University of Maryland
Sociology 498: Homelessness 

Census homelessness data: by poverty rates.

homelessness rate (log)
    |
4.5 +                     Washington
    |
    |          SanFrancisco
    |
    |
    |
    |
4.0 +
    |         Seattle
    |                          Boston
    |            SanDiego       NewYork
    |
    |               Portland
    |
3.5 +
    |                  OklahomaCity
    |                      Denver
    |                              Philadelphia
    |              Phoenix            Chicago
    |                      FortWorth
    |
3.0 +                                  Baltimore
    |                          LosAngeles
    |            Nashville                                      NewOrleans
    |                 KansasCity
    |                        Dallas
    | SanJose
    |                             Tucson                          Detroit
2.5 +
    |           Jacksonville        Houston
    |                      Columbus                      Cleveland
    |                        Austin
    |                     LongBeach
    |                                           ElPaso
    |          Indianapolis              SanAntonio
2.0 +                                   Milwaukee
    |
    |
    |
    |                                     Memphis
    |
    ----+------------+------------+------------+------------+------------+-
       10           15           20           25           30           35
                                  % poor of all persons

The correlation of poverty rates and homelessness rates across these 33 large cities is -.34. This means that cities with more poverty have a slight tendency to have less homelessness -- the reverse of what our theories would predict. This correlation is just statistically significant at the p=.05 level. (This means that we would expect a correlation this large even by chance 5 out of 100 times.)

Across all 200 cities with more than 100,000 population, the poverty - homelessness correlation becomes positive (+.34) which is closer to what our theories would expect. The correlation is not large however.

see also


return to: city data overview Burt overview Sociology 498 home page Sociology 498 schedule

Last updated September 1, 2002
comments to: Reeve Vanneman. reeve@cwmills.umd.edu