Written by Justin Meyer
When I first began my research on large, national art museums and their impact on urban space, I set out to measure and characterize the socioeconomics of their adjacent neighborhoods in cross section (that is, analyzing the socioeconomics at one point in time, such as at the end of a decennial census). By analyzing the socioeconomic character of art museum neighborhoods in the United States in cross section, various agents (researchers, planners, communities, and the museums themselves) could have a better understanding of who lives near to them. Another benefit to urban research: a cross sectional characterization of art museum neighborhoods serves as a baseline for measuring change in these neighborhoods over time.
For my cross sectional analysis of art museum neighborhood socioeconomics, I analyzed 154 neighborhoods adjacent to 59 art museums located across the country in the largest 49 US cities plus Salt Lake City (the 50th largest city it did not have an art museum that fit my criteria, and Salt Lake City was the next largest city that did), based on 2010 census place population. Since no definitive sampling frame exists for art museums in the United States, I selected art museums to analyze from these 50 largest cities in the United States, according to following criteria:
- are fiscally independent of any colleges or universities,
- were founded before 2010 (the date of the most recent U.S. Census),
- had assets of at least $100,000 by the 2013 fiscal year (the most recent accounting year available for the majority of institutions),
- are not located within 0.5 mile radius of another art museum.
Figure 1 maps the resultant art museums when I used this selection criteria. An important implication of my selection process: the selected art museums do not represent the outcome of a probability sample, and thus, the results of these analyses cannot be generalized to a larger population of art museums and neighborhoods.
I used census tracts as a spatial unit for neighborhood analysis, based on the Longitudinal Tract Database (LTDB) created by Logan, Xu, and Stults that normalizes the boundaries of census tracts between the 1970 and 2010 U.S. Censuses. Census tracts have been identified as acceptable approximations for neighborhoods in the academic studies of gentrification by Peter Marcuse and Lance Freeman.
For each art museum, I analyzed socioeconomic data for census tracts within a 0.25 mile radius of an art museum, which I defined as art museum neighborhoods (see Figure 2). I used one quarter mile, because it was a minimum distance at which every art museum had at least one corresponding census tract neighborhood. My analysis also included “comparison census tracts” (aka comparison neighborhoods), defined as census tracts that lie within 0.75 mile radius from each sampled art museum, but no closer than 0.5 mile radius (see Figure 2). I selected comparison census tracts in this way to compare neighborhoods in the same general “part of town” as the museum neighborhood census tracts, but still far enough away from the museums to not be considered museum neighborhoods.
I characterized these census tract neighborhoods in cross-section according to six socio-economic constructs: Wealth, Race/ethnicity, Rent affordability, Home value, Education, and Resident turnover (see Figure 3). I used census data to categorize census tracts according to three characteristic qualities within each construct; one that represents a high value, a low value, and a medium value of the construct (see Figure 3 for the precise definitions for these). High and low values of each construct were judged according to the 75th and 25th percentile values of each neighborhood’s corresponding county census tracts, respectively. This was done to normalize the differences in raw data between geographic locations, where socioeconomic characterizations, such as an “expensive” neighborhood, have different quantitative definitions relative to their regions. For example, a neighborhood in a city like Indianapolis with a median home value of $150,000 would be considered to have high home values; whereas a neighborhood in New York City or San Francisco with a median home value of $150,000 would be considered to have (very) low home values.
I characterized museum and comparison census tracts in this way at five time interval points using census tract level data standardized by the Longitudinal Tract Database for the last five U.S. censuses: 1970, 1980, 1990, 2000, and 2010. I did ensure that all museum census tracts in past censuses were indeed adjacent to an art museum at the time by excluding those whose corresponding museums had not yet been built.
After I characterized each census tract, I calculated proportions for each construct characteristic (for example, the number of museum census tracts that are wealthy divided by the total number of museum census tracts) for both museum and comparison census tract neighborhoods. Then, I graphed these proportions side by side for ease of comparing between censuses and between museum and comparison neighborhood census tracts (see Figures 4 – 6). The x-axis for each graph in Figures 4 – 6 contains the five census data points (1970, 1980, 1990, 2000, and 2010), and the y-axis measures percentage (or proportion) of census tract neighborhoods with a particular characterization.
The results of the cross-sectional analyses suggest that museum neighborhoods at the most recent 2010 census were more likely to:
- be low-income than high income (39% vs. 31%, Figure 4),
- be expensive to rent compared to other neighborhoods (40%, Figure 5),
- have high home values (50%, Figure 5).
Further, a majority of museum neighborhoods at the 2010 census:
- were ethnically/racially diverse (55%, Figure 4),
- were highly educated (57%, Figure 6),
- had high rates of new residents compared to other neighborhoods in their counties (63%, Figure 6).
When compared to the comparison census tracts, museum census tracts in 2010 were more likely to:
- be high income (31% vs. 17%, Figure 4),
- be high white (34% vs. 25%, Figure 4),
- have expensive rent (40% vs. 26%, Figure 5),
- have high home values (50% vs. 37%, Figure 5),
- have high college attainment (57% vs. 39%, Figure 6),
- have high resident turnover (63% vs. 50%, Figure 6).
And less likely to:
- be high nonwhite (11% vs. 23%, Figure 4),
- have affordable rent (30% vs. 37%, Figure 5),
- have low home values (16% vs. 22%, Figure 5),
- have low college attainment (12% vs. 21%, Figure 6).
Though these comparisons reveal telling differences between neighborhoods closest to the selected museums and those farther away, they do not imply that proximity to the selected art museums cause these differences. Unaccounted for variables (such as other neighborhood institutions, actors, or events) and an unspecified order of events (whether the neighborhood led to the formation of the museum, or vice versa) keep this particular analysis from supporting such causal claims. (However, I am in the process of refining an analysis that accounts for other variables and a specified order of events, making an argument that these art museums cause certain changes in their neighborhoods. I hope to publish this analysis later this year.)
When looking across the five censuses, both museum and comparison census tract neighborhoods show signs that they have since 1970 become:
- wealthier (larger proportions of high income neighborhoods and smaller proportions of low income neighborhoods, Figure 4),
- more racially/ethnically diverse and less nonwhite (Figure 4),
- more expensive (larger proportions of expensive rent neighborhoods and high home value neighborhoods, Figure 5),
- home to newer residents (higher proportions of high turnover neighborhoods, Figure 6).
In contrast, both museum and comparison census tract neighborhoods have remained fairly stable in the levels of college attainment of their residents (Figure 6).
This socioeconomic analysis of neighborhoods nearby to a selection of large art museums in the United States shows that these neighborhoods tend to be: home to diverse racial/ethnic groups, poorer, highly educated, and at the same time more expensive than other neighborhoods elsewhere in their counties. Knowing this information can help art museums like those selected understand a little bit about who live nearby to them, and possibly improve how these art museums program and form relationships with their neighbors. This examination of neighborhoods around art museums also underscores a vital role for arts institutions as socially conscious actors negotiating a variety of urban actors (including minority and low income groups) in highly contested urban space.