Parental Self-efficacy and Sensitivity During Playtime Interactions with Young Children: Unpacking the Curvilinear Association

Abstract

Although scholars and service providers typically assume that parental self-efficacy (PSE) facilitates sensitive parenting, this study finds that the association between PSE and observer ratings of parents' sensitivity during free play actually is curvilinear. Parents (33 mothers, 5 fathers) from an intergenerational learning program serving low-income families completed the Self-Efficacy for Parenting Tasks Index and were videotaped during a 15-minute play period with their infant/toddler at program entry. Parents' self-reported PSE is positively, albeit weakly, associated with observer ratings of parental sensitivity from low to moderate levels of self-efficacy but inversely associated with sensitivity from moderate to high self-efficacy. Qualitative analyses show that mothers who report high PSE but are rated as low in sensitivity introduce toys when their child is already engaged, restrict their child's access to toys, physically manipulate their child, and violate their child's proximal space. Parenting education programs need to develop strategies for reaching parents who enter with high levels of confidence but lack knowledge of competent parenting practices.

Keywords:

• Parental Self-efficacy
• Parental Sensitivity
• Observational Methods
• Parent–Child Interaction
• Parenting Programs

Notes

[1] Because only one parent was videotaped playing with their child even when two attended the intake, a total of 46 parents completed the PSE measure, but only 40 were also videotaped playing with their child. Videos of two parents were lost due to equipment failure.

[2] Because the intergenerational learning program served only 16 families at one time, data for this study were gathered over a two-year period and coding was conducted in three phases. The first coding phase (n = 20 families) was completed independently by one of the authors and two trained student coders. For the second phase, the same author who coded in phase one trained three new coders. The new coders practiced rating segments from the first 20 tapes so their ratings could be compared not only with each other but also with ratings from the first phase of coding. Once they achieved sufficient agreement, the three coders independently rated nine additional videotapes. Inter-rater agreement for sensitivity and intrusiveness remained satisfactory, with ICC = .69 in both cases. Agreement for ratings of detachment was somewhat lower for the second group of coders, with ICC = .58. For the third phase, two authors trained four new coders using the same methods as in phase two. Coders in phase three rated the nine remaining tapes. ICC measures were sensitivity = .73, intrusiveness = .75, and detachment = .70.

[3] To assess whether the quadratic association was robust, we removed outliers for both PSE and sensitivity and reran analyses. For PSE, we removed the parent with the lowest SEPTI score (164; see Figure 1). Even with this parent removed, the quadratic association between PSE and sensitivity remained significant after controlling for child sex and the linear effect, β = −.34, p = .049. We also reran the analysis after deleting the parent with the lowest sensitivity score (1.93 on a 1–5 scale). Again, even with this case removed, the quadratic association between PSE and sensitivity remained significant, β = −.40, p = .019.