We predicted low perceived caregiver control over caregiving failure to be related to (a) coercive or abusive parenting and (b) affective reactions to "difficult" children. On the basis of a multidimensional scaling analysis of the Parent Attribution Test (Study 1), we constructed a scale (PCF) that assessed perceived balance of control over caregiving failure (attributed control to caregivers vs. attributed control to children). In Study 2, we found low PCF to predict abusiveness and nonabusive coerciveness among mothers in counseling at a child abuse agency. Additionally, we found low PCF to predict experienced annoyance/irritation among unrelated mothers interacting with children at relatively high risk for abuse (compared with their lower-risk siblings). We interpreted results as demonstrating the potential importance of low perceived control as a moderator of negative affect in response to difficult children.

Autonomic and affective responses to children were assessed as a function of adult perceptions of interpersonal control. Women (N = 160) interacted with and provided feedback to computer-simulated children who "behaved" responsively or unresponsively on a computer game. Women were categorized as low in perceived control (PC) if they attributed high control to children but low control to self over negative events on the Parent Attribution Test. As predicted, low-PC women were maximally reactive to child characteristics, manifesting peak levels of defensive arousal (increased level of heart rate and electrodermal activity) and negative affect with unresponsive children and minimal levels of arousal and negative affect with responsive children. Intermediate response levels were shown by high-PC Ss. We interpreted results as suggesting mediating factors that may operate in dysfunctional interaction patterns previously found for low-PC caregivers.

It was predicted that affective responses of caregivers occur as an interactive function of child characteristics and adult attributions. Mothers in counseling at a child-abuse agency identified one of their children as relatively difficult (and more subject to discipline) and a second child as relatively easy; objective observations revealed that difficult children were more socially unresponsive and acted more inappropriately than did their easier siblings. Sibling pairs were videotaped interacting with unrelated mothers from the general community. Facial and vocal affect directed to difficult children was more dysphoric than that directed to their siblings-in particular among women who attributed relatively high control to children and low control to adults over unsuccessful caregiving. A similar relation between affect and attributions was found for related mothers. Adults with low-control attributions were interpreted as having a low threshold of affective reactivity.

Children between the ages of 5 and 10 years watched a videotape of a child having a routine medical exam. Embedded within the scenes were systematic variations of depicted facial affect shown by doctor and child. Measures were taken of autonomic reactions and information-processing errors in response to positive, neutral, and negative affective cues. For 5-6-year-olds, processing errors were greatest in the negative affect condition. Additionally, peak increases in heart rate (HR) and skin conductance level (SCL) were demonstrated by 5-6-year-olds in response to negative affect shown by the witnessed child; increases in HR were in turn predictive of processing errors. Older children (9-10 years) showed trends reflecting reduced processing errors in response to witnessed negative affect. It was suggested that younger children respond to salient threat cues with a "defensive" response pattern that is relatively adaptive at younger but not older ages.

Children between the ages of 5 and 10 years watched a videotape of a child having a routine medical exam. Embedded within the scenes were systematic variations of depicted facial affect shown by doctor and child. Measures were taken of autonomic reactions and information-processing errors in response to positive, neutral, and negative affective cues. For 5-6-year-olds, processing errors were greatest in the negative affect condition. Additionally, peak increases in heart rate (HR) and skin conductance level (SCL) were demonstrated by 5-6-year-olds in response to negative affect shown by the witnessed child; increases in HR were in turn predictive of processing errors. Older children (9-10 years) showed trends reflecting reduced processing errors in response to witnessed negative affect. It was suggested that younger children respond to salient threat cues with a "defensive" response pattern that is relatively adaptive at younger but not older ages.