Social Cognition, Child Neglect, and Child Injury Risk: The Contribution of Maternal Social Information Processing to Maladaptive Injury Prevention Beliefs Within a High-Risk Sample

Abstract

Objective

Inadequate supervision has been linked to children’s injuries. Parental injury prevention beliefs may play a role in supervision, yet little theory has examined the origins of such beliefs. This study examined whether mothers who perpetrated child neglect, who as a group provide inadequate supervision, have more maladaptive beliefs. Then, it tested a social information processing (SIP) model for explaining these beliefs.

Methods

SIP and injury prevention beliefs were assessed in disadvantaged mothers of preschoolers (N  =  145), half with child neglect histories.

Results

The neglect group exhibited significantly more maladaptive injury prevention beliefs than comparisons. As predicted, SIP was linked to beliefs that may increase injury risk, even after accounting for relevant sociodemographic variables.

Conclusions

Findings support the link of beliefs to injury risk and suggest that specific cognitive problems may underlie these beliefs. Future work should further validate this model, which may inform enhancements to prevention efforts.

Unintentional injuries are a leading cause of child death and disability (CDC, 2013; Corso, Finkelstein, Miller, Fiebelkorn, & Zaloshnja, 2006). Approximately 10% of preschool-aged children experience a nonfatal unintentional injury resulting in emergency department treatment each year, and costs of nonfatal injuries in this age group alone total more than $11.4 billion (CDC, 2013). Rates of injuries not requiring medical attention are likely substantially higher. Children in low-income, single-parent families and those in disadvantaged neighborhoods are at especially high risk (Haynes, Reading, & Gale, 2003; Laflamme et al., 2010), and understanding risk within this group is crucial for prevention efforts. Parental supervision is one factor thought to be crucial to injury prevention (Morrongiello, 2005; Morrongiello & Schell, 2010; Saluja et al., 2004), especially for preschoolers, who are most often injured at home (Morrongiello, Midgett, & Shields, 2001). Inadequate supervision may constitute child neglect (Coohey, 2003); indeed, it is the most frequent reason for referrals (Nelson, Saunders, & Landsman, 1993), and is estimated as a factor in 44% of neglect cases (Ruiz-Casares, Trocmé, & Fallon, 2012). In one study of preschoolers in foster care, almost 90% had experienced supervisory neglect (Pears, Kim, & Fisher, 2008). Death record studies show inadequate supervision as a factor in almost half of maltreatment fatalities (44%; Damashek, Drass, & Bonner, 2014) and child injury fatalities (43%; Landen, Bauer, & Kohn, 2003).

Given its consequences (Coohey, 2003; Morrongiello & Schell, 2010; Peterson & Brown, 1994), research is needed to identify factors that increase risk for providing inadequate supervision. Parents’ beliefs are argued to reflect different supervision styles and, although supervision is determined by many factors, provide the basis for parents’ decisions about supervision across situations (Morrongiello & House, 2004; Morrongiello & Schell, 2010). Beliefs have been shown to be associated with parents’ behavior in naturalistic observations, as well as children’s injuries, suggesting beliefs may be a valid marker for supervision and injury risk (Morrongiello & Corbett, 2006). Given that instruments assessing beliefs are a low-cost way to assess risk, research linking existing measures to risk for lack of supervision is needed. One aim of this study is to do this. In addition, theory-based research that sheds light on the origins of such beliefs is needed to inform prevention efforts. This study sought to extend a cognitive theory, designed to explain parenting risk and shown to be useful in understanding risk for neglect, to childhood injury risk by examining links between social cognitive and neurocognitive problems and parental injury prevention beliefs.

Parental Social Information Processing and Injury Risk

Research on childhood injury is typically descriptive, with a relative paucity of theory-driven studies (Morrongiello & Corbett, 2008). Social cognitive models have been suggested as useful for understanding how parents act (or fail to act) to prevent injuries (Morrongiello & Corbett, 2008), and there is considerable overlap between antecedents of child injury and maltreatment (Peterson & Brown, 1994). Indeed, social information processing (SIP) models of parenting risk that have been used to explain risk for abuse (Azar, 1986; Azar, Reitz, & Goslin, 2008; Milner, 2003) are also being applied to neglect (Azar, Stevenson, & Johnson, 2012; Crittenden, 1993), and SIP theory is hypothesized as relevant for understanding parents’ injury prevention-related beliefs and behaviors (Azar & Weinzierl, 2005).

The SIP model of parenting posits that deficits in parents’ social cognitive and neurocognitive capacities increase risk for maladaptive parenting behaviors by changing the ways in which they attend to, take in, and act on information from the social environment. SIP capacities include schemas (e.g., expectations of children’s capacities), executive functioning (EF) and problem-solving abilities, and attributions (Azar, Robinson, Hekimian, & Twentyman, 1984; Azar et al., 2012; Milner, 2003). Greater SIP deficits have been shown to characterize parents who maltreat their child compared with control samples (Azar et al., 1984, 2012; Azar & Rohrbeck, 1986; Bugental & Happaney, 2004; Haskett, Smith Scott, Grant, Ward, & Robinson, 2003; Milner, 2003). Azar and Weinzierl (2005) argue that parental SIP deficits are also relevant to children’s injury risk, although less empirical work has examined this question. Here we provide a brief overview of how parental SIP capacities are hypothesized to link to children’s injury risk; for a more in-depth discussion, see Azar and Weinzierl (2005).

Parents’ schemas (or expectations) of children must be flexible and complex enough to adjust to their developmental changes (e.g., mobility, reach) and individual differences (e.g., risk-taking) (Azar & Weinzierl, 2005). Parents often overestimate children’s abilities (e.g., for self-care) and knowledge (e.g., of safety rules) (Morrongiello et al., 2001; Peterson, Mori, & Scissors, 1986; Schwebel & Bounds, 2003), and this is likely to be especially pronounced when parents view children as “little adults,” resulting in children being placed in risky situations more often. Indeed, beliefs that children learn from injuries are common and place the responsibility on the child to avoid future injuries; not surprisingly, parents are then less likely to act to minimize risk (Lewis, DiLillo, & Peterson, 2004; Morrongiello & Dayler, 1996; Simpson, Turnbull, Ardagh, & Richardson, 2009). Parents often believe childhood injuries are an inevitable part of development (Morrongiello & Dayler, 1996; Simpson et al., 2009; Whitehead & Owens, 2012) and may also have positive beliefs regarding the impact of being injured on children (e.g., children are toughened by injuries) that contribute to their lack of preventative actions (Lewis et al., 2004). Indeed, unrealistic expectations for children have been linked to both neglect and injury (Azar, 1986; Azar et al., 1984, 2012; Bugental & Happaney, 2004; Morrongiello & House, 2004).

Parents’ EF capacities provide a foundation for higher-order cognitive skills, such as problem-solving, and are essential to learning (Azar et al., 2012; Azar & Weinzierl, 2005; Barrett & Fleming, 2011). Good EF and problem-solving skills are needed to respond to changing environmental cues and child behaviors as children grow and develop. Parents with EF deficits are likely to have difficulty planning and organizing activities, anticipating and identifying safety risks, monitoring children and the environment, and adjusting their responses (e.g., vigilance level) accordingly, leading to inadequate supervision and greater injury risk (Azar & Weinzierl, 2005; Johnston, Mash, Miller, & Ninowski, 2012). For example, poor inhibition may contribute to difficulty maintaining attention to child cues in the face of competing, more rewarding stimuli (e.g., leaving an enjoyable TV show to follow a child into another room) (Johnston et al., 2012). Indeed, parents who have perpetrated maltreatment exhibit poorer problem-solving than comparison ones (Azar et al., 1984, 2012; Hansen, Pallotta, Tishelman, Conaway, & MacMillan, 1989; Milner, 2003).

Although less work has examined more specific EF capacities, there is some evidence that mothers who have maltreated children are less cognitively flexible (e.g., have difficulty shifting set, responding to feedback) than comparison mothers (Azar et al., 2012; Fontaine & Nolin, 2012; Nayak & Milner, 1998), and this predicts their parenting even after considering general intellectual functioning (Azar et al., 2012; Fontaine & Nolin, 2012). More research is needed to link parents’ EF with child injury risk; however, EF in older children has been linked to their attendance to safety cues and behavior in a potential risky situation (i.e., road crossing) (Barton & Morrongiello, 2011), supporting the hypothesis that parents’ EF contributes to the ability to attend to relevant environmental cues. Parents with poorer EF may fail to learn from their children’s injuries and, thus, may develop and maintain more maladaptive beliefs about supervision and the benefits of injuries.

Finally, negative and maladaptive appraisals of child behavior may result from unrealistic expectations and poor EF and problem-solving, further increasing injury risk (Azar & Weinzierl, 2005; Morrongiello & Corbett, 2008). Simplistic, unrealistic schemas are likely to be violated, and poor EF limits generation of response strategies and more benign appraisals, resulting in more hostile attributions, leading to harsher responses, as well as more self-blaming attributions, leading to disengagement from child care (Azar & Weinzierl, 2005). Hostile attributions have been associated with more lax discipline (Leung & Slep, 2006) and child neglect (Azar et al., 2012). Previous research has shown that parents who attribute injuries to external factors are less likely to take action (e.g., teaching, changing the environment) following an injury (Morrongiello & Corbett, 2008; Peterson, Bartelstone, Kern, & Gillies, 1995; Peterson & Brown, 1994), but the idea of a global negative bias toward the child has not received as much attention. Hostile attributions may be most relevant in disadvantaged contexts where environmental stress is high and parents’ own needs are not met, as attributing injuries to child characteristics allows parents to preserve parenting self-efficacy (Morrongiello & Corbett, 2008). Hostile attributions for risky behavior (e.g., “she knows better and is just acting that way to annoy me”) are less likely to lead to preventative action than more adaptive attributions (e.g., “she doesn’t realize how dangerous her behavior is”) and may lead parents to have beliefs that injuries are inevitable and due to fate or luck.

Current Study

Because parents’ supervision beliefs are seen as closely linked to supervision and some evidence supporting this contention is available (e.g., research examining maternal monitoring on the playground, actual child injuries; Morrongiello & Corbett, 2006; Morrongiello & House, 2004), examining SIP predictors of injury prevention beliefs seems worthy of consideration. The study used a sample of mothers of preschoolers who were sociodemographically disadvantaged and therefore are at high risk for unintentional injuries (Laflamme et al., 2010). Furthermore, half of the sample had a history of child neglect, indicating especially high risk for inadequate supervision. Hypotheses were as follows:

Hypothesis 1. Mothers who perpetrated neglect will have more maladaptive and less adaptive injury prevention beliefs than comparisons, further validating the importance of beliefs to risk for inadequate supervision.

Hypothesis 2. SIP deficits will be associated with more maladaptive and less adaptive injury prevention beliefs, after considering IQ and other demographic risk factors, supporting the validity of the SIP model for injury risk.

Methods

Participants

Low-socioeconomic status mothers of 3–5-year-old children (N  =  145) were recruited for a study of parenting from agencies serving disadvantaged families (e.g., Head Start, low-income day cares) and parents involved with child protective services (CPS). Recruitment took place in the poorest neighborhoods of Philadelphia, and the project was described to participants as a study of parenting. With mothers’ permission, all were screened for records of CPS involvement. The study sample includes 69 mothers with CPS histories for neglect and 76 comparison mothers with no CPS histories and no evidence of risk for physical abuse using the Brief Child Abuse Potential Inventory (Ondersma, Chaffin, Mullins, & LeBreton, 2005). This study had university institutional review board approval.

Procedure

Data collection was completed over three home visits of 2.5 hours each. Breaks were given as needed to prevent fatigue. In the first session, participants provided informed consent, consent for CPS record review, and demographic information. To control for literacy difficulties, consent forms and all instruments were read to mothers. Care was taken to ensure that participants understood the use of Likert scales. Participants received $150 for participation.

Measures

IQ

Four subscales of the Wechsler Adult Intelligence Scale (Arithmetic, Matrix Reasoning, Information, and Coding; Wechsler, 2008) were administered, providing a reliable estimate of IQ (r  =  .93; Sattler & Ryan, 2009).

Parent Opinion Questionnaire

The Parent Opinion Questionnaire (POQ) assesses unrealistic expectations for children (Azar et al., 1984). It has 80 agree–disagree items (e.g., “Usually, a 2-year-old can sit and play quietly alone in a room for several hours”). Higher scores indicate more unrealistic expectations. It has good test–retest reliability (r  =  .85; Haskett, Scott, Willoughby, Ahern, & Nears, 2006) and internal consistency (Cronbach’s α  =  .79), and has been shown to differentiate maltreating mothers from comparisons (Azar et al., 1984, 2012; Azar & Rohrbeck, 1986).

Parent Problem Solving Inventory

The Parent Problem Solving Inventory (PPSI) was used to assess problem-solving abilities in parenting situations. Ten typical childrearing problems were presented in story form (Wasik & Bryant, 1994). Participants were read the beginning of each story (e.g., a child begins to tantrum in a grocery store and the mother is embarrassed) and the story end, where the problem is resolved (e.g., the child is calm), and must provide the middle of the story (i.e., solution[s] to the problem). Responses were recorded by interviewers and scored later by independent raters. For this study, the number of different types of solutions provided across stories was used in analyses. Interrater agreement was 90%. In prior studies, the PPSI has distinguished maltreating from non-maltreating parents (Azar et al., 1984, 2012).

Wisconsin Card Sorting Test

The Wisconsin Card Sorting Test (WCST) is a neuropsychological test of “set-shifting,” that is, the ability to display flexibility in the face of changing stimuli and rules (Heaton, 1981). For this study, the number of perseverative errors (i.e., the number of times a participant continued to use a rule that was no longer correct) was used. More perseverative errors indicate less cognitive flexibility and more difficulty “set-shifting.”

Alternate Uses Test

Cognitive flexibility was also assessed using the Alternate Uses Test (AUT; Guilford, Christiansen, Merrifield, & Wilson, 1978). Participants were asked to produce as many uses as possible, other than the typical use, for six familiar objects within a set amount of time. This type of task assesses divergent thinking, and production of new responses is seen as reflecting underlying EF capacities (Gilhooly, Fioratou, Anthony, & Wynn, 2007). In this study, a higher number of unique and acceptable responses indicated greater flexibility. Interrater agreement for a randomly selected subsample (20 cases) was 93%.

Child Vignettes

The Child Vignettes (CV) assesses hostile attributions. Participants were read 18 vignettes of hypothetical aversive child behavior (Plotkin, 1983), asked to imagine that the child is their own, and rated how much they think the child did the behavior to annoy them on a 9-point scale from “not at all” to “very much.” Higher scores indicate more hostile attributions. Cronbach’s α was .87 in this sample. This measure has distinguished mothers with maltreatment histories from control mothers (Azar et al., 2012; Haskett et al., 2006; Plotkin, 1983).

Parent Supervision Attributes Profile Questionnaire

The Parent Supervision Attributes Profile Questionnaire (PSAPQ) is a 29-item questionnaire measuring supervision beliefs thought to underlie situation-specific supervision behaviors (Morrongiello & Corbett, 2006; Morrongiello & House, 2004). Participants were asked to think about being on a playground with their child and rated how often each item is true on a 5-point scale from “never” to “all of the time.” Factor analysis has identified four subscales. Items in the Protectiveness (e.g., “I warn him/her about things that could be dangerous”) and Supervision (e.g., “I have my child within arm’s reach at all times”) subscales assess beliefs about active supervision and harm reduction. Items in the Risk Tolerance subscale (e.g., “I let him/her learn from his/her own mishaps”) assess acceptance of children’s risk-taking behavior, and those in the Belief in Fate subscale (e.g., “When my child gets injured it is due to bad luck”) assess beliefs that children’s injuries are due to fate. The PSAPQ has good test–retest and internal reliability (Morrongiello & Corbett, 2006); Cronbach’s α was .76 in this sample. It has been linked to child injury and parents’ observed supervision (Morrongiello & House, 2004).

Injury Attitudes Questionnaire

The Injury Attitudes Questionnaire (IAQ) is a 30-item measure assessing perceived benefits of child injuries (Lewis et al., 2004). Participants indicated their agreement with each statement (e.g., “After being injured, my child usually learns not to do the same thing again”) on a 7-point scale from “very strongly disagree” to “very strongly agree.” Previous research has found that items load onto two factors—beliefs that children learn from injuries and beliefs that children are toughened by injuries (Lewis et al., 2004). The IAQ has good test–retest reliability and internal consistency (Lewis et al., 2004); Cronbach’s α was .83 in this sample.

Analyses

Given considerable theoretical overlap between supervision beliefs and injury beliefs, a factor analysis of the PSAPQ and IAQ subscales was conducted. The six subscales loaded onto two factors that were used in analyses (Table I). The first factor consisted of the risk tolerance and belief in fate PSAPQ subscales and the toughening and learning IAQ subscales, and was named “Maladaptive Injury Prevention Beliefs.” The second consisted of the Protectiveness and Supervision PSAPQ subscales, and was named “Adaptive Injury Prevention Beliefs.”

Results

Demographics

On average, participants had less than a high school education (M  =  11.49 years, SD  =  1.15) and almost all were single parents (89.0%) with incomes below the federal poverty line (90.3%). The sample was primarily African American (79.3%). They had an average of 2.83 children (SD  =  1.41), with 57% of the preschool children being female. Compared with the control group, the neglect group was slightly older (M  =  28.25 vs. M  =  26.19, t  = −2.32, p  =  .02), had lower IQs (M  =  77.86 vs. M  =  81.59, t  =  2.03, p  =  .04), and had more children (M  =  3.36 vs. M  =  2.36, t  = −4.59, p  <  .01). Mother age and IQ were controlled for in all analyses, as were child age and gender, which have been associated with parental supervision beliefs (Damashek, Borduin, & Ronis, 2014; Morrongiello & Dawber, 1999; Morrongiello, Walpole, & McArthur, 2009).

Hypothesis 1: Injury prevention beliefs of the neglect group mothers were compared with those of comparison mothers using analysis of covariance, with mother age and IQ, and child age and gender as covariates (Table II). The neglect group endorsed significantly greater beliefs in fate as the cause of injuries and showed a trend toward stronger beliefs that injuries toughen children. They also had significantly higher scores than comparison mothers on the overall maladaptive beliefs factor score, supporting the link of maladaptive injury prevention beliefs to parenting risk. No significant group differences were found in adaptive injury prevention beliefs.

Hypothesis 2: Two regressions were run to test the contributions of SIP deficits to maladaptive and adaptive injury prevention belief factors. The first examined maladaptive injury prevention beliefs, with mother age and IQ, child age and gender, and neglect status entered in the first block and SIP variables in the second block (Table III). SIP variables significantly improved model fit, and neglect status was no longer a significant predictor when considered along with SIP variables. In the final model, mother age and IQ, child gender, unrealistic expectations, hostile attributions, and perseverative errors were significant individual predictors. Along with control variables, SIP factors accounted for approximately 31% of the variance in maladaptive injury prevention beliefs. The second regression examined adaptive injury prevention beliefs. Covariates were entered in the first block, and SIP variables were entered in the second block. After accounting for covariates, there was a trend indicating that SIP variables as a group added some predictive power (R² = 11%, F(10, 134)  =  1.71, p  = .09), although no control or SIP variable alone contributed significantly (full results available from authors).

Discussion

Childhood injury risk is clearly multidetermined by an array of contextual, parental, and child factors. This study focused on risk indicators for parental inadequate supervision. It had two goals. First, it examined the idea that supervision beliefs may be a valuable alternative to costly observational approaches (Morrongiello & House, 2004; Morrongiello & Schell, 2010), using two belief instruments. As predicted, neglect group mothers reported more maladaptive injury prevention beliefs. The second goal of this study was to test whether social cognitive and neurocognitive factors are linked to injury prevention beliefs. SIP deficits, specifically unrealistic expectations, EF, and hostile attributions were strongly linked to maladaptive injury prevention beliefs, over and above maternal IQ and other demographic risk factors (e.g., child gender, age). Adaptive injury prevention beliefs, however, did not differ between groups, and there was only limited support for links between SIP deficits and adaptive beliefs.

As beliefs have mostly been studied in populations at lower risk for unintentional injuries, finding neglect/non-neglect group differences in this disadvantaged sample is notable given the high sociodemographic risk in the comparison group. This finding suggests markers may exist within populations most at risk and allow targeted prevention. Preventing unintentional childhood injury requires well-validated etiological models (Azar et al., 2008; Morrongiello & Schwebel, 2008; Peterson & Brown, 1994; Slack et al., 2011). Injury research has already identified social cognitive problems as playing a role in injury (e.g., attributions, expectations; Morrongiello & Corbett, 2008; Peterson & Brown, 1994). Work in maltreatment has also focused in on a specific set of parental social cognitive and neurocognitive determinants, including expectations of children, problem-solving and EF, and attributions, and documented a similar role in parenting risk for neglect (Azar et al., 2012). The findings of this study extend the validity of this global model of parenting risk and elaborate on parental cognitive factors that may negatively affect parental decision-making and actions, and potentially increase the occurrence of childhood injuries. This builds on injury-specific models posited by others (Morrongiello & Corbett, 2008; Peterson & Brown, 1994) that argued for the importance of expectancies and attributions, and also enhances these models by including more basic neurocognitive deficits that affect learning and judgments in the moment.

Findings should be considered with some caution, as direct measures of supervision and childhood injuries were not used. Prospective longitudinal research linking SIP directly to supervision behaviors and injuries would strengthen interpretations. Our sample was entirely disadvantaged urban mothers with children in a particular age group (3–5-year-olds), and along with collecting injury data, a larger sample that is more diverse in social class, context, and child age is needed to assess the broader worth of this risk model and the validity of injury prevention beliefs. For example, adaptive injury prevention beliefs may be more relevant to injury risk in lower-risk populations. Validation of the IAQ with regard to injury is especially needed.

Although we chose in this study to focus on parental determinants of injury risk, parental supervision is just one of many factors that explain childhood injury, and future work should explore how SIP factors may interact with child and contextual factors to affect injury risk. For example, parental SIP problems may be transmitted across generations through both genetic and socialization mechanisms (Friedman et al., 2008; MacBrayer, Milich, & Hundley, 2003), and children’s own SIP deficits may increase their risk for injuries (Barton & Morrongiello, 2011). Environmental hazards are associated with parents’ supervision beliefs (Damashek, Borduin, et al., 2014), and SIP factors may affect how parents perceive or respond to hazards in the environment. Future studies might consider more complex pathways and interactions among parent, child, and contextual factors.

Cognitive techniques have been successfully used as adjuncts to behavioral skills training to reduce maltreatment, and following more validation work, the addition of cognitive strategies targeting specific beliefs (e.g., distorted views on the value of injuries and their attributions to fate) may be useful adjuncts to behavioral injury prevention interventions (Edwards & Lutzker, 2008). Specific efforts to change parents’ attributions regarding child negative intent and cognitive restructuring of expectations and schema about injuries may improve parental motivation and increase the likelihood of behavioral changes (Morrongiello & Corbett, 2008). In addition, given the role of EF in monitoring (Azar & Weinzierl, 2005; Johnston et al., 2012), efforts targeting components of EF (e.g., attentional processes, working memory, set shifting) are particularly worth of exploration.

Acknowledgments

The authors would like to thank the Philadelphia Department of Human Services and the community agencies that made this research possible. We would also like to thank the editor and reviewers of our manuscript for their thoughtful comments.

Funding

This work was supported by funding from the National Institute of Child Health and Human Development (#5R01HD53713), the PSU Child, Youth and Family Consortium and Eunice Kennedy Shriver National Institute of Child Health and Human Development.

Conflicts of interest: None declared.

References