Factor structure of the Postpartum Bonding Questionnaire: Configural invariance and measurement invariance across postpartum time periods

Abstract

BackgroundDespite the research and clinical significance of the Postpartum Bonding Questionnaire (PBQ), its psychometric properties have not been studied intensively. The goodness-of-fit of proposed factor models of the PBQ is poor. Configural and measurement invariance have never been reported. MethodsAs a secondary analysis of the previous paper (Ohashi et al., 2016), we analysed the PBQ data at 5 days and 1 month after childbirth among 247 mothers of a singleton. We created 9 parcels of PBQ items to perform confirmatory factor analysis (CFA). We also examined configural and measurement invariances of the PBQ factor structure between the two observation times. ResultsThe CFI of the 3-factor model of the PBQ was .936 and .968 for 5 days and 1 month after childbirth, respectively. Configural, measurement (metric, scalar, and residual), and structural (factor variance and factor covariance) invariances were accepted. The mean of only ‘anger and restrictedness’ factor was scored higher at 1 month than 5 days after childbirth. ConclusionThe 3-factor model of the PBQ was good in its fit with the data as well as robust in its measurement between the two observation time periods.