«Moving to opportunity voluMe 14, nuMber 2 • 2012 U.S. Department of Housing and Urban Development | Office of Policy Development and Research ...»
the control group, and female youth who moved through MTO experienced declines in some risky behaviors (drinking) and improvements in mental health outcomes compared with female youth in the control group.
Neighborhood and School Environments Ludwig (2012) showed MTO effects on neighborhood poverty and a range of neighborhood characteristics with improved reports of safety, crime rates, and household victimization in the experimental group compared with effects on households in the control group. Whether youth from these households experience the same types of neighborhood differences depends on where they spend their time and, in some cases, whether they continue to live with their parents. Like the adults whose reports about neighborhood safety Ludwig (2012) described, MTO youth reported feeling safer and witnessing less drug use and selling in their neighborhoods, as exhibit 1 shows.
One might expect such effects on neighborhood poverty and safety to translate to improvements in schools. Exhibit 2 shows, however, that MTO had relatively modest and somewhat mixed effects on school environments. We present youths’ average school by combining characteristics for all schools and weighting them by the number of grades attended at each school. Both the experimental and Section 8 treatments had statistically significant effects on the racial and socioeconomic composition and schoolwide student academic performance of the new schools attended by youth in those groups. Youth in the control group, however, attended schools that, on average, had overwhelmingly (about 90 percent) minority enrollments, and most (70 percent) students eligible for
free or reduced-priced lunches, so despite these MTO effects, the new schools differed only very slightly on these characteristics. For example, with a TOT estimate of 7.7 percentage points on share minority and a TOT estimate of 10.1 percentage points on share eligible for free lunch, youth who moved with the low-poverty voucher were still in schools that had 82.3 percent minority enrollments and where more than one-half of the student population was eligible for free or reduced-price lunches. Youth in both treatment groups were also more likely to attend larger schools than were youth in the control group, but an increase in the pupil-to-teacher ratio did not accompany the increase in student body size. Although MTO effects on school characteristics are mixed, youth in the experimental group were more likely to report a favorable climate in their most recent school, and the effects appear to be driven by interactions with teachers; youth in the experimental group were more likely to report that teachers were interested in students and less likely to report that they felt put down by their teachers.8
The analysis is limited to youth younger than age 18 as of December 2007, because older youth had to recall secondary school experiences from multiple years before the interview.
youth mean for the 10-point routing score for reading is about 0.6 points higher than the national mean for eighth graders in the lowest quintile of socioeconomic status; for math, the MTO youth mean is about 0.4 points lower than the comparable national average (Najarian et al., 2009).
Exhibit 4 shows that, in general, MTO also had few detectable effects on measures of educational completion and employment. As of the survey interview (late 2008 to early 2010), about one-fifth of the older youth (ages 15 to 20 as of December 2007) sample was currently idle; that is, neither working nor in school. Another way of characterizing this finding is that about 80 percent of the older youth sample was educationally on track; that is, was currently in school or had received a high school diploma or GED. About one-fourth had attended any college, most at 2-year and public colleges (control group means of 14 percent and 15 percent, respectively). About 40 percent reported being currently employed. Self-reported employment rates appear lower for experimental group members, a finding the UI data corroborated (although variation by site merits further investigation).9 MTO had few detectable effects on grown children’s schooling, completed education, and participation in employment or training, based on proxy reports from adults, UI records, and postsecondary enrollment data from NSC. Parents reported that nearly 40 percent of grown children were idle and that about 70 percent had a high school diploma or GED, and NSC data indicate that about 20 percent had attended college since 2007 (as with the younger youth sample, public and 2-year colleges comprised most of the older youth sample’s enrollment). Both parent reports and UI records showed that about one-half of the grown children sampled were employed.10 On all of these measures, female grown children look better than their male counterparts, with mean differences of 10 to 15 percentage points.
The UI data for this finding came from Maryland, Massachusetts, Illinois, and California; data from New York were not available.
Again, the UI data for this finding came from Maryland, Massachusetts, Illinois, and California; data from New York were not available.
Physical and Mental Health Exhibit 5 shows that MTO’s effects on the physical health measures were similar overall and by gender across treatment groups but that the low-poverty vouchers had some encouraging effects on mental health, primarily for female youth. Nearly 90 percent of youth reported good, very good, or excellent (as opposed to poor or fair) health. Almost 20 percent of youth reported having had an asthma or wheezing attack in the past year, and a similar share had had an accident or injury requiring medical attention in the past year.11 Obesity rates were much higher for females
ITT = intention to treat. M = measured. SR = self-reported. TOT = treatment on the treated.
* = p.05. ~ = p.10.
Notes: Robust standard errors shown in parentheses. The control mean is unadjusted. Unless otherwise indicated, the control mean and effects are expressed as shares of the sample in the category (for example, a control mean of 0.250 for working would indicate that 25 percent of the control group was working). Experimental and Section 8 effects were estimated jointly using an ordinary least squares regression model controlling for baseline covariates, weighted, and clustering on family. Youth and grown children effects by gender were estimated as an interaction with treatment status. See the forthcoming technical appendixes to Sanbonmatsu et al. (2011) for details. Obesity is defined according to the International Obesity Task Force. The psychological distress index consists of six items (sadness, nervousness, restlessness, hopelessness, feeling that everything is an effort, worthlessness) scaled on a score from 0 (no distress) to 24 (highest distress) and then converted to z-scores using the mean and standard deviation of control group youth. Results reported differ slightly from those in Sanbonmatsu et al.
(2011) because here standardization was separate by gender, whereas Sanbonmatsu et al. standardized only on the overall control group mean and standard deviation. The overall (male and female combined) z-score values combine the z-scores by gender and thus are not themselves standardized (the control mean is 0 but the standard deviation is not exactly 1). The serious behavioral or emotional problems measure is based on the strengths and difficulties questionnaire, which consisted of five behavioral and emotional items (obedience, worry or anxiety, unhappiness, getting along better with adults than peers, attention span) scaled on score from 0 (no behavioral or emotional problems) to 12 (severe behavioral or emotional problems).
A score of 6 or higher indicates serious behavioral or emotional problems.
Source: MTO youth long-term survey (27 percent) than for males (19 percent).12 Self-reports of overall health, obesity, asthma, and injury rates were similar among youth in the experimental, Section 8, and control groups. For mental health, however, females in the experimental group were 0.12 standard deviations lower than their counterparts in the control group on the K6. They also were 3.3 percentage points less likely to have serious behavioral or emotional problems compared with a mean of 12.7 percent for These MTO obesity rates appear to be higher than even the low socioeconomic status (SES) group in one national study (Sherwood et al., 2008). That study showed that about 34 percent of females and 32 percent of males in the low SES group were overweight, a less stringent standard than obese. For comparison, MTO youth overweight rates were 48 and 42 percent for females and males, respectively.
the control group, a more than 25-percent reduction in prevalence. These results are consistent with results from the interim impacts evaluation (Kling, Liebman, and Katz, 2007; Orr et al.,
2003) and with MTO qualitative investigations in suggesting that female youth responded more strongly to the new and safer social environments in the neighborhoods to which the families in the experimental group moved (Popkin, Leventhal, and Weismann, 2008).
Via proxy reports from the adults, MTO had few detectable effects on the physical and mental health of grown children. Parents reported less than 5 percent of grown children as having a physical health problem that kept them from normal activities and less than 10 percent as having had a chronic health issue, such as cancer or a heart problem. About 10 percent had depression or another serious mental health problem, and about 5 percent had an alcohol or drug problem.
Physical health problems did not vary by gender, but parents were somewhat more likely to report males as having a mental health or substance abuse problem.
group, and females in the experimental group reported lower rates of drinking than females in the control group.13 In general, the number of violent and property crime arrests for male youth in the control group is about 0.86, more than 3.5 times the number for female youth. MTO increased property crime arrests among female and male youth in the experimental group relative to their peers in the control group, although the point estimates are imprecisely estimated and not quite statistically significant (p =.12). Unlike patterns observed with the followup survey data for the interim impacts evaluation, the number of violent crime arrests, at about 0.48 for males and 0.16 for females, is similar across groups. In results not shown, however, males in the experimental group appear to have a lower number of arrests for drug selling or distribution (roughly 11 percent were ever arrested for drug possession and 8.7 percent were ever arrested for drug distribution).
We have a bit more confidence in this preliminary finding, which calls for further investigation, because the pattern is apparent among both older male youth and grown male children. MTO showed no detectable effects on the property or violent crime arrests of grown children.
Discussion and Conclusions Youth in economically disadvantaged neighborhoods fare less well across a host of educational, health, and behavioral outcomes than do youth who grow up in more economically advantaged neighborhoods. Isolating the extent to which neighborhoods per se contribute to this variation is important for the design of means-tested housing programs and other place-based programs, but isolating neighborhood effects has been challenging in practice because of the difficulty in uncoupling neighborhood effects from so many other aspects of the youth and their families that are associated with moving to and staying in particular places. The MTO study helped overcome these issues of selection bias.
In conducting the long-term survey for the final impacts evaluation, we discovered that MTO had a number of unanticipated effects on youth health (as found with the adults), particularly benefiting female mental health outcomes, but few detectable effects on achievement, education, employment, and a range of health and risky behavior outcomes among either youth or grown children. These patterns of findings are quite similar to what the followup survey for the interim impacts evaluation found. In particular, gender differences in MTO’s effects appear for this cohort of youth in the long-term survey for the final impacts evaluation just as they did for a slightly older cohort of youth in the followup survey for the interim impacts evaluation.
MTO’s effects on achievement and related schooling outcomes were disappointing, particularly among the youngest cohort of children at MTO enrollment, whom we hypothesized would benefit the most from MTO moves into lower poverty neighborhoods. Although many MTO parents seem to believe that the quality of local public schools is an important mechanism through which neighborhood location may influence children’s academic achievement and attainment (Briggs, Popkin, and Goering, 2010), MTO had more modest effects on school quality than on other neighborhood An alternative version of the risky behavior index, not presented here, includes a measure of ever having been pregnant (female) or having impregnated someone (male) instead of the current item on sexual activity. MTO’s effects on the index including the pregnancy item do not statistically differ across groups. Roughly 20 percent of male youth and 27 percent of female youth reported ever having impregnated someone or ever having been pregnant.