Appendix A.

Sample Comparisons at Baseline, Fall of Sixth Grade 2003

	PROSPER (Iowa and Pennsylvania)					United States
	Full 2003 Cohort	In-home Subsample		Analytic Sample		Rural Sixth-Grade Students		PROSPER-Criteriaa Sixth-Grade Students		All Sixth-Grade Students
Receives free or reduced-price lunch	.34	.35		.33		.37	b	.47	b	.41	b
Racial minority	.16	.13	**	.12	***	.21		.29		.41
Male	.49	.47		.48		.47		.47		.48
Friendship nominations made	4.14	4.32	*	4.33	*	—		—		—
Friendship nominations received	2.80	3.20	***	3.23	***	—		—		—
Past-year delinquency	.07	−.02	***	−.03	***	—		—		—
Past-month substance use	.07	.03	**	.02	***	—		—		—
Sixth-grade students	6,165	977		766		789,494		624,864		3,746,944
Sixth-grade students per school (median)	105	105		104		42		93		77

NOTES: PROSPER descriptive statistics shown here are based on baseline respondents only. The median number of PROSPER students per school is based on school roster data (regardless of participation). The analytic sample is limited to observations from in-home survey respondents who met the following criteria: attending participating school district, participated in the in-school survey, valid data on suspension. US data come from the National Center for Education Statistics (NCES) 2003–2004 and include students from regular (not alternative or other) public schools with at least two sixth-grade students (minimum number necessary to be comparable to a peer network study). “Rural” refers to NCES locale codes (census-defined) and is based on school address. PROSPER data are unweighted. Mean comparisons represent dependent-sample t-tests between each of the smaller PROSPER samples and the full 2003 cohort.

^aStudents attending school districts with enrollments between 1,300 and 5,200 students, at least 15% of which are eligible for free or reduced-price lunch (Spoth, et al., 2007).

^bNational data on lunch status are not available by grade. These represent the proportion of all public school students.

^***p<.001

^**p<.01

^*p<.05 (two-tailed)