Issues of Reliability in Measuring Intimate Partner Violence during Courtship

The current paper focuses on problems in conceptualizing and establishing reliability when using self-administered measures of intimate partner violence to assess dating violence. Establishing reliability is an important step in the development of dating violence assessment instruments. However, the nature of dating violence can make it difficult to establish reliability. Most notably, measures of intimate partner violence in courtship yield data that are positively skewed, with almost no one reporting high levels of violence. This could have implications for the calculation of several forms of reliability that assume normality (e.g., Pearson correlations, intraclass correlations). In addition, there are other characteristics of dating violence that could impact reliability. For example, partner violence perpetrators do not necessarily use multiple acts (internal consistency reliability) or repeat specific acts (test-retest reliability). And, gender differences in the perception of partner violence may influence intra-couple reliability in heterosexual couples. Finally, statistical interdependence within couples makes current intra-couple reliability assessment suspect. Research is reviewed and recommendations are made concerning the establishment of test-retest reliability, intra-couple reliability, and internal consistency reliability for measures of dating violence.