352 0.558 3.274 0.668 0.5044
Community: Other 0.616 1.486 1.852 0.101 34.050 0.415 0.6783
Liberal
Political: Moderate -0.126 0.362 0.881 0.433 1.792 -0.349 0.7272
Political: Conservative -0.043 0.406 0.957 0.432 2.122 -0.107 0.9148
Political: Other -16.340 912.114 0.000 0.000 Inf -0.018 0.9857
South
Region: Northeast 0.284 0.457 1.328 0.542 3.252 0.621 0.5349
Region: Midwest -0.269 0.460 0.764 0.311 1.881 -0.585 0.5586
Region: West 0.192 0.401 1.212 0.552 2.658 0.479 0.6319
11
(Intercept) -0.304 0.543 0.738 0.255 2.139 -0.559 0.5759
Age: Continuous -0.068 0.108 0.934 0.756 1.155 -0.628 0.5301
Male Gender: Female 0.124 0.300 1.131 0.628 2.039 0.411 0.6809
Not college grad Education: College grad+ 0.052 0.322 1.053 0.560 1.978 0.160 0.8727
Liberal
Political: Moderate -0.113 0.352 0.893 0.448 1.781 -0.321 0.7480
Political: Conservative -0.048 0.388 0.953 0.445 2.041 -0.124 0.9016
Political: Other -16.052 975.639 0.000 0.000 Inf -0.016 0.9869
“Inf” stands for infinite (where the proportion of people comprising the category was so small, there is infinite uncertainty).
*Associations significant at the 0.05 level

92

4.2.1.5 Pairwise Associations among Preference for Old Buildings, Preservation Attitudes, and
Engagement in Preservation
Association between Preference for Old Buildings and Preservation Attitudes
There was a positive and statistically significant association between preference for old
buildings and preservation attitudes. On average, a one unit increase in the mean score of
preference for old buildings was associated with a .213 unit increase in the mean score of
preservation attitudes. (Table 4.15) This is illustrated below in Figure 4.34, which provides the
plot of the mean score for preservation-related attitudes versus the mean score for old buildings.

Table 4.15: Association between preference for old buildings and preservation attitudes
Variable Estimate SE Lower 95% Upper 95% Z-statistic P-value
(Intercept) 4.463 0.322 3.832 5.094 13.857 < 0.0001
Preference for Old Buildings 0.213 0.080 0.056 0.371 2.650 0.0087*
*Associations significant at the 0.05 level

Figure 4.34: Preference for old buildings and preservation-
related attitudes.

Mean preference rating for old buildings
Mean score for preservation related attitudes
93

Association between Preference for Old Buildings and Engagement in Preservation
When the relationship between preference for old buildings and engagement in
preservation-related activities was investigated, a positive direction of association was found in
both linear and logistic models. In the linear model, a one unit increase in the preference for old
buildings was associated with a participant engaging in 0.16 more preservation activities with
95% CI (-0.025, 0.350). However, with a p-value of 0.0919, statistical significance was not
achieved. (Table 4.16) When the logistic model was considered, where the outcome was whether
a subject participated in two or more activities, the estimated odds ratio was 1.622 with 95% CI
(1.110, 2.372) and a p-value of 0.0125. The point estimate implies that when we compare two
subjects with one unit difference in mean old building ratings, the odds that the subject with the
higher rating would participate in two or more preservation activities was 1.62 times higher than
the subject with the lower rating.

Table 4.16: Association between preference for old buildings and historic preservation activities
(method 1: linear regression)
Variable Estimate SE Lower 95% Upper 95% Z-statistic P-value
(Intercept) 0.690 0.383 -0.061 1.440 1.802 0.0731
Preference for Old Buildings 0.162 0.096 -0.025 0.350 1.694 0.0919

Association between Preservation Attitudes and Engagement in Preservation
Positive and statistically significant associations were identified between preservation
attitudes and engagement in preservation in both linear and logistic regression models. In the
linear model, an increase of one unit in preservation attitude was associated with participation in
an estimated 0.48 more preservation activities with 95% CI (0.325, 0.631) and with a p-value of
less than 0.0001. (Table 4.17) In the logistic model, a one unit average increase in preservation
attitude was associated with the odds ratio of 2.662 with 95% CI (1.724, 4.111) and a p-value of
0.0125.
When compared to the results of the previous section, the estimates were substantially
greater in both models. In other words, the average score for preservation attitude was a stronger
predictor for participation in historic preservation activities than the average score for preference
of old buildings.

94

Table 4.17: Association between preservation attitudes and historic preservation activities
(method 1: linear regression)
Variable Estimate SE Lower 95% Upper 95% Z-statistic P-value
(Intercept) -1.203 0.420 -2.026 -0.379 -2.862 0.0047
Preservation attitude 0.478 0.078 0.325 0.631 6.117 < 0.0001*
*Associations significant at the 0.05 level

4.2.2 Secondary Analysis: Predictors of the Average Preference Rating of a Building
Based on previous research, the primary variables of complexity, condition, and
landscape were expected to be predictors of preference for buildings in the current study.
However, only complexity was found to be significantly associated with the rating of a building.
Further, the secondary variables showed minimal associative relationships with preference, with
only the number of stories a building had being predictive of preference.
The results of analyses for unadjusted associations between average building rating and
the primary and secondary variables are presented in Table 4.18. The main findings are: when
other covariates were not controlled, survey participants preferred buildings with greater
complexity and more stories. There was also a large effect for sky; specifically, a building with a
dreary, gray sky had, on average, a preference score 1.17 units higher than one with a nice, blue
sky. This finding – which is obviously counterintuitive – is likely the result of an unbalanced
distribution of gray skies among the old and new building sets. There was only one photograph
with a bad sky in the new building set, whereas there were four in the old building set.
Considering that subjects tended to give higher ratings to old buildings than new buildings, the
effect of bad sky could be due to this unbalanced distribution. Likewise, buildings located in the
Western region of the United States were rated lower than other regions, but again, this result
was likely due to the unbalanced distribution of old and new buildings in each region (see Table
3.3).

95

Table 4.18: Unadjusted associations between primary / secondary variables and preference (dependent
variable: average preference rating). All models are simple linear regressions.
Model
Reference
Group
Independent
Variable
Estimate SE
Lower
95%
Upper
95%
Z-
statistic
P-value
Old vs.
new
New
buildings
(Intercept) 3.145 0.160 2.831 3.458 19.658 < 0.0001
Old 0.772 0.226 0.329 1.215 3.412 0.0013*
Condition
(Intercept) 2.934 0.902 1.166 4.701 3.253 0.0021
Condition 0.108 0.161 -0.208 0.424 0.669 0.5069
Complexity
(Intercept) 1.577 0.357 0.877 2.277 4.418 0.0001
Complexity 0.527 0.093 0.345 0.708 5.689 < 0.0001*
Landscape
(Intercept) 3.883 0.347 3.203 4.562 11.192 < 0.0001
Landscape -0.119 0.109 -0.333 0.096 -1.087 0.2826
# of stories
(Intercept) 2.752 0.318 2.129 3.375 8.657 < 0.0001
Stories 0.295 0.112 0.076 0.514 2.637 0.0112*
Building
size
Small
(Intercept) 3.143 0.385 2.389 3.897 8.170 < 0.0001
Medium 0.157 0.277 -0.386 0.700 0.566 0.5743
Large 0.500 0.456 -0.393 1.393 1.098 0.2780
Building
orientation
Façade view
(Intercept) 3.365 0.150 3.071 3.658 22.480 < 0.0001
3-quarter view 0.488 0.257 -0.015 0.991 1.900 0.0634
Cars
present
Cars absent
(Intercept) 3.551 0.172 3.215 3.887 20.700 < 0.0001
Cars present -0.044 0.253 -0.540 0.452 -0.173 0.8632
People
present
People
absent
(Intercept) 3.502 0.134 3.239 3.764 26.157 < 0.0001
People present 0.243 0.386 -0.514 1.000 0.629 0.5325
Cladding
type
Red brick
(Intercept) 3.581 0.185 3.219 3.942 19.394 < 0.0001
Yellow brick 1.408 0.477 0.474 2.343 2.954 0.0052*
Brown brick -1.426 0.783 -2.961 0.110 -1.820 0.0761
Painted brick 0.307 0.387 -0.452 1.066 0.793 0.4326
Stone 0.809 0.477 -0.125 1.743 1.697 0.0972
Stucco -0.435 0.387 -1.194 0.324 -1.124 0.2676
Glass -0.400 0.342 -1.070 0.270 -1.169 0.2493
Other -0.485 0.326 -1.125 0.154 -1.487 0.1448
Combination -0.399 0.783 -1.934 1.137 -0.509 0.6135
Quality of
sky
Good sky
(Intercept) 3.414 0.122 3.175 3.653 28.036 < 0.0001
Bad sky 1.167 0.385 0.412 1.922 3.030 0.0039*
Building
region
Northeast
(Intercept) 4.595 0.355 3.899 5.290 12.955 < 0.0001
Midwest -0.770 0.414 -1.582 0.042 -1.859 0.0695
South -0.647 0.445 -1.519 0.224 -1.456 0.1523
West -1.516 0.380 -2.261 -0.772 -3.990 0.0002*
Pacific 0.752 0.793 -0.802 2.306 0.948 0.3480
Award /
designation
No award /
designation
(Intercept) 3.440 0.144 3.157 3.723 23.837 < 0.0001
Award / designation 0.348 0.283 -0.206 0.903 1.230 0.2245
*Associations significant at the 0.05 level

96

The results of a multiple linear regression are presented in Table 4.19. Because of the
small number of buildings used as stimuli in the study, a subset of the primary and secondary
predictors were included in the regression; building size, cladding, award / designation, and
region were excluded based on the outcome of the simple linear regressions in Table 4.18. In this
model, with the covariates controlled, there was a .530 difference in preference for old buildings
in comparison to new buildings. And, here again, complexity and number of stories were
positively associated with preference. A one point increase in complexity was associated with a
.453 increase in preference for buildings of the same age, condition, landscape, number of
stories, etc. Likewise, with the other covariates held constant, a one unit increase in building
height resulted in a .218 increase in the average preference rating.

Table 4.19: Multiple linear regression for adjusted association between primary / secondary variables and
preference (dependent variable: average preference rating).
Independent Variable Reference Group Estimate SE Lower 95% Upper 95% Z-statistic P-value
(Intercept) 1.413 0.701 0.040 2.786 2.017 0.0505
Old buildings New buildings 0.530 0.181 0.176 0.885 2.932 0.0056*
Complexity 0.453 0.092 0.273 0.633 4.924 < 0.0001*
Condition -0.064 0.124 -0.307 0.180 -0.512 0.6112
Landscape -0.044 0.087 -0.215 0.127 -0.506 0.6159
# Stories 0.218 0.103 0.016 0.421 2.112 0.0410*
3-quarter view Façade view 0.163 0.211 -0.251 0.577 0.772 0.4448
Cars present Cars absent 0.030 0.177 -0.318 0.378 0.168 0.8672
People present People absent -0.383 0.324 -1.018 0.252 -1.182 0.2440
Bad sky Good sky 0.578 0.324 -0.056 1.212 1.787 0.0816
*Associations significant at the 0.05 level

The final model investigated if the association between preference and the age of a
building (new versus old) was modified by building complexity. Conditioning on all the
covariates in Table 4.19, the effect modification was tested and the p-value was 0.0062. The
estimated difference in the average rating for an old building and a new building at six different
levels of complexity is presented below in Table 4.20. The results indicate that although, in
general, participants prefer old buildings over new, this is not necessarily true for buildings of
low complexity, where the estimated difference in the average preference ratings was about 0.23
points with 95% CI (-0.854, 0.375). However, as complexity increased to moderate and high
levels, participants preferred old buildings by increasing and significant amounts. At the highest
complexity level, preference for old buildings was, on average, 1.58 points higher than
preference for new buildings with 95% CI (0.796, 2.354).
97

Table 4.20: Association between preference and building age,
modified by complexity
Complexity Estimate Lower 95% Upper 95%
2 (low complexity) -0.239 -0.854 0.375
3 0.214 -0.176 0.604
4 0.668 0.329 1.007
5 1.121 0.605 1.637
6 (high complexity) 1.575 0.796 2.354
The p-value for testing the effect modification was 0.0062.

Given this finding, it is important to note that though there was more extremity in high
complexity among the old buildings, the histograms in Figure 4.35 illustrate that the distribution
of complexity was fairly balanced between the old and new building sets (by the design of the
study). Specifically, whereas the highest complexity rating for a new building was 5.3 on the
seven point scale, three old buildings were rated as more complex, with the highest receiving a
rating of 6.3. On average, old buildings had complexity ratings 0.30 units higher than new
buildings, but this estimated difference was not statistically significant with a p-value of 0.317.

Figure 4.35: Distribution of complexity among old and new building sets.

0
1
2
3
4
5
6
1 2 3 4 5 6 7
Frequency
Complexity
Old Buildings
0
1
2
3
4
5
6
1 2 3 4 5 6 7
Frequency
Complexity
New Buildings
98

5. DISCUSSION

5.1 Major Findings
The main objectives of this study were to determine if people prefer old buildings or new,
if they are mostly supportive of historic preservation, to what extent they engage in preservation-
related activities, and if there are demographic trends in these study areas. The results indicate
that, on average, people do prefer old buildings over new and they have generally favorable
attitudes about historic preservation. These preferences and beliefs, however, do not translate
into significant engagement in preservation-related activities. And although in general,
demographics do not seem to matter much in terms of preference, attitudes, and engagement,
there were some notable associations in regard to age, gender, political affiliation, and
urbanicity.

Major Findings: Preference
Regarding preference, the findings are consistent with the research of Frewald, Herzog
and Gale, and Herzog and Sheir, who have found that with building condition controlled, people
generally have a higher preference for old buildings than new buildings. These findings were
consistent across different types of analyses, including a comparison of the overall means of the
new and old building sets, a comparison of each participants’ average rating for old and new
buildings, and when the buildings were ranked according to their means; collectively, this
provided strong evidence for the preference of old buildings in comparison to new buildings.
There was, however, more variability in the ratings for old buildings, indicating that people tend
to have stronger reactions (either positive or negative) to old buildings than they do for new
buildings. As expected, based on the existing literature, building complexity served as a strong
predictor of preference, with preference for both old and new buildings increasing as complexity
increased. The other variables that were predicted to be associated with preference – building
condition and landscape – were not found to have an association. Regarding building condition,
this could be because only those buildings with moderate to high condition ratings were included
in the survey. And it is possible that since all the buildings had little to no setback, leaving
minimal space for landscaping, it may make sense then that there was no observed correlation
99

between landscape and preference. Also observed was that the number of stories a building had
was a moderate predictor of preference; specifically, the more stories, the more preferred the
building. Lastly, a final finding was that post-World War II buildings were preferred less, on
average, than both pre-World War II buildings and new buildings. This may indicate an age
related aspect to preference of old buildings, but since the post-war buildings tended to have low
complexity ratings, it could also be a matter of a lack of visual complexity.
As expected, some demographic variables were associated with preference. Most notably,
there was a strong correlation between participant age and preference. Older participants, in
comparison to younger participants, were more likely to prefer old buildings, and the opposite
was true, as well; younger participants were more likely to prefer new buildings in comparison to
older participants. This finding was consistent with the research of Richard W. Berman.
1

Gender and political affiliation also appeared to be moderately associated with
preference, with women, on average, preferring old buildings more than men and politically
conservative participants preferring old buildings (relative to new buildings) more than those
who identified as liberals. Let us recall the results of Arthur Stamps’ meta-analysis of
demographic effects in environmental preference in which, among other things, he concluded
that men’s and women’s preferences for environmental stimuli were highly correlated, as were
the preferences of people of various political parties.
2
Although men and women and
conservatives and liberals may share similar preferences in other types of environments as
Stamps concluded, when preferences for old and new buildings, specifically, are considered, it is
possible that gender and political ideology may play a role.
There were no other notable associations between preference and the remaining
demographic variables of education, ethnicity, urbanicity, and geographic location. Interestingly,
it should be noted, whereas other researchers have identified associations between education and
preference for old and new buildings, this finding was not supported by the results of this study.

Major Findings: Attitudes
As hypothesized, participants generally had favorable attitudes about historic
preservation, with most people believing that historic resources are community assets and

1. Berman, “Assessing Urban Design: Historical Ambience on the Waterfront.”
2. Stamps III, “Demographic Effects in Environmental Aesthetics,” 155-175.
100

preservation is an important community service, even in comparison to other services considered
important, such as economic development and public landscaping. These findings are consistent
with the results of Daniel Levi’s study, as well as research conducted by several State Historic
Preservation Offices which have demonstrated a generally high level of support among the
public for preservation. While this is the first nationwide study of preservation attitudes, it is a
promising indication that most people in the United States care about preservation and believe it
is important for communities.
Participants of the current study further appeared to have a broad appreciation of what
preservation is by indicating overall agreement with the ideas that historic landscapes are just as
important to preserve as historic buildings and that preservation is environmentally friendly.
Perhaps not surprisingly, where participants appeared to be in less agreement is in regard to
preservation-related regulation and how preservation should be financed. Still, however, most
people thought that there should be regulation and that taxes should contribute to preservation
efforts.
Interestingly, although participant age and gender were found to be predictors of
preference, they were not predictors of preservation attitudes. And, although politically
conservative people had, on average, higher preference for old buildings than liberals, they were
less likely to be preservation-minded in their attitudes. There was, additionally, some evidence
that residents of urban areas had more favorable attitudes about preservation than people living
in rural areas. This may be related to the lack of development pressure in rural areas where
preservation may happen more passively than actively.

Major Findings: Engagement
Although a large majority of participants engaged in one or more preservation activities
in the last year, for most, this involved only visiting a place because it was historic. Although it
is wonderful that so many people are interested in visiting historic resources, the needs of
preservation extend beyond this type of support. A smaller – but still encouraging – number of
people, 41%, did something preservation related in the last year other than or in addition to
visiting an historic place. Much of this engagement was facilitated by the ease of the internet and
social media, where it is simple and quick to engage in preservation by, for example, liking a
post related to the preservation of an historic place on Facebook. Beyond visiting historic
101

resources and online engagement, only 4% of participants reported being active players in
preservation in more significant capacities. So interestingly, although most participants preferred
older buildings and had supportive attitudes about preservation, these preferences and attitudes
did not translate into considerable and varied engagement in preservation. Why then don’t people
get involved? The results of the preservation obstacles items indicate that, as can be expected, for
many it is simply a matter of a lack of time, or, as was evidenced in the National Trust study,
perhaps a matter of competing priorities.
3
Notably, over a third of respondents would like to get
involved but do not know how.
Who are the people most likely to participate in preservation-related activities?
Respondents who resided in urban areas engaged in more preservation activities, on average,
than those who lived in suburban and rural areas. It is possible that this association is because
there are simply more significant concentrations of historic resources in urban communities and /
or because preservation organizations tend to be established in urban places, making it easier to
get involved. Regarding the other demographic variables, whereas the National Trust identified a
number of demographic trends in terms of the people who they identified as local
preservationists and preservation leaders, urbanicity appears to be the only variable associated
with engagement from this study; the other variables of age, gender, education, ethnicity, and
political affiliation did not demonstrate associations with preservation participation.

Major Findings: Associations between Preference, Attitudes, and Engagement
Positive associations were identified between the three major study areas. Higher
preference ratings were associated with higher attitudinal ratings and more engagement in
preservation activities. Higher attitudinal ratings were further associated with increased
engagement in preservation. In regards to the positive association between preference and
attitudes, this contradicts the results of Daniel Levi’s study, in which no association was found
between preference and attitudes.

5.2 Implications and Recommendations for the Field of Historic Preservation
Coupled with the existing literature, the results of the study present strong evidence for
the value of historic preservation in the United States and the findings from this research can be

3. National Trust for Historic Preservation, Field Guide to Local Preservationists.
102

used by preservationists as a tool in advocacy, and also as a foundation for policy and planning.
So again, I ask: why preserve? One might answer: because 75% of people prefer old buildings
over new, 88% believe that historic places should be treated as community assets, and 89%
believe that with proper planning, both development and preservation goals can be achieved.
4

There is undeniable power and clarity in that response and I hope it serves as an effective
demonstration of how the preference and attitudinal data can be used to support the field of
historic preservation.
One might question why it matters if people like the aesthetics of their communities.
Researchers have identified the perceived beauty of a place as one of the most important factors
in a person’s satisfaction with their community, and further, community satisfaction has been
found to be significantly associated with a person’s overall quality of life.
5
Its effect on quality of
life is second only to marital satisfaction.
6
It might be easy to brush off preference as something
that is a low priority to people in the grand scheme of things, but it appears to actually matter to
people in a deep way, and from a planning standpoint, this humanistic element should be an
important consideration.
That being said, it is important that the results of the preference assessment are not taken
too literally. We must recall that architectural merit is just one of the criteria against which
resources are typically assessed for potential historic designation. And further, preference for a
building does not necessarily equate with the building meeting the architectural criterion of a
designation program. A building could have a low visual preference rating, but be an excellent
and rare example of a type of architecture and therefore be a significant historic resource despite
its low preference score. Or, for example, a building could receive a low visual preference rating,
but because it has a social context that is not obvious the observer’s eye, it may actually have
significant community value as an historic resource. Likewise, a building could have a high
preference rating, but not meet the criterion for designation. The point is that making decisions

4. Planning statistic is from State of Hawai’i Historic Preservation Division. Hawai’i State Historic Preservation
Plan.
5. Richard Florida, Charlotta Mellander, and Kevin Stolarick, Beautiful Places: The Role of Perceived Aesthetic
Beauty in Community Satisfaction (Toronto, Canada: University of Toronto: Martin Prosperity Institute, 2009);
Robin N. Widgery, “Satisfaction with the Quality of Urban Life: A Predictive Model,” American Journal of
Community Psychology 10.1 (1982): 37-48; Michael J. White, “Determinants of Community Satisfaction in
Middletown,” American Journal of Community Psychology 13.5 (1985): 583-597.
6. Marc Fried, “The Structure and Significance of Community Satisfaction,” Population and Environment 7.2
(1984): 61-86.
103

about what buildings to keep or not keep based solely on an age-based preference study would be
an inappropriate use of the results. As a further example, consider the two findings: we know that
it is important that people value the aesthetics of their communities and we also know that, on
average, people appear to dislike mid-century architecture. These statements may be interpreted
by some as a license to destroy post-war buildings. It is important, however, when interpreting
the data to recall that without fail, every architectural movement has lost favor in the proceeding
decades and that it takes time for the public to again appreciate former trends. It is hard to
believe that at one point people disliked Victorian or Art Deco architecture, for example, but it
happened and we now regret the significant resources that were lost during those periods. We
cannot assume that mid-century resources will be the one type of architecture that will be
immune to this trend and not regain preference at a later date. In 10 years, it is entirely likely that
the post-war buildings in the study would receive higher preference ratings than they did today.
Among our many challenges as preservationists is the need to understand this bigger picture,
help people be aware that current public preference may not reflect future preference, and
balance the demands of the present with the needs of the future.
The engagement findings present further opportunities for preservationists to make use of
the data. For instance, while it is certainly a good thing that so many people have intentionally
visited an historic resource and it shows an appreciation for history and preservation, this type of
support alone cannot sustain the needs of the preservation movement. It is important for
managers of historic resources open to the public to be aware that this is our best opportunity to
engage the public in preservation. We also need to consider how this casual interest in visiting
historic places can be converted into additional forms of support and engagement. As
preservation agencies, we are often competing for the same dollars and resources, but if we are
more collaborative, that is better for the entire field. Since historic sites have the most access to
the general public, local preservation organizations and agencies should create partnerships with
popular sites so that they too may interact with this audience. Meeting people where they are
rather than expecting them to come to us may result in greater engagement overall, which would
be better for the field and the culture of preservation in the U.S.
The engagement findings also suggest that beyond visiting historic resources, the easiest
way to get people engaged in preservation is by connecting with them online. This reinforces the
importance of preservation agencies and organizations dedicating time and resources to an online
104

presence. For the 50% of people who said that they would like to get involved in preservation but
do not have the time, this may be how we need to engage them since it is a simple way to support
the cause that does not require much time from a person. And not only is an online presence an
easy way to regularly educate and dispense information to those who are preservation-minded
and connected to the field, but with the simple action of someone “liking” or sharing a post,
preservation can reach new audiences who would otherwise be missed. Just as importantly,
though, is the potential that online engagement can serve as a catalyst for participation in more
involved preservation activities.
Another implication of the data is the knowledge that 37% of survey participants would
like to support preservation in their communities but do not know how to get involved. This is
encouraging, as well as a reminder that we are missing many potential supporters in our current
outreach efforts. In addition to establishing connections with new supporters via our typical
platforms, perhaps a targeted approach would prove to be more worthwhile. As more
quantitative and qualitative research is conducted in the field of preservation by local agencies
and organizations, researchers should take the opportunity to ask participants if they would like
to be involved in preservation in their communities and then follow up with those who respond
affirmatively with suggestions for how they can get connected locally. Rather than casting a
wide net, a direct effort, such as this approach, has the potential to translate into increased
community support.

5.3 Recommendations for Further Research
This study was the first of its kind to be conducted with a nationwide sample of
participants. As such, additional studies with participants from across the U.S. should be
administered to further substantiate the current evidence and to provide more information where
there were discrepancies with the existing literature. Since a limitation of the current study was
the use of an expert panel in order to control for factors that predict preference, a participant
sample is recommended in future research for this aspect of the methodology. In addition to
further nationwide research, some agencies may consider conducting similar research at a
regional or local level as community specific data may be more relevant and useful in local
practice. Further, since the preference assessment of the current study involved only low- to mid-
rise commercial and office buildings, it is also recommended that future researchers of age-based
105

preference consider using other types of buildings in their studies. Additionally, since building
exteriors are just one aspect of preservation, it would be beneficial to study other environmental
stimuli as well, such as old and new building interiors, structures, objects, and landscapes.
Lastly, since the post-WWII old buildings were found to be significantly less preferred than the
pre-war buildings, and because there were only four post-war buildings in the study, it is
recommended that future researchers investigate this finding with a larger sample of post-war
buildings.
In addition to there being a need for further quantitative research, qualitative endeavors
are necessary, as well. With the exception of Jeremy Wells’ study, the related literature has been
exclusively quantitative in nature. Many researchers would agree that quantitative studies tell us
what people are thinking, but qualitative research can help us fully understand why people think
the way they do. There is ample opportunity to enhance the findings of the existing quantitative
literature with qualitative investigations of preference, attitudes, and engagement.
Regarding preservation-related attitudes, specifically, the series of items that were asked
of participants in the current study were rather broad in context in order to provide a baseline
understanding of how people at the national level think about preservation. It is recommended
that more specific quantitative and qualitative investigations are conducted in order to better
understand the depth and complexity of preservation attitudes in the United States and at local
levels. For example, there was only one item in the current study about the relationship between
preservation and sustainability, but entire studies could be conducted on the topic.
Regarding the topic of engagement, the results of the current study tell us how people are
involved with preservation, but what we do not know is what are the most effective ways to
transform a casual interest in preservation into greater levels of engagement. We know
anecdotally that people tend to become motivated to engage in preservation when a resource they
feel strongly about is threatened, but are there actions that can be taken that are effective ways to
motivate people into engaging in historic preservation? For example, what are the most effective
ways to motivate someone who visits an historic site to move on to other forms of engagement?
In a field of limited resources, this is one of the ways in which we can be smart and data-driven
in our practice.
As a final note regarding future research, I want to provide my thoughts regarding a
question I was occasionally asked as I was developing the study and discussing my research
106

plans with preservationists. Sometimes people would question what I would do if the data came
back and it was not “good” for historic preservation. Although one may simply be wondering
how I would talk about “bad” data, the question also communicates a fear that we should not be
conducting research of this sort because we may not like what we hear or because the results
may be used against preservation. A fear of collecting data prevents the field moving forward
with purpose and validity. In order to make genuine progress as a respected discipline, we must
be willing to face the reality in which we are operating, regardless of what the facts are. Hiding
from data is limiting and brushing “bad” data under the rug is unethical. If ever the data come
back and it is not “good” for preservation, we need to be able to accept it and see the silver lining
in it. For example, knowing that people, on average, do not like mid-century architecture and that
a smaller majority of people think that tax dollars should support private preservation efforts is
useful information. The data provide a clear direction for outreach and education, and for a field
that is often stretched and lacking in resources, this is assistance we need. Making informed
decisions and pushing policy with evidence will take the field farther than “bad” data will set it
back.

5.4 Final Conclusions
The results of the study are promising for the field of historic preservation and indicate an
overall level of appreciation among the public for old buildings in comparison to new buildings,
as well as a belief in the value of preservation. The data further indicate that a large majority of
people engage in preservation in some way, although on a somewhat casual level. The findings
additionally point to opportunities for improvement, particularly in regard to increasing
appreciation for mid-century architecture, the need to engage people in preservation in ways that
are not time intensive, and the need to transfer casual engagement in preservation activities to
more active and involved action. The results also highlight future research opportunities,
including additional nationwide and localized assessments of preference, attitudes, and
engagement, as well as more nuanced investigations of these topics. Lastly, the results of the
current study provide an opportunity to support preservation advocacy, as well as to guide
policy, planning, and outreach efforts.

107

BIBLIOGRAPHY

The American Institute of Architects. “America’s Favorite Architecture.” Accessed November
12, 2013. http://www.favoritearchitecture.org/afa150.php.

“Amazon Mechanical Turk.” Accessed May 1, 2014. www.mturk.com.

Avrami, Erica C., Randall Mason, and Marta De la Torre. Values and Heritage Conservation:
Research Report. Los Angeles, CA: Getty Conservation Institute, 2000.

Berinksy, Adam J., Gregory A. Huber, and Gabriel S. Lenz. “Evaluating Online Labor Markets
for Experimental Research: Amazon.com’s Mechanical Turk.” Political Analysis 20 (2012):
351-368.

Berman, Richard W. “Assessing Urban Design: Historical Ambience on the Waterfront.” Ph.D.
diss., University of Pennsylvania, 1999.

Bishop, R. “The Perception and Importance of Time in Architecture.” Ph.D. diss., University of
Surrey, England, 1982.

Budruk, Megha, Kathleen Andereck, and Gautam Prateek. 2013-2014 Arizona State Historic
Preservation Study: Final Technical Report (Draft). Phoenix, AZ: Arizona State Parks
Board, 2014.

Buhrmester, Michael, Tracy Kwang, and Samuel D. Gosling. “Amazon’s Mechanical Turk: A
New Source of Inexpensive, Yet High-Quality, Data?” Perspectives on Psychological
Science 6.1 (2011): 3-5.

Canter, D. and R. Thorne. “Attitudes to Housing: A Cross-cultural Comparison.” Environment
and Behavior 4 (1972): 3-23.

Chandler, Jesse, Pam Mueller, and Gabriele Paolacci. “Nonnaïveté among Amazon Mechanical
Turk Workers: Consequences and Solutions for Behavioral Researchers.” Behavioral
Research Methods 46.1 (2014): 112-130.

Clayton, Kristyn. “Personality and Architectural Preferences: A Search for Patterns.” Master’s
thesis, Washington State University, 2007.

Coeterier, J.F. “Lay People’s Evaluation of Historic Sites.” Landscape and Urban Planning 59
(2002): 111-123.

Florida, Richard, Charlotta Mellander, and Kevin Stolarick. Beautiful Places: The Role of
Perceived Aesthetic Beauty in Community Satisfaction. Toronto, Canada: University of
Toronto: Martin Prosperity Institute, 2009.
108

Frewald, Dori Beth. “Preferences for Older Buildings: A Psychological Approach to
Architectural Design.” Ph.D. diss., University of Michigan-Ann Arbor, 1989.

Fried, Marc. “The Structure and Significance of Community Satisfaction.” Population and
Environment 7.2 (1984): 61-86.

Galindo, M. P. and M. C. Hidalgo. “Aesthetic Preferences and the Attribution of Meaning:
Environmental Categorization Processes in the Evaluation of Urban Scenes.” International
Journal of Psychology 40.1 (2005): 19-26.

Gifford, Robert, Donald W. Hine, Werner Muller-Clemm, and Kelly T. Shaw. "Why Architects
and Laypersons Judge Buildings Differently: Cognitive Properties and Physical Bases."
Journal of Architectural and Planning Research 19.2 (2002): 131-148.

Goodman, Joseph K., Cynthia E. Cryder, and Amar Cheema. “Data Collection in a Flat World:
The Strengths and Weaknesses of Mechanical Turk Samples.” Journal of Behavioral
Decision Making 26 (2013): 213-224.

Henschen, Thomas and David B. Hershenson. “Values, Interests, and Architectural Preferences.”
Man-Environment Systems 5.4 (1975): 239-244.

Herzog, Thomas R. "A Cognitive Analysis of Preference for Urban Nature." Journal of
Environmental Psychology 9 (1989): 27-43.

Herzog, Thomas R. “A Cognitive Analysis of Preference for Urban Spaces.” Journal of
Environmental Psychology 12 (1992): 237-248.

Herzog, Thomas R. and Theresa A. Gale. “Preferences for Urban Buildings as a Function of Age
and Nature Context.” Environment and Behavior 28.1 (1996): 44-72.

Herzog, Thomas R., Lauren E. Gray, Amy M. Dunville, Angela M. Hicks, and Emily A. Gilson.
“Preference and Tranquility for Houses of Worship.” Environment and Behavior 45.4
(2011): 504-525.

Herzog, Thomas R., Stephen Kaplan, and Rachel Kaplan, "The Prediction of Preference for
Familiar Urban Places," Environment and Behavior 8 (1976): 627-645.

Herzog, Thomas R., Stephen Kaplan, and Rachel Kaplan. “The Prediction of Preference for
Unfamiliar Urban Places.” Population and Environment 5.1 (1982): 43-59.

Herzog, Thomas R. and Ronda L. Shier. “Complexity, Age, and Building Preference.”
Environment and Behavior 32.4 (2000): 557-575.

Huff, Connor and Dustin Tingley. “’Who are these People?’: Evaluating the Demographic
Characteristics and Political Preferences of MTurk Survey Respondents.” White paper,
Harvard University, Cambridge, MA, 2014.
109

IP Location. “Hostname / Reverse IP Lookup.” Accessed June 2014.
http://www.iplocation.net/index.php.

Ipeirotis, Panagiotis G. “Analyzing the Amazon Mechanical Turk Marketplace.” XRDS 17.2
(2010): 16-21.

Ipeirotis, Panagiotis G. “Demographics of Mechanical Turk.” White paper, Leonard N. Stern
School of Business, New York University, New York City, NY, 2010.

Levi, Daniel J. “Does History Matter? Perceptions and Attitudes Toward Fake Historic
Architecture and Historic Preservation.” Journal of Architectural and Planning Research
22:2 (2005).

Kaplan, Rachel and Stephen Kaplan. The Experience of Nature. Cambridge: Cambridge
University Press, 1989.

National Park Service. Historic Preservation Fund Grants Manual. Washington D.C.: National
Park Service, 2007.

National Park Service. “National Register Bulletin: How to Apply the National Register Criteria
for Evaluation.” Accessed May 5, 2014.
http://www.nps.gov/nr/publications/bulletins/nrb15/nrb15_2.htm.

National Park Service. “National Register of Historic Places Program: National Register Federal
Program Regulations.” Accessed June 22, 2014. http://www.nps.gov/nr/regulations.htm#604.

National Trust for Historic Preservation. “A Brief History of the National Trust.” Accessed
August 17, 2014. http://www.preservationnation.org/who-we-are/history.html.

National Trust for Historic Preservation. Field Guide to Local Preservationists: Learn Who They
Are, What They Do, and How to Win Their Support. Washington D.C.: National Trust for
Historic Preservation, 2011.

Nasar, Jack L. “Adult Viewers’ Preferences in Residential Scenes: A Study of the Relationship
of Environmental Attributes to Preference.” Environment and Behavior 15.5 (1983): 589-
614.

Nasar, Jack L. "Environmental Correlates of Evaluative Appraisals of Central Business District
Scenes." Landscape and Urban Planning 17 (1987): 117-130.

Nasar, Jack L. “The Evaluative Image of the City.” Journal of the American Planning
Association 56.1 (1990): 41-54.

Nasar, Jack L. “Visual Preferences in Urban Street Scenes: A Cross-Cultural Comparison
between Japan and the United States.” Journal of Cross-Cultural Psychology 15.1 (1984):
79-93.
110

Oostendorp, A. and D.E. Berlyne. "Dimensions of the Perception of Architecture: III.
Multidimensional Preference Scaling." Scandinavian Journal of Psychology 19.1 (1978):
145-150.

Paolacci, Gabriele, Jesse Chandler, and Panagiotis G. Ipeirotis. “Running Experiments on
Amazon Mechanical Turk.” Judgment and Decision Making 5.5 (2010), 411-419.

Preservation Green Lab. The Greenest Building: Quantifying the Environmental Value of
Building Reuse. Washington, D.C.: National Trust for Historic Preservation, 2011.

Preservation Green Lab. Older, Smaller, Better: Measuring How the Character of Buildings and
Blocks Influences Urban Vitality. Washington, D.C.: National Trust for Historic
Preservation, 2014.

Preservation Leadership Forum Blog. “Why Do Old Places Matter: An Introduction.” Last
modified November 13, 2013. http://blog.preservationleadershipforum.org/2013/11/13/old-
places-introduction/#.U_RC-Xck3Nk.

Preservation Leadership Forum Blog. “Why Do Old Places Matter: Architecture.” Last modified
May 23, 2014. http://blog.preservationleadershipforum.org/2014/05/23/old-places-matter-
architecture/#.U_Q9Hnck3Nk.

Preservation Leadership Forum Blog. “Why Do Old Places Matter: Creativity.” Last modified
August 13, 2014. http://blog.preservationleadershipforum.org/2014/08/13/old-places-
creativity/#.U_Q9KXck3Nk.

Rademacher, Eric W. The Ohio Poll – Project Report for: Ohio Historic Preservation Office,
Institute for Policy Research, University of Cincinnati. Columbus, OH: Ohio Historic
Preservation Office and University of Cincinnati, n.d.

Rypkema, Donovan, Caroline Cheong, and Randall Mason. Measuring Economic Impacts of
Historic Preservation: A Report to the Advisory Council on Historic Preservation.
Washington, D.C.: PlaceEconomics, 2011.

Sanoff, Henry. Visual Research Methods in Design. (New York, NY: Van Nostrand Reinhold,
1991.

Shapiro, Danielle N., Jesse Chandler, and Pam A. Mueller. “Using Mechanical Turk to Study
Clinical Populations.” Clinical Psychological Science 1.2 (2013): 213-220.

Stamps III, Arthur E. “Demographic Effects in Environmental Aesthetics: A Meta-Analysis.”
Journal of Planning Literature 14.2 (1999): 155-175.

Stamps III, Arthur E. "Formal and Nonformal Stimulus Factors in Environmental Preferences."
Perceptual and Motor Skills 79 (1994): 3-9.

111

Stamps III, Arthur E. “Mystery, Complexity, Legibility, and Coherence: A Meta-Analysis.”
Journal of Environmental Psychology 24 (2004): 1-16.

Stamps III, Arthur E. "Public Preferences for High Rise Buildings: Stylistic and Demographic
Effects." Perceptual and Motor Skills72 (1992): 839-844.

Stamps III, Arthur E. “Use of Photographs to Simulate Environments: A Meta Analysis.”
Perceptual and Motor Skills 71 (1990): 907-913.

Stamps III, Arthur E. and Jack L. Nasar. “Design Review and Public Preferences: Effects of
Geographical Location, Public Consensus, Sensation Seeking, and Architectural Styles."
Journal of Environmental Psychology 17 (1997): 11-32.

State of Hawai’i Historic Preservation Division. Hawai’i State Historic Preservation Plan:
October 2012 to October 2017. Hilo, HI: State Historic Preservation Division, 2012.

Sullivan, W. C. “Perceptions of the Rural-Urban Fringe: Citizen Preferences for Natural and
Developed Settings.” Landscape and Urban Planning 29 (1994): 85-101.

United States Census Bureau. “American Fact Finder.” Accessed April 8, 2014.
http://factfinder2.census.gov/faces/nav/jsf/pages/community_facts.xhtml.

United States Census Bureau. “Census Regions and Divisions of the United States.” Accessed
June 19, 2014. https://www.census.gov/geo/maps-data/maps/pdfs/reference/us_regdiv.pdf.

United States Census Bureau. “Frequently Asked Questions.” Accessed June 27, 2014.
https://ask.census.gov/faq.php?id=5000&faqId=5971.

Walker, Amanda J. and Robert L. Ryan. “Place Attachment and Landscape Preservation in Rural
New England: A Maine Case Study.” Landscape and Urban Planning 86.2 (2008).

Wells, Jeremy. “Attachment to the Physical Age of Urban Residential Neighborhoods: A
Comparative Case Study of Historic Charleston and I’on.” Ph.D. diss., Clemson University,
2009.

White, Michael J. “Determinants of Community Satisfaction in Middletown.” American Journal
of Community Psychology 13.5 (1985): 583-597.

Widgery, Robin N. “Satisfaction with the Quality of Urban Life: A Predictive Model.” American
Journal of Community Psychology 10.1 (1982): 37-48.

Widmar, Ronald Eugene. "Preferences for Multiple-Family Housing in Small City
Neighborhoods." Ph.D. diss., The University of Michigan, 1983.

Wilson, Margaret A. “The Socialization of Architectural Preference.” Journal of Environmental
Psychology 16 (1996): 33-44.
112

APPENDIX A: Results of the AIA America’s Favorite Architecture Survey

The results of the AIA study are presented here for future researchers because the study is no longer available on the
AIA website. Data reproduced from “America’s Favorite Architecture,” accessed November 12, 2013,
http://www.favoritearchitecture.org/afa150.php.

Ranking
Year
Built
Building Location Architect Age Category
1 1931 Empire State Building New York City, NY William Lamb Old (pre 1956)
2 1792 White House Washington DC James Hoban Old (pre 1956)
3 1990 Washington National
Cathedral
Washington DC George Frederick Bodley
and Henry Vaughn
Mid age (1957 - 1990)
4 1943 Thomas Jefferson Memorial Washington DC John Russell Pope Old (pre 1956)
5 1937 Golden Gate Bridge San Francisco, CA Irving F. Morrow and
Gertrude C. Morrow
Old (pre 1956)
6 1793 US Capitol Washington DC William Thorton;
Benjamin Henry Latrobe;
Charles Bulfinch; Thomas
U. Walter
Old (pre 1956)
7 1922 Lincoln Memorial Washington DC Henry Bacon Old (pre 1956)
8 1895 Biltmore Estate Asheville, NC Richard Morris Hunt Old (pre 1956)
9 1930 Chrysler Building New York City, NY William Van Alen Old (pre 1956)
10 1982 Vietnam Veterans Memorial Washington DC Maya Lin Mid age (1957 - 1990)
11 1878 St. Patrick's Cathedral New York City, NY James Renwick Old (pre 1956)
12 1884 Washington Monument Washington DC Robert Mills Old (pre 1956)
13 1913 Grand Central Terminal New York City, NY Reed and Stern; Warren
and Wetmore
Old (pre 1956)
14 1965 The Gateway Arch St. Louis, MO Eero Saarinen Mid age (1957 - 1990)
15 1935 Supreme Court of the United
States
Washington DC Cass Gilbert Old (pre 1956)
16 1904 St. Regis Hotel New York City, NY Trowbridge and
Livingston
Old (pre 1956)
17 1880 Metropolitan Museum of Art New York City, NY Calvert Vaux; Richard
Morris Hunt
Old (pre 1956)
18 1888 Hotel Del Coronado San Diego, CA James Reid Old (pre 1956)
19 1972 World Trade Center New York City, NY Minoru Yamasaki;
Antonio Brittiochi
Mid age (1957 - 1990)
20 1883 Brooklyn Bridge New York City, NY John Augustus Roebling;
Washington Roebling
Old (pre 1956)
21 1901 Philadelphia City Hall Philadelphia, PA John McArthur, Jr. Old (pre 1956)
22 1998 Bellagio Hotel and Casino Las Vegas, NV Deruyter Butler New (1991 - 2006)
23 1892 Cathedral of St. John the
Divine
New York City, NY Ralph Adams Cram Old (pre 1956)
24 1928 Philadelphia Museum of Art Philadelphia, PA Horace Trumbauer Old (pre 1956)
25 1877 Trinity Church Boston, MA Henry Hobson Richardson Old (pre 1956)
26 1928 Ahwahnee Hotel Yosemite National
Park, CA
Gilbert Stanley
Underwood
Old (pre 1956)
27 1770 Monticello Charlottesville, VA Thomas Jefferson Old (pre 1956)
28 1897 Library of Congress Washington DC John L. Smithmeyer and
Paul J. Pelz
Old (pre 1956)
29 1935 Edgar Kaufman Residence
(Fallingwater)
Bear Run, PA Frank Lloyd Wright Old (pre 1956)
30 1911 Taliesin Spring Green, WI Frank Lloyd Wright Old (pre 1956)
113

Ranking
Year
Built
Building Location Architect Age Category
31 1914 Wrigley Field Chicago, IL Zachary Taylor Davis Old (pre 1956)
32 1909 Wanamaker's Department
Store
Philadelphia, PA Daniel Burnham Old (pre 1956)
33 2000 Rose Center for Earth and
Science
New York City, NY James Stewart Polshek New (1991 - 2006)
34 1941 National Gallery of Art, West
Building
Washington DC John Russell Pope Old (pre 1956)
35 1886 Allegheny County
Courthouse
Pittsburgh, PA Henry Hobson Richardson Old (pre 1956)
36 1903 Old Faithful Inn Yellowstone National
Park, WY
Robert Reamer Old (pre 1956)
37 1903 Union Station Washington DC Daniel Burnham Old (pre 1956)
38 1925 Tribune Tower Chicago, IL John Mead Howells and
Raymond Hood
Old (pre 1956)
39 1947 Delano Hotel Miami Beach B. Robert Swartburg;
Interior Philippe Starck
Old (pre 1956)
40 1894 Union Station St. Louis, MO Theodore C. Link Old (pre 1956)
41 1947 Hearst Residence San Simeon, CA Julia Morgan Old (pre 1956)
42 1974 Sears Tower Chicago, IL Bruce Graham Mid age (1957 - 1990)
43 1882 Crane Library Quincy, MA Henry Hobson Richardson Old (pre 1956)
44 1913 Woolworth Building New York City, NY Cass Gilbert Old (pre 1956)
45 1933 Cincinnati Union Terminal Cincinnati, OH Albert Fellheimer and
Stewart Wagner; Paul
Philippe
Old (pre 1956)
46 1931 Waldorf Astoria New York City, NY Schultze and Weaver Old (pre 1956)
47 1911 New York Public Library New York City, NY Carrere and Hastings Old (pre 1956)
48 1891 Carnegie Hall New York City, NY William B. Tuthill;
Richard Morris Hunt;
Dankmar Adler
Old (pre 1956)
49 1915 San Francisco City Hall San Francisco, CA Arthur Brown, Jr. Old (pre 1956)
50 1788 Virginia State Capitol Richmond, VA Thomas Jefferson Old (pre 1956)
51 1962 Cadet Chapel Colorado Springs, CO Walter Netsch Mid age (1957 - 1990)
52 1909 Field Museum Chicago, IL Daniel Burnham Old (pre 1956)
53 2006 Apple Store Fifth Avenue New York City, NY Bohlin Cywinski Jackson New (1991 - 2006)
54 1888 Fisher Fine Arts Library Philadelphia, PA Frank Furness Old (pre 1956)
55 1967 Mauna Kea Beach Hotel Kohala Coast, Hawaii Edward Charles Bassett Mid age (1957 - 1990)
56 1932 Rockefeller Center New York City, NY Raymond Hood Old (pre 1956)
57 1995 Denver International Airport Denver, CO Fentress Bradburn
Architects
New (1991 - 2006)
58 1879 Ames Library North Easton, MA Henry Hobson Richardson Old (pre 1956)
59 2001 Milwaukee Art Museum Milwaukee, WI Santiago Calatrava New (1991 - 2006)
60 1980 Thorncrown Chapel Eureka Springs,
Arkansas
E. Fay Jones Mid age (1957 - 1990)
61 1972 Transamerica Pyramid San Francisco, CA William Pereira Mid age (1957 - 1990)
62 1983 333 Wacker Drive Chicago, IL William E. Pedersen Mid age (1957 - 1990)
63 1976 National Museum of Air and
Space
Washington DC Gyo Obata Mid age (1957 - 1990)
64 1978 Faneuil Hall Marketplace Boston, MA Benjamin Thompson Mid age (1957 - 1990)
65 1980 Chrystal Cathedral Garden Grove, CA Phillip Johnson Mid age (1957 - 1990)
114

Ranking
Year
Built
Building Location Architect Age Category
66 1908 Gamble House Pasadena, CA Greene and Greene Old (pre 1956)
67 1922 Nebraska State Capitol Lincoln, NE Bertram Grosvenor
Goodhue
Old (pre 1956)
68 2006 New York Times Building New York City, NY Renzo Piano New (1991 - 2006)
69 2003 Salt Lake City Public Library Salt Lake City, Utah Moshe Safdie New (1991 - 2006)
70 1990 Dolphin and Swan Hotels Orlando, FL Michael Graves Mid age (1957 - 1990)
71 1927 Hearst Tower New York City, NY George P. Post and Son;
Addition Foster and
Partners
Old (pre 1956)
72 1903 Flatiron Building New York City, NY Daniel Burnham Old (pre 1956)
73 1968 Lake Point Tower Chicago, IL Schipporeit - Heinrich,
graham, Anderson, Probst
and White
Mid age (1957 - 1990)
74 1959 Guggenheim Museum New York City, NY Frank Lloyd Wright Mid age (1957 - 1990)
75 1939 Union Station Los Angeles, CA John Parkinson and
Donald B. Parkinson
Old (pre 1956)
76 1901 Willard Hotel Washington DC Henry Janeway
Hardenbergh
Old (pre 1956)
77 1880 Sever Hall Cambridge, MA Henry Hobson Richardson Old (pre 1956)
78 1918 Broadmoor Hotel Colorado Springs, CO Warren and Wetmore Old (pre 1956)
79 1998 Ronald Reagan Building and
International Trade Center
Washington DC Pei Cobb Freed and
Partners
New (1991 - 2006)
80 1972 Phillips Exeter Academy
Library
Exeter, NH Louis I. Kahn Mid age (1957 - 1990)
81 1907 The Plaza Hotel New York City, NY Henry Janeway
Hardenbergh
Old (pre 1956)
82 2002 Sofitel Chicago Chicago, IL Jean-Paul Viguier New (1991 - 2006)
83 1887 Glessner House Chicago, IL Henry Hobson Richardson Old (pre 1956)
84 1923 Yankee Stadium New York City, NY Osborn Architects and
Engineers
Old (pre 1956)
85 1991 Harold Washington Library Chicago, IL Hammond, Beeby and
Babka
New (1991 - 2006)
86 1962 Lincoln Center New York City, NY Wallace K Harrison Mid age (1957 - 1990)
87 1884 The Dakota Apartments New York City, NY Henry Janeway
Hardenbergh
Old (pre 1956)
88 1893 Art Institute of Chicago Chicago, IL Shepley, Rutan and
Coolidge
Old (pre 1956)
89 1906 Fairmont Hotel San Francisco, CA Julia Morgan Old (pre 1956)
90 1895 Boston Public Library Boston, MA McKim, Mead and White Old (pre 1956)
91 1924 Hollywood Bowl Hollywood, CA Lloyd Wright; Allied
Architects; Hodgetts and
Fung Design Associates
with Gruen Associates
Old (pre 1956)
92 1888 Texas Capitol Austin, TX Elijah E. Myers Old (pre 1956)
93 1954 Fontainebleau Hotel Miami Beach, FL Morris Lapidus Old (pre 1956)
94 1931 Legal Research Building, U
Michigan
Ann Arbor, MI Gunnar Birkerts Old (pre 1956)
95 1997 Getty Center Los Angeles, CA Richard Meier New (1991 - 2006)
96 1983 High Museum Atlanta, GA Richard Meier; addition
Renzo Piano
Mid age (1957 - 1990)
97 2000 Federal Building at Islip Islip, NY Richard Meier New (1991 - 2006)
115

Ranking
Year
Built
Building Location Architect Age Category
98 1986 Humana Building Louisville Michael Graves Mid age (1957 - 1990)
99 2003 Disney Concert Hall Los Angeles, CA Frank Gehry New (1991 - 2006)
100 1932 Radio City Music Hall New York City, NY Edward Durell Stone Old (pre 1956)
101 2000 Paul Brown Stadium Cincinnati, OH NBBJ New (1991 - 2006)
102 1988 United Airlines Terminal,
O'Hare
Chicago, IL Helmut Jahn Mid age (1957 - 1990)
103 1967 Hyatt Regency Atlanta Atlanta, GA John C. Portman Mid age (1957 - 1990)
104 2000 AT & T Park San Francisco, CA Hellmuth, Obata,
Kassabaum
New (1991 - 2006)
105 2003 Warner Center New York City, NY David Childs New (1991 - 2006)
106 1976 Washington DC Metro Washington DC Harry Weese Mid age (1957 - 1990)
107 1972 IDS Center Minneapolis, MN Phillip Johnson Mid age (1957 - 1990)
108 2004 Seattle Public Library Seattle, WA Rem Koolhaas New (1991 - 2006)
109 1995 SFMOMA San Francisco, CA Mario Botta New (1991 - 2006)
110 1925 Union Station Chicago, IL Daniel Burnham Old (pre 1956)
111 1947 United Nations Headquarters New York City, NY Wallace K Harrison;
Oscar Niemeyer;
LeCorbusier
Old (pre 1956)
112 1887 National Building Museum Washington DC Montgomery C. Meigs Old (pre 1956)
113 1912 Fenway Park Boston, MA Osborn Architects and
Engineers
Old (pre 1956)
114 1904 Dana-Thomas House Springfield, IL Frank Lloyd Wright Old (pre 1956)
115 1962 TWA Terminal, Kennedy
Airport
New York City, NY Eero Saarinen Mid age (1957 - 1990)
116 1979 The Athenaeum New Harmony, IN Richard Meier Mid age (1957 - 1990)
117 2005 Walker Art Center Minneapolis, MN Edward Larrabee Barnes New (1991 - 2006)
118 2001 American Airlines Center Dallas, TX David M Schwarz New (1991 - 2006)
119 1929 Arizona Biltmore Resort and
Spa
Phoenix, AZ Albert Chase McArthur Old (pre 1956)
120 1926 Los Angeles Central Library Los Angeles, CA SOM; Del Campo and
Maru Architects; Michael
Willis Architects
Old (pre 1956)
121 2002 San Francisco International
Terminal
San Francisco, CA Bertram Grosvenor
Goodhue
New (1991 - 2006)
122 1992 Oriole Park at Camden Yards Baltimore, MD Bertram Michael Willis
Architects; Grosvenor
Goodhue
New (1991 - 2006)
123 1937 Telesin West Scottsdale, AZ Frank Lloyd Wright Old (pre 1956)
124 1993 US Holocaust Memorial
Museum
Washington DC James Ingo Freed New (1991 - 2006)
125 1977 Citicorp Center New York City, NY Hugh Stubbins and
Assoc., Emery Roth and
Sons
Mid age (1957 - 1990)
126 1948 V.C. Morris Gift Shop San Francisco, CA Frank Lloyd Wright Old (pre 1956)
127 1914 Union Station Kansas City, MO Jarvis Hunt Old (pre 1956)
128 1888 Rookery Building Chicago, IL Burnham and Root Old (pre 1956)
129 1990 Weisman Art Museum Minneapolis, MN Frank Gehry Mid age (1957 - 1990)
130 1973 Douglas House Harbor Spring, MI Richard Meier Mid age (1957 - 1990)
131 1917 Hollyhock House Los Angeles, CA Frank Lloyd Wright Old (pre 1956)
116

Ranking
Year
Built
Building Location Architect Age Category
132 1976 Pennzoil Place Houston, TX Phillip Johnson Mid age (1957 - 1990)
133 1988 Royalton Hotel New York City, NY Gruzen Samton
Steinglass; interior
Philippe Stack
Mid age (1957 - 1990)
134 1964 Reliant Astrodome Houston, TX Hermon Lloyd and W.B.
Morgan
Mid age (1957 - 1990)
135 1999 Safeco Field Seattle, WA NBBJ New (1991 - 2006)
136 1951 Corning Museum of Glass Corning, NY Gunnar Birkerts Old (pre 1956)
137 1934 30th Street Station Philadelphia, PA Graham, Anderson, Probst
and White
Old (pre 1956)
138 1901 Robie House Chicago, IL Frank Lloyd Wright Old (pre 1956)
139 1979 Williams Tower (Transco
Tower)
Houston, TX Phillip Johnson Mid age (1957 - 1990)
140 1959 Stahl House Los Angeles, CA Pierre Koenig Mid age (1957 - 1990)
141 2002 Apple Soho New York City, NY Bohlin Cywinski Jackson;
Ronnett Riley Architect
New (1991 - 2006)
142 1976 John Hancock Towers Boston, MA Henry Cobb Mid age (1957 - 1990)
143 1910 Penn Station New York City, NY McKim, Mead and White Old (pre 1956)
144 1973 Hyatt Regency San Francisco San Francisco, CA John C. Portman Mid age (1957 - 1990)
145 1903 Carson Pirie Scot Chicago, IL Louis Sullivan Old (pre 1956)
146 1939 MOMA New York City, NY Philip Goodwin and
Edward Durell Stone
Old (pre 1956)
147 1889 Auditorium Building Chicago, IL Adler and Sullivan Old (pre 1956)
148 1892 Brown Palace Hotel Denver, CO Frank E. Edbrooke Old (pre 1956)
149 1958 Ingalls Arena New Haven, CT Eero Saarinen Mid age (1957 - 1990)
150 1911 Battle Hall, UT Austin, TX Cass Gilbert Old (pre 1956)

117

APPENDIX B: Survey Instrument

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

APPENDIX C: Mean Preference Ratings for Old and New Buildings, Ranked from 1 to 50

1. Old building. Mean: 5.35

2 (tied). New building. Mean: 5.07

2 (tied). Old building. Mean: 5.07

4. Old building. Mean: 5.02

5. Old building. Mean: 4.93

6 (tied). Old building. Mean: 4.77

6 (tied). Old building. Mean: 4.77

8. Old building. Mean: 4.76
Photo credits from top left: Adam Smith, Brandon Bartoszek, Brandon Bartoszek, Adam Smith, Marc Belanger,
Paige Miller, Flickr – OzinOH, Phil Squattrito.
181

9. Old building. Mean: 4.61

10 (tied). Old building. Mean: 4.58

10 (tied). Old building. Mean: 4.58

12. Old building. Mean: 4.34

13. Old building. Mean: 4.25

14. Old building. Mean: 3.89

15. Old building. Mean: 3.83

16. New building. Mean: 3.81
Photo credits from top left: Phil Squattrito, John Chambers, Jr, Adam Smith, Jim Peacock, Adam Smith, Brandon
Bartoszek, Sandra Shannon, Dan Bertolet.

182

17. New building. Mean: 3.78

18. New building. Mean: 3.75

19. Old building. Mean: 3.74

20. Old building. Mean: 3.67

21. Old building. Mean: 3.57

22. Old building. Mean: 3.55

23. New building. Mean: 3.48

24. Old building. Mean: 3.45
Photo credits from top left: Sandra Shannon, Sandra Shannon, Flickr – OzinOH, Marc Belanger, Brandon
Bartoszek, Brandon Bartoszek, Sandra Shannon, Sandra Shannon.

183

25. New building. Mean: 3.38

26. New building. Mean: 3.31

27. New building. Mean: 3.29

28. Old building. Mean: 3.28

29. New building. Mean: 3.22

30. New building. Mean: 3.18

31. New building. Mean: 3.14

32 (tied). New building. Mean: 3.12
Photo credits from top left: Brian Phelps and Chris Whitis, Sandra Shannon, Geomorph, Sandra Shannon, SOLA
MOB, Sandra Shannon, www.flickr.com/photos/7542656@N02/1150704543/, Sandra Shannon.

184

32 (tied). New building. Mean: 3.12

34. New building. Mean: 3.10

35. Old building. Mean: 3.07

36. New building. Mean: 3.02

37. New building. Mean:2.91

38 (tied). New building. Mean: 2.88

38 (tied). New building. Mean: 2.88

40. New building. Mean: 2.86
Photo credits from top left: Sandra Shannon, Brandon Bartoszek, Sandra Shannon, Sandra Shannon, Sandra
Shannon, Sandra Shannon, Lea Suzuki / The San Francisco Chronicle, Sandra Shannon.

185

41. New building. Mean: 2.84

42. Old building. Mean: 2.79

43. New building. Mean: 2.62

44. New building. Mean: 2.58

45. New building. Mean: 2.57

46. New building. Mean: 2.43

47. New building. Mean: 2.31

48. Old building. Mean: 2.16
Photo credits from top left: Sandra Shannon, Brandon Bartoszek, Sandra Shannon, Sandra Shannon, Yichen Lin,
Sandra Shannon, Robert Benson / 4240 Architecture, Sandra Shannon.

186

49. Old building. Mean: 2.05

50. Old building. Mean: 1.90
Photo credits: Sandra Shannon.

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Historic preservation in the United States Air Force: exploring new frontiers

PDF

Preserving California City: an exploration into the city plan preservation of a mid-century, master-planned community

PDF

Heritage conservation in post-redevelopment Los Angeles: evaluating the impact of the Community Redevelopment Agency of the City of Los Angeles (CRA/LA) on the historic built environment

PDF

Sawtelle reexamined: a preservation study for a historic California Japantown

PDF

SoCal ski hills: a typological analysis of a cultural landscape

PDF

Engaging today's youth at house museums: forging a vision for the Gamble House

PDF

Celebrating conformity: preserving Henry Doelger's midcentury post-war suburb

PDF

Mobilizing heritage conservation as a tool for urban resilience: linkages and recommendations

PDF

Empowering communities through historic rehabilitation: creating a maintenance plan for public housing developments in Los Angeles

PDF

Would community design overlay (CDO) be a tool for revitalizing Los Angeles Chinatown?

PDF

Maintaining historic integrity and solving a rehabilitation dilemma: the history of hollow clay tile and an argument for its preservation

PDF

Collaborative conservation: how the heritage and environmental movements can and should unite for progress

PDF

Defining the City of Gardens: the conservation of Pasadena's bungalow courts

PDF

Being a part of the narrative: how can we preserve Black heritage in the U.S. South while mitigating violence and facilitating change?

PDF

Capturing the layers of the Arroyo Seco landscape: documenting a cultural landscape using digital storymaps

PDF

Conserving Compton: Identifying potential landmarks and recommendations for conservation

PDF

"It's important to remember what started it": conserving sites and stories of racial violence in Los Angeles, 1943-1992

PDF

Money in the bank: three case studies in the adaptive reuse of midcentury bank branch buildings

PDF

Identifying and conserving Pacoima: a heritage conservation study of a minority enclave in the San Fernando Valley

PDF

Citius, altius, fortius: filling a void in the identification and designation of historic venues from the 1932 Los Angeles Olympics

Asset Metadata

Creator Shannon, Sandra (author)

Core Title A survey of the public: preference for old and new buildings, attitudes about historic preservation, and preservation-related engagement

School School of Architecture

Degree Master of Heritage Conservation

Degree Program Heritage Conservation

Publication Date 09/19/2014

Defense Date 09/05/2014

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

Tag Architecture,attitudes,engagement,Heritage Conservation,Historic Preservation,OAI-PMH Harvest,Planning,visual preference

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sandmeier, Trudi (committee chair), Von Winterfeldt, Detlof (committee member), Wells, Jeremy (committee member)

Creator Email sandra.shannon@gmail.com,sandrash@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-482725

Unique identifier UC11287131

Identifier etd-ShannonSan-2963.pdf (filename),usctheses-c3-482725 (legacy record id)

Legacy Identifier etd-ShannonSan-2963.pdf

Dmrecord 482725

Document Type Thesis

Format application/pdf (imt)

Rights Shannon, Sandra

Type texts

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

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

Repository Name University of Southern California Digital Library

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