The Census Tract That Looks Healthy
A county health department maps "wellbeing" using three indicators: median household income, homeownership rate, and life expectancy. A census tract scores well above county averages on all three, so it is designated a "healthy community" and excluded from a new outreach program. A sociologist visits and finds: the tract contains a large university and a hospital that employ many residents; a significant share of residents are graduate students and medical residents who live there temporarily; and the tract's boundaries cut through a low-income neighborhood whose residents shop and send children to school in the tract but live just outside it. The map shows a sharp line between high and low scores at the tract boundary.
1. Identify the methodological problem with using these indicators and tract boundaries, and explain how it could produce a wrong policy decision.
Read the worked answer
The core problem is the ecological fallacy combined with boundary and composition artifacts. Tract-level averages describe places, not the people within them, and the tract's high median income and life expectancy may reflect a transient, institutional population (students, residents) rather than a stable community. The hospital and university concentrate high earners and healthy young adults, pulling averages up. Life expectancy in a tract with many young temporary residents is especially misleading because it reflects who lives there now, not the life course of a settled population.
The boundary compounds this: a real low-income neighborhood is split by the tract line, so its needs are invisible on the map. Excluding the tract from outreach therefore withholds services from people who live adjacent and use the tract's institutions. The decision is wrong not because the numbers are false but because the unit of analysis (the tract) does not match the unit of need (households and the cross-boundary community).
2. A colleague says, "If the data show high wellbeing, the community is fine." Respond using the idea of the unit of analysis.
Read the worked answer
The claim confuses the unit of analysis with the unit of interest. The data are aggregated at the census-tract level; the question is about households and individuals. Aggregation can hide internal variation-a tract can have high median income and contain a substantial poor subpopulation if the distribution is skewed by a large institutional employer. Medians and rates are summary statistics; they do not certify that every resident is thriving.
There is also a selection issue: who is counted as a resident? Temporary students and medical residents inflate the averages while having different needs and timelines than settled families. The colleague's inference assumes homogeneity within the tract, which the site visit contradicts. The correct move is to disaggregate-examine distributions, not just averages-and to check whether tract boundaries align with the community people actually live in. "The data show high wellbeing" is true at one level of analysis and misleading at the level where policy acts.