TABLE 2—

Neighborhood Characteristics at Follow-Up by Treatment Group: MTO Demonstration: New York City, 1998–2000

	Experimental (n = 220)	Section 8 (n = 181)	In-Place Control (n = 149)	Difference: Experimental vs Controla	Difference: Section 8vs Controla
1990 US Censusb
Median family income, $	23 277 (14 684)	17 922 (9283)	14 808 (6531)	+8469†	+3114**
Fraction poor	0.34 (0.20)	0.40 (0.14)	0.45 (0.12)	−0.11†	−0.05**
Fraction rental units	0.82 (0.25)	0.92 (0.15)	0.94 (0.15)	−0.12†	−0.02
Fraction Black	0.45 (0.25)	0.40 (0.24)	0.41 (0.21)	0.04	−0.01
Fraction Latino	0.41 (0.25)	0.48 (0.23)	0.51 (0.20)	+0.10†	−0.03
Fraction White	0.12 (0.21)	0.09 (0.18)	0.06 (0.16)	+0.06**	+0.03
Parental reportc
Disorderd	2.05 (0.73)	2.17 (0.65)	2.38 (0.56)	−0.33†	−0.21***
Satisfactione	3.07 (1.42)	2.83 (1.33)	2.58 (1.39)	+0.49†	+0.25*
Interviewer observationb
Poor external environmentf	2.27 (1.40)	2.45 (1.31)	2.72 (1.16)	−0.45***	−0.27

Note. MTO = Moving to Opportunity for Fair Housing Demonstration. Means (SDs) are weighted by date of random assignment because the assignment ratio for the 3 groups changed throughout the randomization period.

^aSignificance levels indicate significant difference compared with in-place controls.

^bTest statistic computed via analysis of variance, with mean comparisons Bonferroni-adjusted.

^cMeans adjusted for baseline characteristics: parental sex, race/ethnicity, age, education, employment status, marital status, and number of children in household; missing baseline characteristics were imputed to the mean of the nonmissing sample. Test statistics were computed via ordinary least squares regression.

^dParents reported “how big a problem” 5 types of events in the neighborhood were, on a 3-point scale from “a big problem” (3) to “not a big problem” (1). For disorder scale, range = 1–3.

^eFor satisfaction, range = 1–5.

^fInterviewer observed the street block on 4 attributes, and items were recoded dichotomously, with higher scores reflecting lower quality (range 0–4).

*P < .10;

**P < .05;

***P < .01;

†P < .001.