MacKenzie Gaddy, Porter Jones, Craig Dittmann
The focus of our project is on gentrification's impact on K-12 students in Oakland, California. We hope to take polygon data stating the gentrification risk for a given area from the Urban Displacement Project and analyze various educational markers for that space (e.g. looking at student records that indicate whether the student is on track or not). Student groups of the same areas can be matched based on demographic data. If data is available, we will ideally be able to do this analysis at a few different points in time to analyze the changes over time in neighborhood gentrification impact on school achievement as well.
Gentrification is defined by the Urban Displacement Project as:
"A process of neighborhood change that includes economic change in a historically disinvested neighborhood —by means of real estate investment and new higher-income residents moving in - as well as demographic change - not only in terms of income level, but also in terms of changes in the education level or racial make-up of residents."
The Bay Area has experienced rapid gentrification in the past decade, with 1/3 of poor neighborhoods experiencing some form of gentrification between 2013 and 2017, the highest rate in the country.
Gentrification is difficult to measure, as there are a lot of factors that could be used to classify a neighborhood as being "gentrified" some easier to track than others. The Urban Displacement Project has produced a system for classifying neighborhoods in varying stages of gentrification, which we will use for this project.
Since gentrification is a powerful transformation of a space, we wanted to explore any trends in schools/students related to that transformation. Oakland Unified School District has some public facing datasets which we will analyze in the context of our gentrification data.
Oakland Unified School District (OUSD) is a highly diverse school district serving 35,000 students across a mosaic of campuses, including public and charter schools. The majority of the students OUSD attend “Title 1” schools and receive free and reduced lunch.
While gentrification has provided a mixed bag of outcomes, the changes have been real. Students, teachers and administrators have been forced to confront and grapple with low enrollment, budget cuts and a declining quality of education. While gentrification may bring money and resources into communities and change cities rapidly, schools have largely remained unchanged, outdated and underfunded.
One of the major determinants of resource supply and demand is OUSD’s school choice policy. Students can theoretically apply to any school in the district, creating a fairly fluid network of choices within the district. This policy has been studied extensively and has been shown to create segregation across the district further exacerbating issues around funding and equity.
Our research will examine three different comprehensive district high schools. We chose these schools due to their gentrification categorization (Urban Displacement Project) while controlling for relative school size.
Our research hopes to examine the relationship between three key relationships:
- 1: How has school enrollment demand changed across our study area over the last five years?
- 2: What community resources help create communities that are resilient to gentrification?
- 3: How has the increase in average home value across our study areas impacted standardized test scores (CAASPP MAth and English) across five years?
Create compelling geospatial visual representations of data that accurately show how our study areas have changed over time.
Conduct statistical analysis that can give insight into how gentrification can be identified and studied.
Use a mixture of qualitative and quantitative data to contextualize and provide nuance for our analysis.
A few datasets we may use:
- Urban Displacement Project
- Oakland School Locations
- Oakland School Educational Data
- American Community Survey Data
A few tools and packages we are considering using:
Below we have defined our inital steps with possible types of analyis.
-
Merge school location data w/gentrification classifications to determine what stage of gentrification schools are classified as.
-
Determine a dataset of educational data that we wish to explore in the context of gentrification.
-
Merge our educational dataset with the school/gentrification dataset.
-
Analyze the data in an exploratory manner. This would involve making plots, calculating aggregate statistics, and maybe some other simple statistical methods for seeing if there are any interesting trends in the data.
-
If an interesting trend is found, further explore that trend using more involved methods. Maybe related to linear regression, clustering, removing outliers, etc.
-
Determine the type of tests we will use for our analysis.
- Probabliity Density Function
- Quantile Plot
- Linear Regression
- Kernal Density Estimator
- Interactive Plot
We do not have any strong expected outcomes at the moment. Our objective is to approach this project with an exploratory mindset. We will try to see if there are any interesting patterns or correlations, but given how complicated educational environments are, we will likely not be able to claim any causations in the scope of this project.
We do suspect that gentrification interacts with schools and students in a variety of ways, but we aren't sure if the data will show those interactions or necessarily what they are.
There are three main directories within the project:
notebooksincludes notebooks we used to conduct analysis and organize our thought processes.scriptsincludes a few random scripts for data processessing.dataincludes datasets we used throughout the project.
Overall we found it very hard to draw conclusions mostly because there are a lot of different factors at play and our datasets are fairly limited. Individual notes and observations are below.
No patterns emerged when plotting school size based on gentrification typoloy. In other words, all the small schools or all the big schools are not in one specific typology or even on the same group of being gentrified at risk versus not at risk. Similarly, there was no pattern among schools' average change in enrollment over the past five years based on typology. Notably, the majority of schools at the high end of enrolmment changes were elementary schools.
If I were to keep working on this project, I would want to do more focused analyzes among schools in similiar typologies and among school types (elementary, midlle and high). It would also be helpful to bring in other datasets like the American Community Survey to see what other socio-economic factors could be contributing to students' achievement and satisfaction.
Education data is oftentimes difficult to analyze due to an array of variables that must be accounted for in addition to inconsistent data collection. Perhaps the most difficult part of our project was identifying quality data that had been collected for a continuous time period. My analysis centered around two specific domains, School Policy/Planning and Student Centered outcomes.
Oakland Unified School District operates under an "Open Enrollment" policy, meaning that students can theoretically attend any school in the district. The intention behind this policy was to allow individual families to be able to send their students to other schools in the district if transportation, student interests or quality of education did not align with their needs. In practice this means that students travel close to a 250,000 miles every day to attend school while destabilizing enrollment for individual schools.
Our analysis of school demand indicated that two of the four schools saw a statistically significant increase (R^2>0.5) in demand across our 10 year study period, while the other two saw no change. Our ANOVA indicated that one site had a median enrollment demand value that was higher than the other three sites. However when we looked at the dataset as a whole there was little to no change in enrollment demand across the district. Similarly when analyzing the graduation data our dataset could not confirm an increase in graduation rates across the district, while individual sites saw fluctuations during this period.
To make any definitive conclusions would be reckless, however it would be interesting to look further into these relationships to determine if this expensive and complicated enrollment policy achieves the stated goals.
I looked at how students sense of belonging changed in relation to gentrification. There weren't any strong trends, but one thing that seemed interesting is the two schools w/highest positive change in sense of belonging were located on the edge of an "Advanced Exclusive" area.
I found it hard to draw meaning/relationships in my analysis. A student's sense of belonging has a lot to do with the school environment/individual staff members, which may be affected by gentrification but probably very indirectly.
Our datasets are also relatively small, so any outliers make a huge difference and it is tough to run traditional statistical analyses robustly.
I think that in general, we struggled a lot with data transparency and aggregation, which I think may be common in many disciplines, but seemed stark contrast to some of the other geographic datasets we looked at. While all data requires processing, we struggled a lot with just downloading comprehensive datasets for educational statistics, which slowed us down quite a bit, yet alone filtering/analyzing those as best we could.
Some questions we thought would be interesting to explore in the future:
- Does learning about gentrification in local communities affect students? Could look at affects on interests, goals, mental health, grades, etc.
- Does gentrification/red lining affect school districts with more neighborhood-based enrollment policies (e.g. Seattle)?
- Does gentrification affect staff retainment?
- What factors contribute to changes in enrollment and demand if not gentrification?
We are currently looking for related research on gentrification's impact on schools and will update our outline as we find things.