930
Susan J. Popkin, Larry F. Buron, Diane K. Levy, and Mary K. Cunningham
of locating former participants, only a handful of participants who had
either moved back to the city or lost their Section 8 assistance were
included in these samples, meaning that it is likely that many unsuc-
cessful movers were excluded. Thus, any findings about positive effects
may be inflated. Finally, as many critics have pointed out, the Gautreaux
studies were not a random experiment but rather took advantage of a
“natural experiment” to compare outcomes for city and suburban movers.
For all of these reasons, it would be a mistake to view this research as
conclusive evidence of the potential benefits of dispersal programs for
very low income public housing residents.
HUD’s 10-year MTO demonstration was intended to address some of
the shortcomings of the Gautreaux research. MTO was implemented
in five cities (Baltimore, Boston, Chicago, Los Angeles, and New York)
and differed from Gautreaux in that it used the poverty rate of the
receiving neighborhood rather than its racial composition as the dis-
persal criterion. In each city, samples of public housing residents were
randomly assigned to one of three groups: In the experimental group,
families were offered a special MTO certificate, which could be used
only in census tracts in which less than 10 percent of the households
were below the poverty level, along with counseling and housing search
assistance. In the second group, families were offered a regular Section
8 certificate or voucher with no additional support, and in the in-place
control group, families received no Section 8 assistance at all.
While MTO should answer many of the questions about the effects of
dispersal strategies, it will be some years before the full results on par-
ticipant outcomes are available (HUD 1996, 1999). However, reports
from small studies conducted on the individual sites (Katz, Kling, and
Liebman 2000; Leventhal and Brooks-Gunn, forthcoming; Ludwig and
Duncan, forthcoming; Pettit and McLanahan 1998; Rosenbaum and
Harris, forthcoming) indicate a number of positive short-term outcomes
for the experimental households. The studies generally find that house-
holds in the experimental group that succeeded in moving are living in
safer neighborhoods, with lower levels of drug trafficking and violence
and higher levels of social organization. Individual studies have found
evidence that children in these families seem to have fewer behavior
problems, a reduced incidence of arrest and convictions, fewer injuries,
and fewer episodes of asthma, and that mothers hold positive views of
their children’s new schools and teachers. Studies in Boston and New
York indicate that adults seem to have improved mental health. Some
families, however, have reported being concerned about their children’s
interaction with new schoolmates and teachers. While these results are
exciting, it is not clear that they are generalizable: The MTO population
suffers from the same self-selection biases as the Gautreaux partici-
pants in that the families volunteered for this special experiment and
therefore may differ in unknown ways from the larger population of
public housing residents. Further, only about half of the experimental