Deterring Domestic Violence
Summary. In this study, Sherman and Berk (1984) looked at the effect of three interventions on repetition of domestic assault. The video segment contains a description of the interventions used in the study. This is a good video for stimulating discussion of lab vs. field research, research ethics, random assignment, and practical vs. statistical significance.
Video. Against All Odds: Inside Statistics (1989), Program 12, Experimental Design. The segment on the domestic violence study begins about 17 minutes into the program, and is about 6 minutes long. You can show the entire segment; the study was still in progress when the video was made, thus results are not discussed. The video can be viewed here (email me at firstname.lastname@example.org for the page password).
Original Research. Sherman, L. W. & Berk, R. A. (1984). The specific deterrent effects of arrest for domestic assault. American Sociological Review, 49, 261-272.
Design. The study can be presented as a 2 x 3 cross-tabulation (see Table 1).
Statistical Analysis. A 2 x 3 chi-square test of independence produces a value of 5.76 (p = .056). A 2 x 2 chi-square test comparing just the control and arrest groups yields a value of 5.74 (p = .017), showing that the proportion of repeat assaults in the arrest group (13%) is significantly lower than in the control group (26%).
Published Results. The authors report results from a linear regression analysis, resulting in t-tests for each factor (see their Table 3). “Compared to the baseline treatment of separation, which had the highest recidivism rate in the police data, the arrest treatment reduced repeat occurrences by a statistically significant amount (t = -2.38). . . . The mediation treatment was statistically indistinguishable from the other two.” (p. 267).
Published conclusions. “The official recidivism measures show that the arrested suspects manifested significantly less subsequent violence than those who were ordered to leave. The victim report data show that the arrested subjects manifested significantly less subsequent violence than those who were advised.” (p. 261).
Table 1. Number of participants who did and did not have a repeat arrest for domestic assault within 6 months of the original assault, by treatment condition.
Control Arrest Advise
No repeat assault 66 118 73
Repeat assault 23 18 16
The Excel file for this activity contains the 314 cases in random order.