Abstract

In this study we explore how gentrification has impacted student outcomes vis-a-vis public schooling in Nashville, TN. Academic literature on gentrification’s impact on schools is varied and there is no consensus on what gentrification could mean for a given neighborhood school, not in the least those of North and East Nashville. We use American Community Survey 5-year estimates from 2013-2018 to identify gentrifying tracts. We then find high schools in these tracts and compare two key metrics, associated with student success, to measure the impact of gentrification on student outcomes. We find a complicated relationship; while higher rates of gentrification are broadly associated with higher graduation rates, we find a quadratic relationship with ACT proficiency. Measuring this effect is crucial, as gentrification disproportionately displaces minority families and could potentially harm Nashville’s predominantly Black schools. Thus, this study has important ramifications for exploring racial and socioeconomic equity within the city, and will be essential, as school advocates and local policymakers determine an appropriate response to gentrification-induced displacement. 

Key words: Gentrification, Displacement, Socioeconomic Inequality, Urban Redevelopment, Racial Stratification 

‍

Figure 7  Source: Metropolitan Government of Nashville & Davidson County

Introduction 

Nashville has been one of the nation’s fastest growing cities throughout the last decade, being hailed as the “It” city in 2013 and becoming the fifth fastest-growing metropolitan area in the country. Even now, as growth slows, the city is still ranked as the 22nd fastest growing metropolitan area in the country (Korolevich 2023). As part of this growth, from 2013-2022, median household income increased from $47,150 to $72,473, while median gross rent almost doubled, from $856 to $1,463. While some Nashville residents are experiencing upward social mobility, it is theorized that lower income families who previously lived in Nashville are being displaced, as they can no longer afford to live in the city, and their children are being sent to other schools outside of the school district (Cantrell 2023). This introduces the potential for decreases in state funding for local public schools in neighborhoods most impacted by gentrification, due to the decreases in enrollment. This could directly lessen the resources available to low SES students, in schools that might already lack adequate resources. Already, a Nashville Chamber of Commerce report found that the average ACT composite score in MNPS fell from 18.5 in 2019, to 17.4 in 2022.  

‍

My personal motivation for undertaking this study comes from growing up in Antioch, a neighborhood in South Nashville. From early in my childhood, I was sent to schools far outside of my own neighborhood, in hopes that other, better-funded public schools would equip me with the tools to attain what my mother hoped for me. During my formative years, gentrification accelerated at a rapid pace, transforming Nashville from a small city into one where cranes litter every corner. However, in the latter half of the 2010s, the onus of this growth shifted from North and East Nashville, to South Nashville. Being forced to confront these changing dynamics, such as the displacement of Black community members, in our daily lives, led to many questions; chief among them, what will this mean for our public schools?  Would schools be depopulated, as students left Davidson County? Would teachers be able to afford to continue living in the city? 

‍

Academic literature on the subject is varied, with one study finding that the reduction in crime and violence in a neighborhood, as well as the removal of other “physiological” barriers to promote educational success among lower SES students (Green et al. 2022). In fact, traditional research posits that gentrification is a positive force in local schools, arguing that the new high SES residents bring with them increased financial resources and parental attention. On the other hand, some studies claim that gentrifying residents are less apt to send their children to local public schools, instead opting for local charter schools, and decreasing the enrollment and resources at schools in gentrified areas (Keels et al. 2013). Literature has also traditionally been centered on research looking at gentrification in larger cities, such as Chicago and Philadelphia, which have populations of 2.6 million and 1.5 million, respectively. Nashville, being both a Southern city and having a population of about only 683,000 people, may see vastly different dynamics play out. Thus, the effects of Nashville’s rapid demographic change on schools attended by low SES students, who live in gentrification hot spots, are virtually unknown, and too complex to predict. Yet, this information is crucial, as gentrification disproportionately displaces Black families and could potentially harm Nashville’s predominantly Black schools. Thus, this study has important ramifications for exploring racial and socioeconomic equity within the city and will be essential as school advocates and local policymakers determine an appropriate response to gentrification-induced educational changes. For this reason, we have decided to track the preliminary effects of gentrification on school performance, as measured by student outcomes. We have used American Community Survey 5-year estimates to track a number of variables commonly associated with gentrification, across every tract in the Nashville Metropolitan area, to first establish which tracts have the most rapid SES increases – likely meaning lower SES families are being displaced at higher rates. We then found high schools in the area and compared outcomes in graduation rates, ACT averages, and school finance data. 

‍

In this study we specifically test how higher levels of gentrification correlate to “student success” metrics, as defined by local schools’ graduation rates and the percentage of students scoring a 21 or above on the ACT. To answer this question, we compiled a composite gentrification score with data from ACS 5-year data and found graduation and ACT data on public high schools in Davidson County on the State of Tennessee website. We then conducted a Principal Components Factor Analysis to create a gentrification score for each census tract, then administered a regression analysis to find the correlation between the gentrification level and our “student success” factors - graduation rates and ACT proficiency. We found that there is a weak but slightly positive correlation between gentrification level and graduation rates, yet a quadratic relationship between the gentrification and ACT proficiency. The varied nature of the correlation not only echoes the complex, yet interlocked effects of gentrification on previously low SES neighborhoods, but also serves as an intriguing result and the inspiration for future academic work on the intersection of gentrification and educational institutions. 

‍

Literature Review 

The process which we call “gentrification" is multifaceted, with a series of complicated contradictory impacts on local public schools. Dr. Francis A. Pearman of Stanford University simply describes it as “the in-migration of relatively affluent households into disinvested central city neighborhoods,” while professor of geography and anthropology - Dr. Neil Smith - defines it as “the process by which central urban neighborhoods that have undergone disinvestments and economic decline experience a reversal, reinvestment, and the in-migration of a relatively well-off middle- and upper middle-class population” (Pearman 2019; Smith 1998).These differing yet similar definitions of gentrification encapsulate its essence, yet only focus on the new, middle-class residents.  

‍

An important facet of this process is the idea of displacement. A side effect of the migration of new residents into previously low SES neighborhoods is the increase in rent costs and housing values (Hwang 2016; Pattillo 2008; Smith 1996). The rising cost of living often can force low SES inhabitants of a gentrifying neighborhood to leave. Thus, more recent literature also includes this “displacement” in the definition of gentrification. Geographer, urban planner, and housing policy analyst Samuel Stein’s 2019 definition of term denotes gentrification as the “process by which capital is reinvested in urban neighborhoods, and poorer residents and their cultural products are displaced and replaced by richer people and their preferred aesthetics and amenities” (Stein 2019). This demographic shift happens most overtly through eviction, neglect of rental units inhabited by lower SES residents, residents being offered money to sell their homes to developers, or related increases in financial strain (Marcuse 1986; Newman & Wyly 2006; Brown-Saracino 2017). Yet comparing the previous and current socioeconomic compositions of gentrified neighborhoods reveals important economic, racial, and ethnic implications in the demographic outflows. In many studies, gentrified neighborhoods see an influx of middle and upper-class residents and a simultaneous increase in the number of White families and decrease in the number of families of color. (Wellman 1979; Zukin 2010; Hyra 2014). This has led many researchers to claim that in accordance with economic forces, socioeconomic implications, governmental action (or the lack thereof), the process is deeply linked to racial stratification and historical racial hierarchies.  (Boyd 2005; Irby 2015; Lees et al, 2013; Moskowitz 2017; Pattillo 2010; Smith and Stovall, 2008; Stein 2019, Hwang and Ding 2020). Thus, in a separate study, Stanford University’s Dr. Jackelyn Hwang finds US census tracts that are considered gentrifying have larger increases in the share of college-educated residents, renters, White residents, median home values, and rents, while experiencing more rapid decreases in Black populations and poverty rates (Hwang 2020).  

‍

With these effects, some have claimed that gentrification is a net positive for local schools. In a 2006 interview-based study, Dr. Lance Freeman finds that gentrification led to the improvement of local services, and this, in tandem with related reductions in violent crime could foster an environment more conducive to student success (Freeman 2006). In fact, Dr. Pearman’s 2019 study theorizes as much, claiming that “neighborhood gentrification impacts the social ecology of students’ lives, which in turn can affect academic outcomes” (Pearman 2019). Others claim that though gentrifying families bring benefits to gentrifying schools, such as financial resources and advocacy for higher level classes, these benefits may be “siphoned off”, only going to those students that are middle-class and White, predicated on the exclusion of lower income students (Pattillo et al. 2014; Posey-Maddox et al. 2014). In a 2009 National Longitudinal Study of Adolescent Health study, researchers found that as inequality increased in school, educational gains for low SES students in math and science actually decreased (Crosnoe 2009). Other studies have found that increased inequality in schools led to new parents pushing for educational agendas that would only benefit middle-class students, less emotional support for lower SES women, and a greater share of lower SES students dropping out of school (Cucchiara 2013; Curley 2009; Kearney & Levine 2014). In fact, a greater number of families competing for enrollment in the same school could mean that low SES families, who have previously lived in the neighborhood, are forced to send their children elsewhere (Posey-Maddox et al. 2014). Even if higher SES and lower SES students are able to attend the same schools, the leverage newer families will have, due to sizable financial contributions and fundraising might provoke racial and class tensions (Cucchiara & Horvat, 2009; Posey-Maddox, 2014).  

‍

Therefore, there is evidence that some aspects of the gentrification process may, at best, be inconsequential and, at worst, potentially harmful to the academic achievement of children whose families have historically called these neighborhoods home. In some studies, scholars have found that gentrification has no impact on urban schools or student success, such as Dr. Micere Keels's 2013 study, where they find “no aggregate academic benefit on student outcomes” (Keels et al. 2013) This can be explained by gentrifier families not always sending their children to local public schools, instead choosing to enroll their children in charter schools, private schools, or public institutions in more well-off neighborhoods (Keels et al. 2013; Bridge 2006). In fact, academic literature has found out that gentrifier families are more likely to opt-out of sending their children to neighborhood schools (Candipan 2020; Pearman & Swain 2017). In some gentrifying neighborhoods, this has caused the student populations at neighborhood schools to decline, due to the two-pronged effect of low SES families being displaced, and gentrifying families opting out of sending their children to neighborhood schools. Since enrollment is a large determinant for school funding this would imply that gentrification, paradoxically, decreases the resources available to urban schools.  

‍

In sum, the effect of gentrification on neighborhood public schools varies so much that it should be studied on a case-by-case basis. The process may prove beneficial for remaining low SES students, who see decreases in violence and improvements in local services. However, benefits from more overt sources (fundraising, a more active PTA, new class offerings, etc.) will likely only be distributed amongst newcomer families, at the expense of low SES students. Alternatively, one may see no change in these schools or even a deterioration in student outcomes, as gentrifying families opt out of sending their children to neighborhood schools, decreasing available resources. 

‍

Methodology 

Census data relating to gentrification was first obtained from Social Explore. We identified five variables to track from sets of American Community Survey datasets - Median Household Income, Median Gross Rent, Median House Value (for all owner-occupied housing units), Percentage of Residents 25 years or over with at least a bachelor's degree, and Percentage of Renter Occupied Housing Units. After identifying the variables of interest, we then recorded their values from ACS 5-year datasets for the years of 2013-2018. We then imported the data into Stata and conducted a Principal Components Factor Analysis to create a summary measure of gentrification potential, and afterwards calculated a gentrification score based on the difference in each tract’s gentrification potential between 2013-2018. These gentrification change scores were then imported into GIS. 

‍

Along with the gentrification change scores, we imported shapefiles showing census tracts and school clusters/districts in Davidson County, into GIS. Gentrification change scores were assigned to their corresponding census tract, using the Spatial Join feature. This data was then converted into Raster data, using the Polygon to Raster conversion command. Dimensions included a 0.000083333 Decimal Degree cell size. The Zonal Statistics as Table Tool, was then utilized, designating school clusters as the zonal polygon, to record the average Gentrification change scores in each tract.  

‍

We combined the Zonal Statistic data, with our school data, by using the Spatial Join feature. With this done, we are able to also import data on student success outcomes, such as the Graduation Rate and Percentage over Students Scoring a 21 or Higher on the ACT. We examined this new data, through the use of the Correlation feature in Stata, to determine if there was indeed a correlation between our Gentrification scores and Student Success Outcomes. We asked two specific questions: Are our Gentrification Change scores correlated with graduation rates in 2018? And Are our Gentrification Change scores correlated with the percentage over Students Scoring a 21 or Higher on the ACT in 2018? We then created two scatter plots and used the Regression command in Stata to estimate linear regression models predicting each of our outcomes. 

‍

Results 

Nashville is split into several neighborhoods, largely corresponding to council district boundaries. In the center of Davidson County is Downtown Nashville, surrounded by North Nashville, East Nashville, South Nashville, and Midtown-Green Hills. To the North of Nashville are the neighborhoods Bordeaux-Whites Creek, Joelton, and Parkwood-Union Hill. To the East are Madison and Donelson-Hermitage-Old Hickory. To the South of Nashville are Antioch and Brentwood-Antioch (here labeled as “Southeast”). Lastly, the neighborhoods to the West of Nashville include West Nashville, Midtown-Green Hills (as it extends from Downtown to the county border), and Bellevue.  

‍

It is also important that when we examine student success outcomes, we distinguish the sum data from the zoned school data. “Optional” schools, in which students have to meet certain academic, professional, musical, or other criteria to be eligible for application/admission to these institutions, tend to skew overall student trajectory upward in each school cluster. For example, when removing non-zoned schools from the dataset, the graduation rate in the Pearl-Cohn and Glencliff clusters fall from 92.05% and 85.03% to 76.95% and 73.8%, respectively. Thus, we decide to record both the mean values for student success metrics, and re-record the same metrics, removing the optional schools.  

‍

According to our data, we see North, East, and neighborhoods to the West of Nashville facing the most gentrification. The neighborhoods of North and East Nashville, represented by the Pearl-Cohn and Stratford clusters lead the way, with Green Hills and the Brentwood-Antioch area (represented by the Hillsboro and Overton clusters) closely following. We also find that over this period, my home neighborhood of Antioch (a combination of the Cane Ridge and Antioch clusters) ranked lowest. The Stratford, Pearl-Cohn, and Hillsboro clusters, respectively saw the highest levels of gentrification. Stratford saw a gentrification level over ½ ( 0.616049) a standard deviation above the mean gentrification score, while Pearl-Cohn and Hillsboro saw levels about ½ (0.484042 and 0.46286, respectively), a standard deviation above the mean gentrification score. Looking at our student success metrics – in specific, the mean high school graduation rate, we see that school clusters surrounding Downtown and to the West of Nashville generally had better graduation rates than those on the Northern, Eastern and Southeastern edges of the city. The Pearl-Cohn (North Nashville), Hillsboro (Green Hills), and Lawson (Bellevue) clusters had the highest rates (89.02-92.68%), while Whites Creek (Bordeaux-Whites Creek), McGavock (Donelson-Hermitage-Old Hickory), Stratford (East Nashville), Hunters Lane (Madison), and Antioch saw the lowest rates (75.53-79.63%). Yet, once we removed choice schools from our dataset, we find that schools surrounding Downtown, to the North, and to the East of Nashville tended to have the lowest graduation rates, while those in the West and South had the highest. In this set, the Pearl-Cohn (North Nashville), Glencliff (South Nashville), and Whites Creek (Bordeaux-Whites Creek) saw the lowest rates (75.3-76.95%), while the Hillsboro (Green Hills), Lawson (Bellevue), and Overton (Brentwood-Antioch) clusters led the group. 

‍

Looking at ACT proficiency rates, we see that the clusters in the East Nashville Neighborhood and those to the North struggle the most. Hunters Lane (Madison), Stratford (East Nashville), and Maplewood (East Nashville) see rates between 12.53% and 13.81% of students scoring a 21 or above on the ACT. Whites Creek (Bordeaux-Whites Creek) stands out as an outlier, with only 6.86% of students scoring a 21 or above. On the opposite side, students at schools to the East and West of Nashville tended to score better on the test. Clusters to the West of Nashville lead the way, the mean ACT proficiency percentage in the Hillsboro (Green Hills) and Lawson (Bellevue) clusters being 41.18% and 38.28%, respectively. In this respect Pearl-Cohn (North Nashville) also stands out as an outlier, with a mean percentage of 61.96% of students. When removing choice schools from this dataset, we see similar trends, albeit different in a few key ways. In this set, students in North Nashville, in addition to those in East Nashville and  to the North of Nashville, struggle the most on the ACT. The Pearl-Cohn (North Nashville) cluster ranks lowest, with 6.53% of students scoring a 21 or higher. It is accompanied by Whites Creek and Maplewood (6.86% and 7.7%). On the opposite side, Hillsboro and Lawson still lead the ACT proficiency rates (41.18% and 31.46%), while the McGavock cluster (Donelson-Hermitage) has the highest percentage. 

‍

When comparing gentrification rates to student success metrics, graduation rates and ACT Proficiency, we find interesting implications. Overall, we find that the level of gentrification in a school cluster has a weak, slightly positive correlation (0.255) with graduation rates.  Furthermore, with an R-squared value of 0.197, we find that our regression model only predicts 19.7% of variance in the observed data. Looking at ACT success, we find a quadratic, nonlinear relationship between level of gentrification and ACT proficiency. Thus, the variable also has a very weak correlation (0.063) to the level of gentrification. With an R-squared value of 0.065, our model only predicts 6.5% of variance in the observed data. 

‍

Discussion 

This study was undertaken to discover what impacts gentrification in Nashville, Tennessee (Davidson County) had on student success metrics in public high schools, in the 2013-2018 period. Literature studying these implications is notably varied. Overall, we find gentrification to have a slightly positive correlation with graduation rates. On the other hand, we find it to have a quadratic relationship with the percentage of students who score above 21 or higher on the ACT. This means that schools which experience the lowest levels of gentrification and those that experience the highest, perform worse in ACT proficiency, than schools in moderately gentrified areas. In sum, we find that schools in areas that see higher levels of gentrification have slightly higher graduation rates than their peers but perform worse in ACT proficiency than schools in areas of only moderate gentrification.  

‍

‍

Though a lack of stronger conclusions, these results are salient because they add more depth to the complex relationship gentrification has with school related metrics. Why would high levels of gentrification in an area be correlated with increased graduation rate, yet also decreased ACT scores? Perhaps, due to gentrification, students who might otherwise not graduate are more likely to be displaced, bringing up the mean graduation rate. In tandem, the declining student population at schools which are in highly gentrified areas could mean a decrease in school funding, resulting in fewer resources to prepare students for the ACT. Perhaps, in this way, schools in moderately gentrified areas are able to see higher rates of graduation yet also higher ACT scores, as students from disadvantaged backgrounds are displaced, yet not at a rate that would severely affect school funding. Examining the complexity of the relationship between these variables could serve as the grounds for future research on the intersection between gentrification and the education system. The zone-optional school dichotomy is important to mention again in this section. When taking optional schools out of the dataset, we see even less of a correlation between gentrification level and graduation rates, yet a similarly nonlinear (albeit less pronounced) relationship between gentrification level and ACT proficiency. The potential unequal distribution of effects on education institutions, due to gentrifications, and how they may impact optional schools differently than zoned schools should also be the topic of further research on gentrification in Nashville. 

‍

Limitations 

While this study contributes notable insights to the sociological study of gentrification and its effects on local educational institutions, there were a number of factors that hampered the dearth of analysis that could be done in this study. The first of which was the limited size of the data analyzed. By solely looking at the seventeen public high schools in Metro Nashville Public schools, it made complementary analyses - such as separating data for zoned and non-zoned schools - much more difficult to justify, since removing so much of the data would make it harder for a clear correlation to be seen. In a similar way, there were over thirty new US census tracts that were created in the 2020 American Community Survey. This means that during our period of focus (2013-2018), our analysis was constrained by a lack of sizable amounts of tract data, which could have altered our findings if accounted for. Another issue was the suppression of ACT data by the state of Tennessee. For certain schools, where metrics had either ranked at the top or the bottom, data was hidden from the general public, which meant that their values could not be recorded for this project. By only recording data from 2013 onward, there were too few values to attempt to extrapolate the suppressed ones. Ultimately, these factors severely limited the data range we could work with decreasing the original proposed range from 2013-2022 to 2013-2018. This, resultantly, decreased the depth by which we could complete our analysis, working with a more limited data set, that might not have accounted for gentrification over an extended period of time. 

‍

Another key finding is the need for a more comprehensive analysis to be undertaken. Regression tables for this study show that our model only predicts 19.7% of variance in graduation rate data, while it also only predicts 6.5% of variance in ACT proficiency. Due to the highly complex nature in which gentrification affects school systems, and its close relationships with related metrics, it is imperative that this study is replicated with more comprehensive models, to better understand how salient gentrification is in student success. 

‍

Lastly, it is also important that future studies on gentrification and student success include data disaggregated by race and data from local charter schools. Due to the brief nature of this project, data of this nature could not have been secured in time to analyze it, but it is important to do so. A number of studies have found that gentrification can have negative (or positive effects) for low SES minority students, and, more importantly, that any benefits from the process will disproportionately go to newer students. Regarding charter schools, a sizable number of Davidson County students do attend these institutions and may be affected by gentrification in salient ways. Additionally, it was theorized that gentrifying families might choose to send their kids to charter schools, providing an additional reason for future research on the subject to take a more holistic approach to determining schools of interest. Thus, to fully explore the topic and all its implications, this data should be added for future studies. 

‍

Conclusion 

In this study we have utilized American Community Survey and State of Tennessee data to uncover the effect gentrification has had on student success metrics. Our findings suggest that school clusters with very high and very low gentrification scores do worse on the ACT, while students from schools in the middle generally score better, along with seeing a slight increase in graduation rates in areas of higher gentrification. However, we also uncover significant limitations, such as inequalities between zone and optional schools, small sample sizes, lack of inclusion of charter schools, data suppression, and the need to have data disaggregated by race and socioeconomic status. This study can serve as the base for future academic work exploring the impacts of gentrification on the school system and recent socioeconomic developments in Nashville, Tennessee. 

‍

‍