Improving risk prediction accuracy for new soldiers in the U.S. Army by adding self-report survey data to administrative data

Table 1 Pearson correlations between composite risk scores based on the HADS and the NSS by month in the NSS sample (n = 18,838 men and 2952 women)^a

	Physical violence perpetration^b	Sexual violence perpetration^b	Sexual violence victimization^c
Month	r	r	r
0	0.46	0.04	0.34
1	0.45	0.03	0.35
2	0.44	0.02	0.34
3	0.42	0.03	0.33
4	0.37	0.03	0.33
5	0.39	0.04	0.30
6	0.41	0.04	0.26
7	0.42	0.04	0.27
8	0.42	0.05	0.28
9	0.41	0.05	0.28
10	0.39	0.05	0.27
11	0.39	0.06	0.26
12	0.38	0.06	0.27
13	0.37	0.06	0.26
14	0.36	0.07	0.25
15	0.35	0.07	0.24
16	0.34	0.07	0.24
17	0.34	0.07	0.24
18	0.34	0.07	0.25
19	0.34	0.07	0.26
20	0.34	0.07	0.27
21	0.34	0.06	0.27
22	0.34	0.06	0.27
23	0.34	0.06	0.26
24	0.34	0.06	0.26
25	0.35	0.07	0.25
26	0.36	0.07	0.24
27	0.35	0.07	0.26
28	0.36	0.07	0.25
29	0.36	0.07	0.24
30	0.36	0.07	0.22
31	0.36	0.07	0.22
32	0.37	0.06	0.24
33	0.38	0.06	0.24
34	0.36	0.06	0.25
35	0.34	0.05	0.28
36+	0.36	0.06	0.24
25% Quartile	0.34	0.05	0.24
Median	0.36	0.06	0.26
75% Quartile	0.39	0.07	0.27

^aThe NSS respondents considered here were surveyed between April 2011 and November 2012. Administrative data were available through December 2014 (25-44 months after the survey). The sample size decreased with duration both because of attrition and because of variation in time between survey and end of the follow-up period. The sample included 18,838 men (decreasing to 16,479 by 12 months, 15,306 by 24 months, and 3,729 by 36 months) and 2,952 women (decreasing to 2,300 by 12 months, 2,094 by 24 months, and 687 by 36 months)
^bMales only
^cFemales only

ISSN: 1471-244X