1. How many studies in the meta5. How should the overall mean analysis used official measures d be interpreted in the studies of repeat victimization as the that used official measures of outcome variable? repeat victimization? 2. How many studies in the meta- 6. How should the overall mean analysis used victim report of d be interpreted for the studdomestic violence as the out- ies that used victim report of come measure? domestic violence? 3. What is the largest effect size 7. Does the 95% confidence for the studies using official interval for the overall mean measures of repeat victimiza- d in Figure 1 include 0 ? Why tion? What is the smallest? does this matter? (Note: If you What is the Overall mean d ? are unfamiliar with confidence 4. What is the largest effect size intervals, please see Exercise for the studies using victim 5.15.) report of domestic violence? 8. How many of the studies in What is the smallest? What is Figure 2 have a d in the negathe overall mean d ? tive range? 146 Questions for Discussion 9. What do you make of the dif- report and victim report). ference between the Overall Does this seem like a small, Mean d in Figure 1 and medium, or large set of studFigure 2? ies for a meta-analysis? 10. There are seven studies per outcome variable (official The aim of this systematic review is to assess the effects of postarrest court-mandated interventions .... for domestic violence offenders ... with the aim of reducing their future likelihood of re-assaulting above and beyond what would have been expected by routine legal procedures. We searched numerous computerized databases and websites, bibliographies of published reviews of related literature and scrutiny of annotated bibliographies of related literature. Our goal was to identify all published and unpublished literature that met our selection criteria ... ... For ... experimental ... designs, control conditions could be no-treatment or treatment as usual. That is, the no-treatment Figure 1: Effect Size (d) and 95% Confidence Interval for Official Measure from Experimental (Random) Studies Author and Year N Favors Comparison Favors Treatment Figures 1 and 2 control condition could include routine legal interventions such as probation or a short jail stay ... This systematic review used standard meta-analytic methods... Effects measured on continuous type measures ... were encoded as standardized mean difference type effect sizes (d)... Official reports were either official complaints made to the police that may or may not have resulted in an arrest, or actual arrests for domestic violence... Figure 1 indicates a general pattern of positive effects on official reports of repeat victimization in these experimental studies... Victim Reported Outcomes: A concern with official measures is that they may not accurately reflect the amount and severity of ongoing violence. Research consistently indicates that official reports capture only a small fraction of this abuse... As such, the victim is viewed as the best source for information on the offender's continued abuse. Given that, we turn our attention to the ... estimates we have from these studies on the effect of these programs according to the victim's reports of abuse ... The distribution of effects is shown in Figure 2. 1. How many studies in the meta5. How should the overall mean analysis used official measures d be interpreted in the studies of repeat victimization as the that used official measures of outcome variable? repeat victimization? 2. How many studies in the meta- 6. How should the overall mean analysis used victim report of d be interpreted for the studdomestic violence as the out- ies that used victim report of come measure? domestic violence? 3. What is the largest effect size 7. Does the 95% confidence for the studies using official interval for the overall mean measures of repeat victimiza- d in Figure 1 include 0 ? Why tion? What is the smallest? does this matter? (Note: If you What is the Overall mean d ? are unfamiliar with confidence 4. What is the largest effect size intervals, please see Exercise for the studies using victim 5.15.) report of domestic violence? 8. How many of the studies in What is the smallest? What is Figure 2 have a d in the negathe overall mean d ? tive range? 146 Questions for Discussion 9. What do you make of the dif- report and victim report). ference between the Overall Does this seem like a small, Mean d in Figure 1 and medium, or large set of studFigure 2? ies for a meta-analysis? 10. There are seven studies per outcome variable (official The aim of this systematic review is to assess the effects of postarrest court-mandated interventions .... for domestic violence offenders ... with the aim of reducing their future likelihood of re-assaulting above and beyond what would have been expected by routine legal procedures. We searched numerous computerized databases and websites, bibliographies of published reviews of related literature and scrutiny of annotated bibliographies of related literature. Our goal was to identify all published and unpublished literature that met our selection criteria ... ... For ... experimental ... designs, control conditions could be no-treatment or treatment as usual. That is, the no-treatment Figure 1: Effect Size (d) and 95% Confidence Interval for Official Measure from Experimental (Random) Studies Author and Year N Favors Comparison Favors Treatment Figures 1 and 2 control condition could include routine legal interventions such as probation or a short jail stay ... This systematic review used standard meta-analytic methods... Effects measured on continuous type measures ... were encoded as standardized mean difference type effect sizes (d)... Official reports were either official complaints made to the police that may or may not have resulted in an arrest, or actual arrests for domestic violence... Figure 1 indicates a general pattern of positive effects on official reports of repeat victimization in these experimental studies... Victim Reported Outcomes: A concern with official measures is that they may not accurately reflect the amount and severity of ongoing violence. Research consistently indicates that official reports capture only a small fraction of this abuse... As such, the victim is viewed as the best source for information on the offender's continued abuse. Given that, we turn our attention to the ... estimates we have from these studies on the effect of these programs according to the victim's reports of abuse ... The distribution of effects is shown in Figure 2
