University of Maryland
Sociology 498: Homelessness 

Comparing the currently homeless to those who have left homelessness

To compare the currently homeless with the formerly homeless, we use Martha Burt's 1996 national survey of homeless service users. While her currently homeless sample comes about as close to a representative sample of the U.S. homeless as is practical, her formerly homeless sample all continue to use some services often intended for the homeless (e.g., soup kitchens, health care). Formerly homeless persons who continue to use homeless services are undoubtedly poorer than a truly representative sample of the formerly homeless would be. But to get a large enough representative sample of the formerly homeless would be a major undertaking. Burt's sample gives us a conservative test of the difference between the current and formerly homeless since the differences would be even larger if we had a better sample.

graph of mean income for US 1996 homeless
samples

The formerly homeless have about 30% higher incomes than the currently homeless. So exits from homelessness appear to be associated with less poverty.
 

This association is a clue that more income can cause exits from homelessness, but the evidence is hardly decisive:

While these alternative interpretations are plausible, we do not have the data to test them adequately. If we followed homeless persons over time and observed the changes in their income during and after homelessness, we would have better data to test whether increased incomes cause exits from homelessness.
 

Conclusion:

Despite the reservations above, we have to make do with the data we do have. These data are consistent with an interpretation that higher incomes do provide a path out of homelessness.
 
return to: mean incomes overview of income Sociology 498 home page 498 schedule 498 exam questions

Last updated February 13, 2006
comments to: Reeve Vanneman. reeve@cwmills.umd.edu