Date of Award

Fall 1-5-2018

Degree Type


Degree Name

Master of Public Health (MPH)


Public Health

First Advisor

Daniel J. Whitaker, Ph.D

Second Advisor

Wendy P. Guastaferro, Ph.D.

Third Advisor

Jessica Rogers Brown, MPH


Background: Most common forms of child maltreatment are neglect and physical abuse. In coerced populations, children are at risk of child maltreatment leading to a variety of negative outcomes including social, emotional and behavioral problems. Interventions for coerced populations are provided and focused on parenting programs. Ultimately, these interventions can help improve parenting behaviors and benefit children. Interventions for parents must be evaluated using standardized assessment tools, and thus one important question that arises is how best to assess parenting. Observational methods (observing a parent and child interact) are often regarded as the gold standard in the assessment of parental behaviors, but are cumbersome to administer. Self-reports of parenting behaviors are the most commonly used measure due to ease of administration, but their validity may be questioned, especially among populations such as parents with substance use disorder who may be coerced into treatment by the child welfare or criminal justice system. The goal of this study is to examine the relationship between two observational measures of parenting and several self-report measures.

Methods: Participants (n=133) were either parents who were receiving treatment at metro-Atlanta drug courts or other caregivers. All completed self-report measures of parenting, completed a videotaped interaction task with a child. Self-report measures collected include the Parent-Child Communication Scales and the Alabama Parenting Scale. The observational task included a play activity, structured discussions, or both, depending on the age of the child. Videos were coded for a variety of behaviors, and two of those behaviors (affection/warmth and involvement) matched constructs that parents reported on in a self-report battery.

Results: Correlations between self- report and observational measures for the constructs affection and involvement for the whole sample ranged from r = -.03 to.06 for affection, and r = -.05 to .08 for involvement, but none were statistically significant. Relationship between self-report and observed parenting by adult type and child age were also examined, however none of the correlations were statistically significant.

Conclusion: Although there were no significant correlations found between self-report and observational measures, it does not undermine the importance of the study question, as current research suggests that self-reports are not interchangeable with observational methods. Future directions of the research would be to examine CAIC (observational tool) in more detail. Also when comparing constructs in self-reports and observational methods, items should be associated according to concept targeted rather than using the generalized subscales.