Running head: GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE

CORRELATION BETWEEN NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Gentrification and Education at a Crossroads:

The Correlation between Neighborhood Change and Student Achievement

Tasia S. Harris

The University of North Carolina at Chapel Hill

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 2
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Abstract

This thesis explores the relationship between gentrification and student achievement in

New York State. Gentrification, established by Ruth Glass (1964) as a class issue contextualized

by urban geographies, has historically been investigated from the perspective of the gentrifier.

Much of the early gentrification research has been focused on defining the gentrifier and the factors

that cause gentrification to come about. More recently, researchers have begun to turn their

attention to the impact on the existing populations where gentrification takes place. However, these

efforts are primarily concerned with the impact gentrification has on the accessibility of affordable

housing or how it affects the business landscape. The literature on its relationship with education

is very scarce.

This thesis involves a quantitative analysis of the relationship between gentrification (as

measured by median income, educational attainment, and racial composition) and student

achievement (as measured by mean test scores on third grade statewide standardized tests and third

grade proficiency rates on statewide standardized tests) in New York from 2006-2012. After a

series of bivariate regressions and two multivariate regressions, I find that there is no statistically

significant impact of measures of gentrification on math scores and proficiency; the impact of

gentrification on English Language Arts scores and proficiency, while statistically significant has

no practical influence’ and; lastly, over time student achievement is decreasing in New York.

Although the findings regarding gentrification are insightful they are limited by both time and

space. I suggest that future quantitative research be conducted for longer periods of time with more

aspects of the design aimed at isolating the “stayers,” those who lived in areas prior to the onset

gentrification. Lastly, I urge policymakers in New York to act to reverse the trend of decreasing

student achievement.
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 3
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Acknowledgements

I would like to thank my advisor Dr. Steven Hemelt for his encouragement and guidance

from the beginning to the end of this process. His continued support helped me throughout the

conception, research, and writing of this thesis. I could not have imagined a better advisor and

mentor for my undergraduate research experience.

I would also like to thank my friends and family who have provided emotional support

inspiration, and encouragement throughout my time at Carolina and without whom this thesis

would not be possible.

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 4
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Table of Contents

Chapter 1: Introduction…………………………………………………………………………5
Context and Significance………………………………………………………………….5
Research Question………………………………………………………………………...7
Roadmap…………………………………………………………………………………..8
Chapter 2: Literature Review…………………………………………………………………...9
Core Tenets of Gentrification Research…………………………………………………...9
Quantifying Student Achievement……………………………………………………….12
Gentrification and Student Achievement………………………………………………...14
Qualitative Research……………………………………………………………..15
Quantitative Research……………………………………………………………17
Significance and Implications of this Study……………………………………………..18
Chapter 3: Data and Methods………………………………………………………………....20
Data Sources……………………………………………………………………………..20
Variables of Interest……………………………………………………………………...21
Analysis………………………………………………………………………………….22
Bivariate Regression……………………………………………………………..22
Multivariate Regression #1………………………………………………………22
Multivariate Regression #2………………………………………………………23
Measurement……………………………………………………………………………..23
Chapter 4: Results………………………………………………………………………………25
General Findings…………………………………………………………………………25
Effect by Subject…………………………………………………………………………27
Effect of Gentrification on Student Achievement as Measured by Mean Test Scores…..27
Effect of Gentrification on Student Achievement as Measured by Proficiency…………30
Chapter 5: Conclusions………………………………………………………………………...32
The Relationship Between Gentrification and Student Achievement…………………...32
Limitations……………………………………………………………………………….33
Policy Implications………………………………………………………………………34
References……………………………………………………………………………………….35

Table of Figures

Table 1: Methods of Measuring Student Achievement…………………………….……………13

Table 2: Association Between Measures of Gentrification and Reading and Math Test Scores for
39 counties in New York State Over 7 Years (2006 to 2012)……………………………..….....26
Table 3: Association Between Measures of Gentrification and Reading and Math Proficiency for
39 Counties in New York State Over 7 Years (2006-2012)………………………………....…..28
Table 4: Descriptive Statistics of New York State Over 7 Years (2006-2012)……........……….31
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 5
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Chapter 1: Introduction

Context and Significance

In the summer of 2014, protesters lined up on the busy street of Flatbush Avenue in

Brooklyn, New York, inhibiting the construction of a 23-story high-rise luxury condominium. Just

a few weeks prior, they had gathered at the steps of City Hall voicing their concerns surrounding

the implications of this development project. One protestor Leah Margulies asked of City Hall and

the development companies, “our median income is $41,000 for a family of four—does it make

sense to insert a building where studios cost $1900 [per month]?” (Curbed, 2014). Though they

were able to secure a restraining order to halt the construction, a month later, a judge lifted the

restraining order allowing for the continued development of the tower.

This story is not unique to Flatbush Avenue. To answer Ms. Margulies question, building

studios that cost $1900 per month makes sense when they are meant to attract new residents who

contribute to a steadily increasing median income. In the Williamsburg neighborhood of Brooklyn,

NY, the median income range rose from $25k to $55k in 1990 to $53k to $80k by 2012. The West

Harlem/Morningside Heights neighborhood of Manhattan experienced a similar upward trend in

median income over the same 22 years. Heat maps created by the United States Census Bureau

show that over this same time frame, cities like Oakland, Washington, DC and Philadelphia have

seen their median household income increase at similar rates. These changes in median income

and the rate of development projects are just two indicators of the larger urban geographical

phenomenon of gentrification.

Gentrification, the process by which working class neighborhoods are transformed into

upper middle class neighborhoods, is arguably one of the most well studied processes of urban
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 6
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

development and urban policy. Spanning several spheres of social issues, such as demographic

composition, housing, and economics, it has drawn numerous researchers over the past four to five

decades to examine its characteristics, causes, and, more recently, its impact within these different

realms of social policy. While researchers have examined the implications gentrification holds for

affordable housing, the business landscape, or the tension it inevitably creates between original

residents and newcomers as neighborhoods undergo significant demographic changes, there has

been very little done to examine gentrification in relation other equally pertinent public issues.

Originally described by sociologist Ruth Glass as a class issue embedded in the context of

urban geographies (Glass, 1964), gentrification has continued to be depicted as a class issue. Issues

of class are not without warrant for change, but gentrification as a class issue has not garnered

enough public attention to demand that policymakers take an active role in mitigating any

consequences of these processes of neighborhood change. In addition to its reputation as a class

issue, part of this neglect to address gentrification as an issue of public policy is due to the fact that

the process has historically been seen from one perspective: that of the gentrifier. Early research

on gentrification was primarily focused on describing the phenomenon and determining the cause

of what seemed to be irrational behavior on the part of young wealthy individuals (Kozak 2014).

This perspective did not speak to what the impacts of these class-based geographical decisions

might be on people and structures of existing working class neighborhoods.

More recent research and discussions have provided two new compelling perspectives on

gentrification. The first is the perspective that seeks to fill this gap in the work done to examine

the impact of gentrification as it relates to existing working class neighborhoods. Some researchers

have examined the social tensions that result from such abrupt demographic changes (Hwang,

2014). Several others have detailed the impact that gentrification has on the housing sphere,
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 7
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

namely its role in the elimination of affordable housing and the displacement of working class and

minority individuals (Kozak, 2014). The second perspective has undertaken the investigation of

the role of the public sphere in facilitating the advancement of gentrification, as there exist certain

policies, or lack thereof, that provide a platform for housing and business development projects to

draw young upper class individuals into traditionally working class neighborhoods (Bates, 2013).

These two major additions to the gentrification debate have moved the discourse away from its

descriptive tradition towards a critical perspective that demands the attention of policymakers.

Research Question

The goal of this thesis is to examine the impact of gentrification on the academic

achievement of public school students. I will examine the association between specific indicators

of gentrification (median income, racial composition, and education attainment levels) and student

achievement (standardized test scores in third grade English Language Arts and Math) in counties

across New York State over a period of seven years (2006 through 2012). These three indicators

of gentrification capture common themes in the characterization of gentrification: the influx of

upper middle class, predominantly white, highly educated individuals. In terms of student

achievement, the purpose of using third grade test scores is to capture the impact of the independent

variables as much as possible by focusing on students who would have standardized achievement

data, but who would also be young enough to not have experienced several years of schooling

before gentrification began (Keels, Burdick-Will, & Keene, 2013).

This study makes both theoretical and practical contributions to the literature. First, this

study will help expand the growing and changing conception of gentrification as it determines

metrics by which to measure gentrification. Secondly, the analysis of gentrification’s impact on

student performance has not yet been widely studied, allowing this thesis to contribute to an
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 8
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

emergent subset of gentrification studies. Lastly, if I find that these processes are having any

impact at all on student achievement in public schools, it will reaffirm claims made by researchers

in the housing, economic, and social realms that gentrification is a public issue to be addressed in

policy.

Roadmap

The following chapters of this these will provide a more detailed synthesis of salient

gentrification research; an overview of the data sources, variables, and methods by which they will

be analyzed; the results of this analysis; and final thoughts on the findings and implications for

policymakers moving forward. Chapter 2 reviews prior research on gentrification with regard to

its evolving definition, the changing perspective of researchers, as well as an overview of the

quantitative and qualitative work done with special attention to research that examines the

intersection between gentrification and education. Chapter 3 discusses the data methods and

sources that will be employed in this thesis. Chapter 4 will include a description and analysis of

my results and statistical tests. It will also discuss the relevance, strength, and weaknesses of my

findings as well as contextualize them in the broader scope of earlier studies discussed in Chapter

2. Chapter 5 will provide a summary of my findings.

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 9
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Chapter 2: Literature Review

Core Tenets of Gentrification Research

Glass’ idea of gentrification as a class issue that plays out within the context of urban

geographies (Glass, 1964) has remained at the root of gentrification studies and research. Over the

past several decades, researchers, regardless of the goals of their studies, have continued to define

and measure gentrification as a function of class difference and displacement. Since the term’s

debut in Glass’ 1964 piece, researchers have rearticulated her findings. Zukin (1987) defines

gentrification as, “the conversion of socially marginal and working-class areas of the central city

to middle-class residential use” (Zukin, 2). Atkinson (2000) in a study on displacement in London

defines it as, “a process of class succession and displacement in areas broadly characterized by

working-class and unskilled households” (Atkinson, 1). Later Baxter (2009), continued in Glass’

trend by defining the process as

…the ‘upgrading’ of geographic space so that it reflects middle class

values…gentrification is a gradual process, occurring over years and even decades. It
entails middle class households moving into a disinvested, economically-depressed area,
buying real properties, renovating them and, by virtue of doing so, increasing property
values. The property appreciation in turn increases contract rents and property tax bills,
often resulting in a lower class household displacement…(Baxter, 2009)

As recently as 2014, scholars defined gentrification as a process that “traditionally occurs when

middle-class families move in and renovate economically depressed, inner-city neighborhoods,

which often results in the displacement of the existing, working-class residents” (Kozak, 2014).

That this trend of working class displacement by the middle class ever transpired, and

continues to occur, defies neoclassical urban models of the Chicago School. These models

predicted a “natural process” whereby consumers would sacrifice distance to urban centers for

cheaper land in the surrounding suburbs (Kozak, 2014). Often called the “Back to the City
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 10
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Movement” this anomaly initiated what is now another prominent aspect of gentrification research:

defining the type of middle class individuals who would choose to go against the natural process

and occupy these spaces.

This recognition that gentrification is comprised of transformations that encompass both

physical changes (i.e. deindustrialization, suburbanization, and the disinvestment in ethnic

enclaves and lower class areas of the inner city), as well as social and cultural changes (i.e. the

unexpected preferences of the middle class), prompted David Ley to forge the term “the new

middle class” (Lees, 2012). Though David Ley’s gentrification research focused primarily on

Canadian cities, his findings are applicable to international gentrification trends. In a more recent

piece since Ley first posited the term, Ley describes the new middle class using a stage model:

In the model, a lower-middle class of professionals and pre-professionals (students) seeking

inexpensive housing in the inner city comprises the first wave of gentrification. The presence of
artists among this group has been a strong predictor of subsequent gentrification (Ley, 1996, 2003).
Artists gravitate towards central locations with low rents and high degrees of social diversity. Soon
after their arrival, they are joined by a broader stratum of social and cultural workers, the ‘cultural
new class’ including arts, design and media workers, educators and social and health care workers.
Although generally highly educated professionals, this sub-group of the new middle class shares
something of the artists’ valuation of bohemian landscapes and urban authenticity (Lloyd, 2006).
These residents, in turn, are succeeded in a third stage by increasingly affluent gentrifiers in an
ascending economic hierarchy, including corporate lawyers, medical specialists, business people
and capitalists (Danyluk & Ley, 2006).

The idea of the new middle class as a progression along a gradient of groups with different

characteristics and preferences provide a method of measuring the process of gentrification that is

primarily driven by the individual.

The establishment of a widely accepted definition of gentrification and of the general

characteristics and demographics of gentrifiers is crucial to research on gentrification, as efforts to

measure its beginnings, rate of neighborhood change, or its impact depend on a clear definition of

the phenomenon as well as its agents of change. Though it seems that class based explanations of

this phenomenon and Ley’s emphasis on the preferences of the gentrifiers suffice, differences in
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 11
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

thought regarding the factors driving middle class preferences as well as the unique role of race in

the United States, complicate both the definition of gentrification and theories of its cause.

Due to policies and practices that have historically favored whites over blacks and other

people of color in the United States, income and class divisions fall along racial lines. This frequent

intersection of race and other social identifiers means that gentrification is not only witnessed and

interpreted as a class struggle, but equally as one of color, where neighborhoods with

predominantly minority residents become predominantly white in tandem with the class status

change of the neighborhoods.

The relationship between race and other social identifiers that have been used to account

for gentrification (i.e. class status and income) has prompted several researchers to also include

race as a measure of gentrification. Glick (2008) highlights the intersection of class and race as he

explores the impact of gentrification by white middle class individuals on Black and Latino

homeowners. Hwang (2014) echoes this intersection as she states “gentrification in US cities has

been problematic for low-income minorities” (Hwang, 2). Formoso, Weber, & Atkinson (2010)

also conflate class and race in their study on the impact of gentrification on children’s well-being:

…before change occurs, the condition of a gentrifying neighborhood is deficient relative

to other parts of an urban area in terms of median household income, aggregate property
value, and crime, and vacancy rates…these neighborhoods, prior to the change, often were
predominantly comprised of ethnic minorities (Formoso, Weber & Atkinson, 2010).

The persistent intersection of race and class in gentrification research suggests that race remains

as much a key indicator of gentrification as class and income.

In addition to race, defining gentrification is also complicated by the competing ideas of

production-based theories and consumption-based theories of the cause of gentrification.

Consumption-based theories reaffirm what Ley has declared as the agent of gentrification:

preferences of middle class individuals. Production-based theories, however, point to conditions

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 12
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

created by capitalism as the factor driving investors towards gentrification (Kozak, 2014).

Billingham (2013) defines gentrification as “the manifestation of efforts by municipal

governments to spur economic development by making their cities more amenable to use by

members of the professional classes and by the businesses that employ them.” In this production-

based conceptualization of gentrification, rather than individuals and their preferences driving

gentrification, conditions created by the government and subsequent economic development drives

gentrification.

Given the various social, cultural, and economic factors that contribute to gentrification,

any definition or measurement of the process must simultaneously account for categorical

indicators, quantitative measures, (i.e. race, class, income), and resolve or control for the

possibility that the agent of gentrification could be either individual or structural.

Quantifying Student Achievement

One of the major contributions of this study is its intent to expand the scope of

gentrification research to include the potential impact of gentrification on education. An obvious

indicator of the quality of education is student achievement. Best practices concerning the

measurement of student achievement continue to be contested today, as student achievement is

used to characterize school quality, teacher effectiveness, or factors beyond the school setting, like

neighborhood safety.

In 2012, the Center for Educator Compensation Reform of the US Department of Education

published Understanding the Basics of Measuring Student Achievement. This piece outlines broad

categories of student achievement measurement models: 1) student attainment, 2) gain, 3)

percentile growth, 4) standard value-added, 5) customized value-added, and 6) student learning

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 13
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

objectives (SLOs) (Scott and Miller, 2012). Table 1 below portrays each method as described by

the authors:

Table 1: Methods of Measuring Student Achievement

(1) Attainment “Attainment scores reflect student performance on a

particular assessment at a single point in time. These
measures are easy to compute and widely used in school
systems to determine performance related to
benchmarks, such as Adequate Yearly Progress (AYP).”
(2) Gain “…gain measures take a longitudinal approach. Gain in
student test scores is the difference between student
performance on a post-test and the same group of
students’ performance on the corresponding pre-test.”
(3) Percentile Growth “Student growth percentiles…compare test score growth
across groups of academic peers, which are students with
similar test score histories in the same grade and
subject…percentile growth [ranks] each student’s
growth with all other student who have similar student
achievement histories.”
(4) Standard Value-Added “By controlling for prior student test scores, standard
models will take into account some of the non-school
factors that contribute to student achievement.”
(5) Customized Value Added “Unlike standard value-added models, customized value-
added models may consider the effects of non-school
factors that contribute to student achievement in specific
states and/or districts…value-added may also take into
account many…school and classroom level factors…”
(6) Student Learning Objectives “…SLOs are goals set by teachers that specify what
(SLOs) students will know, or able to perform, after completing
a quarter, semester, or school year.”

Source: US Department of Education - Center for Educator Compensation Reform (2012)

Although Scott and Miller describe approaches of assessing student achievement holding

educator effectiveness as the independent variable, understanding the different methods available

for measuring student achievement are especially relevant to this study, where the intention is to

measure the impact of gentrification on student achievement. Having a reliable, quantifiable

method of measuring student achievement will allow for a strong analysis. While some

approaches, such as gain models, value added, or percentile growth models, might capture the
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 14
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

changes in student achievement over a longer period of time, my research will be conducted in a

relatively short period of time and the conditions required of these models (pre- and post-testing

of the same group of students, or the ability to compare test growth across academic peers) are not

achievable in a shorter period of time. Additionally, these methods are likely to have significant

internal validity gaps, as attrition rates may be of concern in gain models, value-added models

have come under scrutiny for issues of accurate implementation (Amrein-Beardsley, 2012; Bonk

et al, 2012) and teacher portfolios are extremely vulnerable to the personal bias of teachers.

Among research that uses student achievement as a dependent variable, the most common

methods for measuring student achievement, regardless of the independent variable, are student

test scores. This study will also follow the practice of student achievement measures as it seeks to

explore the quantifiable impact of gentrification on student achievement. Test scores provide

accessible, quantifiable resources for gauging student achievement that can be incorporated into a

quantitative analysis

Gentrification and Student Achievement

From the 1950s to today, researchers have explored the impact of gentrification on the

urban housing landscape, examining the implications for affordable housing or the prevalence of

high development projects (Kozak, 2014; Billingham, 2013). Research on gentrification has also

studied the influence it has on business and the economics of the affected neighborhood. However,

there is very little work on the impact gentrification has on neighborhood schools or student

achievement specifically. My goal is to ask what this phenomenon means for the students attending

schools in areas where there is an influx college educated, middle class individuals. How might

their achievement be impacted by these demographic changes in their neighborhood?

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 15
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

The two strands of research on gentrification and education are divided by whether the

researchers conduct a qualitative analysis or quantitative analysis. Across the limited work on this

topic there are no real common themes or methodologies as the purpose of each researchers work

varies significantly.

Qualitative Research

A prime example of the qualitative research on the intersection of gentrification and

education has been conducted by urban sociologist, Judith N. DeSena. DeSena (20092) seeks to

study, qualitatively, the impact of gentrifying parents on schools. Her research explored the

manifestations of social class distinctions in the Greenpoint neighborhood of Brooklyn, New York

in its initial stages of gentrification, i.e. during the period of transition from residents of primarily

low-income, working class demographics, to largely middle class residents. Although most

research portrays gentrifiers as young, childless individuals, DeSena focuses on gentrifying parents

and families. It is this focus on the gentrifying family that leads to her study and conclusions about

how the process of gentrification can impact schools in gentrifying neighborhoods. Through a

series of interviews aimed at determining how gentrifying families engage with neighborhood

schools, she finds that gentrfying parents are more likely than not to seek out education

opportunities outside of the neighborhood. This active rejection of neighborhood schools, leads to

segregation by social class of children with local schools (DeSena, 2009). Overall, DeSena finds:

The gentry are more accustomed to relative privilege and have made a judgment about
what constitutes quality education. Their lower income neighbors accept public schools
and believe that their only other choice is to pay tuition…They are largely unaware of the
options within the pubic school system, such as entrance by lottery or applying to
specialized schools…and the additional strategies used by their gentry neighbors. These
practices result in negative consequences for community cohesion.

These findings are extremely relevant to this study considering that other researchers have

suggested gentrification as a means to improve neighborhood schools through the anticipated

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 16
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

advocacy and involvement of gentrifying parents who are dissatisfied with the quality of

neighborhood schools (Formoso, Weber, & Atkins, 2010).

Other qualitative work on this area of gentrification suggests that by-products of

gentrification, such as improved institutional resources and collective socialization work together

to improve children’s well-being, a larger umbrella of indicators that includes performance in

schools (Formoso, Weber, & Atkins, 2010). This claim rests heavily on the assumption that

gentrifying families and original residents interact formally in institutional spaces like the school,

or informally.

The possibility that these parents would instead find other options for their children, which

DeSena explains by saying it is easier to commute children than leverage their human and social

capital to affect change in schools, hold serious implications for the impact of gentrification on

student achievement (DeSena, 2009). If gentrifying parents choose to educate their children

through avenues other than neighborhood schools, gentrification may give way to an even more

segregated schooling landscape than we already have, where even students who live in the same

neighborhood will have a schooling experience that is segregated by class, income, and/or race.

DeSena also conducted a study in conjunction with George Ansalone that takes another

qualitative look at the impact of gentrification on education. The study expands on her initial

findings in her book, and verifies, by means of interviews, that gentrification contributes to the

accentuation of between-school tracking (DeSena & Ansalone, 2009). Although the study finds

that over time, more than 50 percent of gentry families choose to send their children to schools

and education programs outside of their neighborhood, these findings were based on interactions

with 21 families in one neighborhood in Brooklyn, NY. The sample size and the specificity of the
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 17
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

location pose significant limitations with regards to external validity and the generalizability of

these findings.

Quantitative Research

Of the already slim pool of research on gentrification and education is a slimmer pool for

which the variables of interest and findings are quantified. In Chicago, it was determined that,

overall, gentrification has little effect on neighborhood public schools (Keels, Burdick-Will &

Keene, 2013). If there is any impact, it may be slightly negative as neighborhoods skew to meet

the needs and preferences of higher income residents. For example, as the median income in a

neighborhood increased, students experienced a lower than average increase in test scores.

The researchers of this study establish gentrification as their independent variable and third

grade standardized test scores as their dependent variable. They measured and defined

gentrification using a categorical and linear method. Using Taylor and Puente’s work from 2004

on gentrification in Chicago, they determine that a neighborhood has gentrified if:

…during a given decade…it undergoes at least two of the following: 9 percent increase in
the percent of residents with a college education, 29 percent consumer price index (CPI)-
adjusted increase in average household income, 65 percent CPI-adjusted increase in
average home value, or 11 percent increase in CPI-adjusted median rent (Keels, Burdick-
Will, & Keene, 2013).

The linear method draws on Griffith’s definition of gentrification as the “in-migration of middle-

and upper-income households into existing lower income urban neighborhoods and the upgrading

of the housing stock therein.” By using this linear method, the researchers are able to eliminate the

perceived difference between consumption based theories and production-based theories for the

occurrence of gentrification. Regardless of whether individuals are driving the change or

conditions are created to attract these individuals, the ultimate situation is a change in residents’

demographics and housing stock. The rationale for focusing on third grade is based on the
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 18
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

assumption that children in earlier grades are more impacted by the effects of gentrification than

older children who would have completed several years of school before the neighborhood was

gentrified.

The analysis that is created with these variables is a linear growth model that tests the extent

to which the rates of growth in test scores for each school are associated with level of neighborhood

change.

Researchers of this study found negligible improvements in student achievement: for

reading, a 1 percent increase in the fraction of residents with a bachelor’s degree is associated with

an annual increase in reading scores 0.002 points higher than Chicago Public Schools average—

less than one half of 1 percent of the average growth each year. These insignificant findings

extended across subjects. Further, there were high rates of student mobility in Chicago Public

Schools. The prevalence of student mobility suggests that even if schools were to improve, the

original low-income students would not benefit, as they would no longer reside in neighborhoods

with these improved schools.

Though this study concluded with largely insignificant findings, the research design and

methodology provides a solid framework for conducting a similar study in counties in New York

State.

Significance and Implications of this Study

This study has the potential to make three major contributions to the existing research on

gentrification. First, it will contribute to the developing definition of gentrification. By continuing

to incorporate race as an equally essential indicator and measure of gentrification in the state of

New York, as other researchers have done in other locations, this study may reaffirm the need to
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 19
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

include race, or if it shows race to have an insignificant impact, may suggest the decreasing

importance of race as a factor in the process of gentrification.

Secondly, this study will examine associations of changes in student achievement with

neighborhood change, and add to the growing knowledge about this relationship. There are three

major camps of gentrification research: 1) descriptive studies that dominated most of the initial

research, 2) studies that seek to explain the causes of gentrification (consumption- and production-

based theories), and 3) studies that examine the impact of gentrification on other sphere social

policy. Because the third camp mostly consists of research on housing and displacement, the study

on the impact gentrification has on student achievement will diversify the research on impact and

foster a more multifaceted analysis of the implications of gentrification.

These potential contributions hold significant implications for policies concerning

gentrification. Although a substantial amount of work has been to substantiate gentrification’s

negative impact on the availability of affordable housing, if this study reveals that there are positive

or negative impacts of gentrification on education, policymakers would then have to consider the

tradeoffs (or gains) in both the housing and education spheres.

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 20
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Chapter 3: Data and Methods

Data Sources

In order to address the general question of the impact of gentrification on student

achievement, I will use data from two sources: Student achievement data will be collected from

the New York State Department of Education on the standardized test performance of all third-

graders attending neighborhood schools in the 63 counties of New York. The standardized test

scores include both Math and English Language Arts statewide assessment over a period of 7 years

(2006 through 2012). This specific time frame is due to data constraints, but fortunately also allows

for the analysis of change in student achievement during the period of the most rapid gentrification

in Brooklyn. From 2006 to 2012, several Brooklyn neighborhoods lost significant percentages of

Black and Latino residents as interactive maps on the Census Explorer indicate.1 The intent behind

using third grade test scores is consistent with the research conducted in Chicago that determined

third grade to be most appropriate given that students will not yet have been conditioned to

standardized testing, nor been in the public school system long enough to be biased (Keels et al,

2013).

In terms of quantifying gentrification, I will collect data on demographic and population

trends in New York’s counties using the United States Census Bureau’s American Community

Survey. More specifically, I will focus on the educational attainment levels of residents, their

household income, and the racial composition of each county. Educational attainment, income,

and race will be used. These indicators seem to be the defining characteristics of gentrifying

individuals according to the literature on this topic (Atkinson, 2000; Baxter, 2009; Billingham,

1
Census Explorer: http://census.socialexplorer.com/pop-flash/
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 21
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

2013; DeSena, 2009; Danyluk & Ley, 2006; Formoso, 2010; Glass, 1964; Glick, 2008; Hwang,

2014; Keels et al, 2013; Kozak, 2014; Lees, 2012; & Zukin, 1987).

Ultimately I will be constructing my own data set of New York State that will consist of

standardized test scores by county, and the demographic and population trends in the 63 counties

from 2006 through 2012.

Variables of Interest

The dependent variable of interest, student achievement will be measured by student test

scores on statewide standardized tests. Test scores are given in both raw scores and on a scale of

1 to 4. Scores of 470-680 are below standard (1), scores 661-683 meet basic standard (2), scored

684-706 meet proficiency standard (3), and scores 707-770 exceed proficiency standard (4). This

thesis will make use of the percentage breakdown of the categorical scores for each county

provided by the New York State Department of Education.

The independent variable of interest is broadly referred to as gentrification, but in this thesis

takes the specific form of increased educational attainment levels, increased income levels, and

decreasing levels of Black and Latino residents. Each of these indicators will be interpreted as

continuous variables, since they are all measured as percentages by the US Census Bureau’s

America Community Survey. In addition to these independent variables, I will include the poverty

rate, unemployment rate, and population density of each county as control variables. Each of these

designated control variables are significant confounding variables whose impact should be

clarified as each of them is related to student achievement and the other social identifiers I have

included in my definition of gentrification.

Analysis
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 22
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

My analysis will include bivariate regressions for each aspect of gentrification, and two

separate multivariate regressions that seek to achieve more precise measures of the relationship

between each indicator of gentrification and student achievement.

Bivariate Regression:

Υ𝜅𝜏 = 𝛽0 + 𝛽1 Ν𝜅𝜏 + 𝛽2 (𝑦𝑒𝑎𝑟)𝜏 + 𝜀𝜅𝜏

In this bivariate regression, Υ𝜅𝜏 represents the test score in ELA or Math in county 𝜅 in year 𝜏, Ν

represents one of the three gentrification indicators in county 𝜅 in year 𝜏. This regression will serve

as a preliminary measure on the association between the dependent variable (ELA and Math test

scores) and each independent measure of gentrification from 2006 to 2012. I also included a

dummy year variable to control for time. Since change over time is at the center of my analysis,

this variable will be used in regression analyses.

After conducting these bivariate regressions, I will conduct the following two multivariate

regressions:

Multivariate Regression #1:

Υ𝜅𝜏 = 𝛽0 + 𝛽1 (𝑒𝑑𝑎𝑡𝑡𝑛)𝜅𝜏 + 𝛽2 (𝑖𝑛𝑐𝑜𝑚𝑒)𝜅𝜏 + 𝛽3 (𝑟𝑎𝑐𝑒)𝜅𝜏 + 𝛽4 (𝑦𝑒𝑎𝑟)𝜏 + 𝜀𝜅𝜏

In this regression, Υ𝜅𝜏 represents the test score in ELA or Math in county 𝜅 in year 𝜏, (𝑒𝑑𝑎𝑡𝑡𝑛)𝜅𝜏

represents the educational attainment level of residents in county 𝜅 in year 𝜏 , (𝑖𝑛𝑐𝑜𝑚𝑒)𝜅𝜏

represents the household income in county 𝜅 in year 𝜏 , and (𝑟𝑎𝑐𝑒)𝜅𝜏 represents the racial

demographics of county 𝜅 in year 𝜏. This regression allows for the assessment of the relationship

between each component of gentrification and student achievement while controlling for the other

measures. It also contributes to the growing literature about which aspect of gentrification is most

impactful. The findings from this regression will hold implications for policy makers, as this
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 23
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

knowledge on which factors of gentrification most impact student achievement will influence how

they address gentrification.

Multivariate Regression #2:

Υ𝜅𝜏 = 𝛽0 + 𝛽1 (𝑒𝑑𝑎𝑡𝑡𝑛)𝜅𝜏 + 𝛽2 (𝑖𝑛𝑐𝑜𝑚𝑒)𝜅𝜏 + 𝛽3 (𝑟𝑎𝑐𝑒)𝜅𝜏 + 𝛽4 (𝑝𝑜𝑣𝑒𝑟𝑡𝑦)𝜅𝜏

+ 𝛽5 (𝑢𝑛𝑒𝑚𝑝)𝜅𝜏 + 𝛽6 (𝑝𝑜𝑝)𝜅𝜏 + 𝛽1 (𝑦𝑒𝑎𝑟)𝜏 + 𝜀𝜅𝜏

In this regression, I will add control variables for the poverty rate, unemployment rate, and the

population density of each county 𝜅 in year 𝜏. Each of these additional controls all have the

potential to correlate with the observed test scores as well as the measures of gentrification.

Including them in this regression helps to net out the influence of potentially confounding factors.

These regressions serve to increasingly measure the effect of these different aspects of

gentrification on counties throughout the state and will help to discern which aspect of

gentrification is most relevant regarding student achievement or if the demographic changes

associated with gentrification have any impact at all in counties in New York.

Measurement

The benefit of using data source like the New York State Department of Education is that

the data are not self-reported, which eliminates some sources of bias. Both the Census and the

NYS Department of Education encompass large populations that build a significant sample size,

minimizing noisy data. Using data from the Census Bureau is also beneficial due to the yearly data

collection provided by the American Community Survey, which offers a more precise look at the

changes in student achievement and neighborhood demographics over time.

My data sources and methods also present significant challenges considering the

granularity lost when measuring changes in student achievement and demographics changes at the

county level rather than at the district level within the county. The Census Bureau does collect
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 24
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

demographic information on lower geographic levels than the county, however, that information

is only available in five-year aggregates. Since the key aim of this thesis is to examine changes in

student achievement within the context of changes in neighborhood demographics, I have chosen

to use the county level data, which is provided on a yearly basis.

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 25
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Chapter 4: Results

General Findings

The impact of gentrification on student achievement is most apparent in the income aspect

of gentrification. Of all three measures of gentrification median income consistently had the largest

and most statistically significant association with mean ELA scores and ELA proficiency rates.

However, the association between median income and these measures of student achievement is

positive while other aspects of gentrification (educational attainment and racial demographics)

have a negative association with student achievement. Although the median income variable was

intended to capture the trends in income as it relates to gentrification, the fact that it has this

positive relationship suggests that it may be more representative of the higher scores of new higher

income students rather than of higher scores among students who remain in a place that is

undergoing gentrification. Unfortunately, my data and methods are not designed in a way that

allows me to discern which perspective of median income is being captured. The second most

significant aspect of gentrification in relation to student achievement is race, as the percent of

Black and Latino residents consistently has a statistically significant impact (also in relation to

ELA scores and proficiency).

As control variables are factored in in the progression from bivariate analyses to Model 1

and Model 2, the impact of each aspect of gentrification fluctuates. Relationships between

gentrification measures and student achievement appear more significant when student

achievement is measured by the percent of students who are proficient. For example, in Table 2,

Model 2 shows that a $10,000 increase in median income would result in a 0.193 increase in mean

ELA scores, but in Table 3, that same change is associated with a 0.247 percentage point
 Running head: GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN NEIGHBORHOOD
CHANGE AND STUDENT ACHIEVEMENT

Table 2: Association Between Measures of Gentrification and Reading and Math Test Scores for 39 counties in New York
State Over 7 Years (2006 to 2012)

Mean English Language Arts Test Score Mean Mathematics Test Score°

Bivariate Bivariate Bivariate Model 1 Model 2 Bivariate Bivariate Bivariate Model 1 Model 2
Percent with Bachelor’s Degree or 0.102 0.001 0.0 21 1.175 -3.759 -4.372
Higher (0.063) (0.051) (0.051) (2.777) (3.590) (3.956)

Median Income 0.215*** 0.231*** 0.193*** 3.594 5.118* 3.960

(0.037) (0.034) (0.046) (1.847) (2.372) (3.864)

Percent Black or Latino -0.064* -0.103*** -0.076** 0.443 0.470 -2.352

(0.026) (0.020) (0.024) (1.492) (1.511) (2.364)

Population -0.010 1.255*

(0.007) (0.624)

Unemployment Rate -0.154 7.171

(0.119) (13.443)

Poverty Rate -0.110 -4.835

(0.105) (10.216)

Year -1.0431*** -1.138*** -0.980*** -1.104*** -0.970*** 4.121 3.612 4.293 4.169 5.115
(0.098) (0.096) (0.098) (0.140) (15.669) (15.662) (15.581) (18.264)

R2 0.202 0.423 0.222 0.583 0.596 0.001 0.017 0.000 0.022 0.041

N 39 39 39 39 39 39 39 39 39 39

* p-value < 0.05 **p-value < 0.01 ***p-value < 0.001

Source: New York State Department of Education, US Census Bureau – American Community Survey

° Mean math scores were unavailable for 2006. These findings reflect 6 years of observation (2007-2012)
Running head: GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE
CORRELATION BETWEEN NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

increase in ELA proficiency rates. Across the majority regression models, the passage of time

maintains a statistically significant negative impact on student achievement. Though the impact

on scores and proficiency levels suggests that the aspects of gentrification most relevant to student

achievement are the median income and racial composition of the area, the real impact of both

aspects is minor.

Effect by Subject

Surprisingly, the effect of gentrification appeared very different depending on the subject.

While educational attainment seemed to have a slightly positive association with mean ELA

scores, it had a negative association with mean Math scores. The statistical significance of the

relationship between mean ELA scores and the percentage of Black and Latino residents was much

higher than for mean Math scores. However, as with mean test scores, the percentage of Black and

Latino residents had a more statistically significant association with ELA proficiency than it did

with Math proficiency. Overall, ELA scores and proficiency rates have more statistically

significant results than math scores and proficiency rates, and are therefore referred to more

frequently when analyzing the impact of gentrification on student achievement.

Effect of Gentrification on Student Achievement as Measured By Mean Test Scores

Table 2 presents results from the bivariate and multivariate models discussed in Chapter 3,

where outcomes are mean test scores in English Language Arts and Mathematics. Through a series

of bivariate regressions, I examined the relationship between the three separate measures of

gentrification and mean test scores, while controlling for a linear time trend.

These initial bivariate regressions indicate that median income has the greatest impact on

test scores (with a coefficient of 0.215) in relation to the other gentrification measures. They also

show that a one percent increase in the fraction of residents who hold a Bachelor’s degree or
Running head: GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN NEIGHBORHOOD CHANGE AND
STUDENT ACHIEVEMENT

Table 3: Association Between Measures of Gentrification and Reading and Math Proficiency for 39 Counties in New York State Over 7
Years (2006-2012)

English Language Arts Proficiency Mathematics Proficiency

Bivariate Bivariate Bivariate Model 1 Model 2 Bivariate Bivariate Bivariate Model 1 Model 2
Percent with Bachelor’s Degree or 0.136 -0.055 -0.058 0.167* -0.074 -0.114
Higher (0.093) (0.066) (0.068) (0.069) (0.084) (0.092)

Median Income 0.325*** 0.350*** 0.247*** 0.242*** 0.269*** 0.198*

(0.052) (0.044) (0.066) (0.044) (0.056) (0.091)

Percent Black or Latino -0.146** -0.184*** -0.151*** -0.075 -0.088* -0.085

(0.041) (0.028) (0.039) (0.037) (0.036) (0.058)

Population 0.004 2.41x10-6

(0.010)
Unemployment Rate -0.438* -0.547
(0.217) (0.336)

Poverty Rate -0.273 -0.240

(0.168) (0.242)

Year -3.036*** -3.183*** -2.934*** -3.108*** -2.707*** -5.446*** -7.422*** -5.369*** -5.497*** -5.025***
(0.200) (0.197) (0.193) (0.250) (0.335) (0.336) (0.316) (0.388)

R2 0.367 0.506 0.418 0.616 0.629 0.616 0.651 0.610 0.660 0.669

N 39 39 39 39 39 39 39 39 39 39

* p-value < 0.05 **p-value < 0.01 ***p-value < 0.001

Source: New York State Department of Education, US Census Bureau – American Community Survey
Running head: GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE
CORRELATION BETWEEN NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

higher to increase mean ELA scores by 0.102 points, and a one percent increase in the fraction of

Black and Latino residents to decrease mean ELA scores by 0.064 points. Of these three bivariate

relationships, only median income and the fraction of Black and Latino residents are statistically

significant. Interestingly, results also illustrate a downward trend in ELA test scores over time.

This relationship is much larger than any of the gentrification measures’ and it maintains high

statistical significance in each bivariate model.

Also in Table 2 are the findings from the first multivariate regression (labeled Model 1).

These findings factor in each aspect gentrification while controlling for secular trends in test scores

over time. The impact (or lack thereof) of median income increased slightly, while the impact of

educational attainment levels (i.e. the percentage of those with a Bachelor’s degree or more)

decreased significantly. Lastly, the impact of the fraction of Black and Latino residents increased

in both magnitude and statistical significance.

Finally, in the second multivariate regression (labeled Model 2), I included several control

variables that are highly correlated with the dependent and independent variables. These controls

include the total population, the poverty rate, and the unemployment rate. None of these controls

proved to be statistically significant in themselves, however, after factoring them in, the magnitude

of the percentage Black and Latino residents as well as median income dropped slightly (in

comparison to the first multivariate regression). The significance of the percentage of Black and

Latino residents also decreased. Given that poverty and unemployment are often racialized, this

decrease in the impact of fraction Black and Latino is not unexpected. The racialization of poverty

and unemployment are part of the expected correlation between these additional controls and

aspects of gentrification and student achievement. In addition to racialized poverty the impact of

poverty on student achievement has also been well studied and documented.
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 30
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Effect of Gentrification on Student Achievement as Measured By Proficiency

In addition to examining the effect of gentrification on mean test scores, I also examined

its effect on proficiency in both English Language Arts and Mathematics. These statistical analyses

revealed similar results as the mean test score findings. While the bivariate analyses indicated that

the percentage of those with a Bachelor’s degree or higher might be the most pressing factor, as

controls were increasingly introduced, the impact of the percentage of residents with a Bachelor’s

degree or higher decreased in magnitude. Median income had the largest and most statistically

significant impact, with the percentage Black or Latino having the next largest influence (Table

3).

It is important to note that when looking at gentrification’s impact on student achievement

through the scope of mean test scores, regardless of the statistical significance, the real impact was

miniscule. For example, for mean ELA scores, with a mean of approximately 667 and a standard

deviation of about 5 (Table 4), increasing or decreasing mean test scores by a fraction of a point is

hardly practically significant. The impact on proficiency was similarly small. For example, a one

percent increase in Black or Latino residents is associated with a 0.151 decrease in ELA

proficiency (Table 3: Model 2). With a mean of ~65.2 and standard deviation of ~10.4 (Table 4),

ELA proficiency would only change significantly if there were a drastic change in the racial

demographics.
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 31
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Table 4: Descriptive Statistics of New York State Over 7 Years (2006-2012)

Descriptive Statistics

Variable Mean Standard Deviation

Mean ELA Score 667.907 5.954
Mean Math Score° 714.538 399.816
ELA Proficiency 65.160 10.441
Math Proficiency 75.118 15.377
Percent with Bachelor’s Degree or Higher 28.655 9.650
Median Income (increments of $1,000s) 54.841 14.357
Percent Black or Latino 17.503 17.749
Total Population (increments of 10,000) 48.078 62.080
Poverty Rate 13.447 4.784
Unemployment Rate 7.851 2.606

N = 39 counties
° Mean math scores were unavailable for 2006.

Source: US Census – American Community Survey, New York State Department of Education
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 32
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Chapter 5: Conclusions

The Relationship Between Gentrification and Student Achievement

The main research question for this thesis was to explore the relationship between

gentrification and student achievement. In doing so, this thesis sought to contribute to the growing

definition of gentrification as well as to the growing literature on the association between

gentrification and student achievement. Given prior research on gentrification itself, I settled on a

three-pronged definition of gentrification that consisted of racial demographics, median income,

and educational attainment levels. Measuring student achievement by standardized test scores was

also motivated by prior literature, which indicated that numerical representations of student

achievement were least biased and most generalizable.

A series of multivariate regressions indicated that gentrification has little to no real impact

on student achievement. Although income and racial factors appear to have the most impact

(Model 2 in Table 2 and Table 3), even these associations are not practically significant given the

mean and standard deviation of student test scores (mean: ~668; st. dev.: ~6) or proficiency levels

(mean: ~65l st. dev.: ~10) (increases in median income and percentages of Black and Latino

residents are associated with changes of less than half a point in test scores and less than half a

percentage point in proficiency). Although these results did not completely confirm my working

hypothesis that gentrification would have a negative association with student achievement, these

regressions did provide some direction for moving forward with gentrification studies, as to which

aspect of gentrification might be more pertinent to future research on the subject. The fact that the

racial makeup of a county (conditional on equal income and educational attainment levels) was a

prominent aspect of gentrification in its relationship with student achievement indicates that race
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 33
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

is not to be discredited. The study of gentrification, though rooted in class-based and racial

discourse, has recently tended to place more emphasis on class, but these findings show that race

remains relevant.

Limitations

Ideally, future work should be modeled after the Keels et al (2013) study on neighborhood

change and student achievement in Chicago. These authors focus on the relationship between

gentrification and student performance on a much smaller geographic level (i.e., census tracts)

over a longer period of time. While gentrification eventually manifests across an entire county, it

starts on much smaller geographic scales (blocks or neighborhoods within a county). These authors

were also able to isolate the students whose families remain in neighborhoods where gentrification

is occurring. This part of their design addresses the possibility that observed changes in student

achievement might be the result of newly arrived students of gentrifying families rather than the

students who were there before the change. Additionally, Keels et al (2013) were able to observe

the relationship between neighborhood change and student achievement over the span of decades.

Due to the fact that my main source, the Census Bureau, did not maintain demographic data on

similarly granular geographic levels and did not maintain annual data consistently (the American

Community Survey, which provides yearly data just started in 2005), this thesis was limited to

county-level observations for seven years. Furthermore, of the 63 counties in the state, only 39

counties had yearly demographic data.

Another significant limitation revealed itself in the analysis stage. Although median

income had the largest and most significant impact on student achievement, as stated in Chapter

4, it is not clear whether this impact is of the students’ own household income or that of
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 34
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

surrounding gentrifiers. This uncertainty exists for all of variables I used to measure gentrification

as there is no way to make this distinction in the design of my data or models of statistical analysis.

While these data constraints presented limitations in certain ways, it also led to the creation

of a statewide data set that can be used for future study involving these variables. Also the decision

to use county-level data for seven years allowed for a larger, less biased data set than would have

been possible if I had attempted to use neighborhoods and census tracts as my level of analysis. In

order to address these limitations, this thesis could be continued for at least another three years

using data from the American Community Survey to capture the relationship between

gentrification and student achievement over ten years.

Policy Implications

Though the intent of this thesis was to determine the effects and policy implications of

gentrification as it relates to student achievement, results indicate that a more pressing issue is the

decline of student achievement over time in New York. In each regression model, regardless of

the number of variables incorporated, results showed that over time student achievement (in mean

test scores and proficiency) decreased gradually. These findings hold serious implications for

reforming and improving education in New York. The association between education and other

socioeconomic indicators such as employment status or income have been well established by

social scientists. Therefore, although declining student achievement was an unexpected finding, it

nevertheless demands attention from policymakers in the state to implement practices and changes

that will reverse these trends.

GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 35
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

References

Amrein-Beardsley, A. (2012). Methodological Concerns about the Education Value-Added

Assessment System. Educational Researcher (Vol. 37). Retrieved from
http://www.jstor.org/stable/30137966

Atkinson, R. (2000). The hidden costs of gentrification: Displacement in central London.

Journal of Housing and the Built Environment, (Vol. 15). Retrieved from
http://www.jstor.org/stable/41107148

Bates, L. (2013). Gentrifcation and Displacement Study: implementing an equitable inclusive

development strategy in the context of gentrification (Doctoral dissertation). Retrieved
from https://www.portlandoregon.gov/bps/article/454027

Baxter, H. (2009). Toward a theory of gentrification (Doctoral dissertation). Retrieved from

OhioLINK (447)

Billingham, C. M., (2013). The state, the school, and the family in the gentrification of the
American city (Doctoral dissertation). Retrieved from IRis (20)

Bonk W. et al (2012). Growth Models: Issues and Advice from the States. Retrieved from
http://nces.ed.gov/programs/slds/pdf/guide_growth-model.pdf

DeSena, J. N. (2009). Gentrification and inequality in Brooklyn : the new kids on the block.
Lanham, MD: Lexington Books.

DeSena, J. N., & Ansalone, G. (2009). Gentrification, Schooling and Social Inequality.
Educational Research Quarterly, 33(1), 60-74.

Danyluk M. & Ley D. (2006). Modalities of the New Middle Class: Ideology and Behaviour
in the Journey to Work from Gentrified Neighborhoods in Canada. Urban Studies
Journal Foundation (Vol. 44). Retrieved from
http://usj.sagepub.com/content/44/11/2195.full.pdf+html

Formoso, D., Weber, R. N., & Atkins, M. S. (2010). Gentrification and Urban Children’s
GENTRIFICATION AND EDUCATION AT A CROSSROADS: THE CORRELATION BETWEEN 36
NEIGHBORHOOD CHANGE AND STUDENT ACHIEVEMENT

Well-Being: Tipping the Scales from Problems to Promise. American Journal of

Community Psychology (Vol. 10). Retrieved from
http://link.springer.com/article/10.1007/s10464-010-9348-3/fulltext.html

Glass, R. (1964). London: Aspects of Change. London: MacGibbon & Kee.

Glick, J. (2008). Gentrification and the Racialized Geography of Home Equity. Urban Affairs
Review (Vol. 44). Retrieved from http://uar.sagepub.com/content/44/2/280.full.pdf+html

Hwang, J. (2014) How “gentrification” in American cities maintains racial inequality and
segregation: Research brief. Retrieved from
http://www.scholarsstrategynetwork.org/sites/default/files/ssn_key_findings_hwang_on_
how_widespread_gentrification_maintains_racial_inequality_and_segregation_2.pdf

Keels, M., Burdick-Will, J. and Keene, S. (2013). The Effects of Gentrification on

Neighborhood Public Schools. City & Community, 12: 238–259.
doi: 10.1111/cico.12027

Kozak, S. (2014). From Section 8 to Starbucks: The Effects of Gentrification on Affordable

Housing in Pittsburgh, Pennsylvania (Doctoral dissertation). Retrieved from
kuscholarworks.ku.edu/.../1/Kozak_ku_0099D_13316_DATA_1.pdf

Lees, L. (2012). Gentrification. In Andrew T. Carswell (Ed.), The encyclopedia of housing.

(2nd ed., pp. 241-244). Thousand Oaks, CA: SAGE Publications, Inc. doi:
http://dx.doi.org/10.4135/9781452218380.n82

Scott J. & Miller J. (2012). Understanding the Basics of Measuring Student Achievement.
Retrieved from http://cecr.ed.gov/pdfs/Understanding_Basics.pdf

Zukin, S. (1987). Gentrification: Culture and Capital in the Urban Core. Annual Review of
Sociology, (Vol. 13). Retrieved from http://www.jstor.org/stable/2083243