Mothers' child abuse potential as a predictor of maternal and child behaviors during play‐time interactions

Abstract

This paper explores the effect of mothers' child abuse potential on maternal and child behaviors during play‐time interactions. Forty‐two mothers participated in a 10‐min play period followed by a clean‐up task along with one of their children (range=3–8 years old), after which they completed Milner's (1994) Child Abuse Potential (CAP) Inventory. Mothers' behaviors were coded using a 21‐category scheme, which was collapsed into five behavioral clusters: describing, directing, criticizing, soliciting/affirming, and silent play. Children were rated for levels of involvement (activity) and cooperation. Mothers' child abuse potential scores were inversely associated with their own rate of soliciting/affirming behaviors during the play period. Children of higher‐risk mothers displayed higher levels of involvement combined with lower levels of cooperation compared children of lower‐risk mothers. Implications for understanding and responding to child physical abuse are discussed.

Keywords:

• Child Abuse Potential
• Maternal Behavior
• Child Cooperation/Child Involvement
• Play‐Time Interactions

Notes

Steven R. Wilson and Wendy M. Morgan are at Purdue University. Javette Hayes is at California State University, Fullerton. Carma Bylund is at the University of Iowa. Andrew Herman is at the State University of New York, Geneseo. An earlier version of this paper was presented at the meeting “Victimization of Children & Youth: An International Research Conference,” August 4–7, 2002, Portsmouth New Hampshire. Correspondence to Steve Wilson, Department of Communication, Purdue University, 100 N. University Street, Beering Hall 2114, West Lafayette, IN 47907‐2098, USA. Email: [email protected]

Most prior studies of mother–child interaction in physically abusive versus matched‐comparison families fail to report information about statistical power; however, our sample size (N=42) is similar to earlier observational studies. Bousha and Twentyman (1984) report statically significant differences in maternal and child behavior based on home observations with physically abusive, neglectful, and comparison mothers (n=12 per group). Oldershaw et al. (1986) report significant differences based on lab observations with physically abusive and comparison mothers (n=10 per group). Dolz et al. (1997) report significant differences in maternal and child behavior during home observations with families in which mothers scored high or low on Milner's 1994 Child Abuse Potential (CAP) Inventory (n=10 per group).

Mothers' CAP scores were inversely associated with overall rates for each of the three behavioral categories composing the “soliciting/affirming” cluster (although these correlations did not reach conventional levels of statistical significance): child praise, r (39)=−.25, p=.12, questions about the child's mental state, r (39)=−.29, p=.08, and questions about the toy, task, or setting, r (39)=−.28, p=.09 (all two‐tailed). Thus, the statistically significant correlation between mothers' child abuse potential and their rate of soliciting/affirming behavior appears to be driven by all three categories composing this cluster rather than by only one of the categories.

The child cooperation×involvement interaction remained statistically significant even after controlling for the effects of mothers' SES and marital status. In a subsidiary analysis, we entered the main effects for child cooperation and involvement across the six phases as well mothers' SES and marital status simultaneously into the regression analysis at Step 1, and then the cooperation×involvement interaction at Step 2. The four variables entered at Step 1, as a set, approached conventional levels of statistical significance, R² =.23, F(4, 33)=2.50, p=.061. Martial status (scored 0=not married, 1=married) was a significant predictor of mothers' CAP scores, β=−.37, p<.05, whereas mothers' SES and the main effects for child involvement and cooperation were not significant predictors. The cooperation×involvement interaction still was statistically significant at Step 2, R² Δ=.11, FΔ (1, 32)=5.14, p<.05.

Additional information

Notes on contributors

Steven R. Wilson

Steven R. Wilson and Wendy M. Morgan are at Purdue University. Javette Hayes is at California State University, Fullerton. Carma Bylund is at the University of Iowa. Andrew Herman is at the State University of New York, Geneseo. An earlier version of this paper was presented at the meeting “Victimization of Children & Youth: An International Research Conference,” August 4–7, 2002, Portsmouth New Hampshire. Correspondence to Steve Wilson, Department of Communication, Purdue University, 100 N. University Street, Beering Hall 2114, West Lafayette, IN 47907‐2098, USA. Email: [email protected]