<< News Scan | Main | News Scan >>

Studies Show ... Or Maybe Not

Eric Jaffe has this piece in the Atlantic Cities, Do Foreclosures Increase Crime After All?

A few months ago our Emily Badger reported on a study that found no direct correlation between home foreclosures and crime. The researchers drew that conclusion from a thorough analysis of housing markets and crime statistics in 142 metro areas across the United States. "Moving forward, researchers should begin to think about why the foreclosure crisis is not directly linked to rates of violent and property crime," they wrote.

Before other scholars move forward too quickly, they might consider another study -- this one set to appear in the Journal of Urban Economics -- that reaches the opposite conclusion. The new work, led by Ingrid Gould Ellen of New York University, found that recent foreclosures in New York City led to a 1 percent increase in crime on the same block. The largest effects occurred when properties went all the way through the foreclosure process, either to auction or bank ownership.

Just as important as the foreclosure-crime issue is the lesson about overreliance on studies.  All studies involve assumptions and simplifications.  If they are well-written, these are detailed in the "Limitations" section.  Go there first, just like the footnotes in a financial disclosure.  That's where the bodies are buried.  Different researchers get different and sometimes conflicting answers to the same questions because they made different assumptions or simplifications or used different methods.

Far too often, the "bottom line" of a study is cited in a policy advocacy piece as if it were the definitive word without any indication the study has limitations.  Take such citations with a heavy dose of skepticism.  What "studies show" "ain't necessarily so."

Leave a comment

Monthly Archives