A discrete choice task to measure preferences for harsh discipline among parents of young children
Maltreatment in early childhood is difficult to measure. Self-report surveys of parents and guardians are the most common method used, but in many settings social desirabil- ity may lead to underestimates of prevalence. There is also reason to be concerned about response bias in the context of intervention trials. To diversify the tools available to in- tervention researchers, we created and tested a discrete choice experiment to elicit parent preferences for harsh discipline. This study was conducted in Liberia as part of a random- ized controlled trial of a positive parenting program. Baseline data were collected from 609 parents and guardians living in Monrovia and caring for a child age 3 or 4. Participants completed a discrete choice experiment that consisted of 12 parenting vignettes, in addition to a survey of parenting attitudes and behaviors. The vignettes were presented to parents as digital comic strips. Each scene could vary on four attributes: child gender; child offense; setting; number of adults present. For each scene, participants selected 1 of 5 discipline strategies that they would in that situation. The visual discrete choice task was easy to implement, well understood by participants, and has evidence of construct validity as a measure of parent preferences for harsh discipline. Tools like this expand the options for researchers studying the maltreatment of pre-school age children, particularly in the con- text of program evaluations where post-intervention observations may be at increased risk of response bias. It may also be useful in a clinical context.