غفلت مزمن و تجاوز / بزهکاری: بررسی طولی

Abstract Neglect is the most common form of maltreatment in the United States, yet its impact on development remains understudied, especially for chronic neglect. Chronic neglect is also one of the most costly burdens on child welfare systems. This study examines the effects of chronic neglect, including two subtypes (Failure to Provide and Lack of Supervision) on adolescent aggression and delinquency using a diverse longitudinal sample of youth. Chronic neglect and chronic failure to provide (ages 0–12) predicted aggression/delinquency (age 14) even after controlling for the effects of other maltreatment (ages 0–12). Chronic lack of supervision, however, did not. Gender significantly moderated these effects, suggesting that males are more likely to respond to neglect by becoming aggressive/delinquent. Finally, social problems (age 12) partially mediated for boys, and fully mediated for girls, the connections between chronic neglect and aggression/delinquency, bolstering theorizing that neglect impairs social functioning broadly. Implications include the need for further research on chronic neglect, especially in providing guidance for child welfare systems. Interventions for chronically neglected youth should include social skill development.

Results Prevalence of Chronic Maltreatment in Sample Sample descriptive for chronic maltreatment variables are shown in Table 1. As expected, neglect was the most common form of chronic maltreatment, and the overall average of chronicity is higher than other forms of maltreatment. Lack of Supervision (LOS) was slightly less common than Failure to Provide (FTP) neglect. The least common form of maltreatment was sexual abuse, with emotional maltreatment and physical abuse allegations comparable in terms of prevalence and chronicity. No child had reports of any single form of maltreatment in all six possible time periods. Also as expected, most youth had not experienced chronic maltreatment of any kind; the highest sample mean for any form was that of overall neglect, at 1.04. This indicated that most of the sample had experienced neglect in no more than one time period, although 30% had experienced neglect in 2 or more time periods. 20.9% of the sample had experienced FTP allegations in 2 or more time periods, versus 15.3% for LOS allegations. The percentage of the sample with allegations of other forms of maltreatment in two or more time periods were lower—13.1% for physical abuse, 3.7% for sexual abuse, and 13.1% for emotional abuse. Bivariate Correlations Bivariate associations, shown separately by gender, were generally significant and in the expected directions (see Table 2). All maltreatment variables were significantly associated for both boys and girls. All maltreatment variables were significantly associated with both aggression and delinquency for boys, however for girls chronic lack of supervision was not associated with aggression, and delinquency was only associated with chronic physical and sexual abuse. Social problems had mixed associations by gender as well. For males, social problems were not associated with chronic LOS or sexual abuse; for females, social problems were not associated with chronic FTP. Table 2. Bivariate associations amongst study variables by gender. 1 2 3 4 5 6 7 8 9 1. Chronic neglect 0.894 0.772 0.470 0.253 0.554 0.133 0.066a 0.098 2. Chronic FTP 0.867 0.540 0.368 0.170 0.484 0.132 0.074a 0.034a 3. Chronic LOS 0.821 0.552 0.451 0.243 0.508 0.090a 0.032a 0.103 4. Chronic physical abuse 0.504 0.421 0.475 0.361 0.506 0.197 0.184 0.172 5. Chronic sexual abuse 0.243 0.130 0.251 0.308 0.319 0.150 0.133 0.158 6. Chronic emotional abuse 0.547 0.470 0.480 0.618 0.279 0.149 0.062a 0.143 7. Aggression 0.243 0.263 0.154 0.239 0.170 0.159 0.726 0.503 8. Delinquency 0.215 0.210 0.169 0.227 0.163 0.157 0.761 0.376 9. Social problems 0.165 0.164 0.092a 0.194 0.093a 0.103 0.430 0.313 Note: FTP = failure to provide, LOS = lack of supervision. Males on bottom left, females on top right. All associations are significant at p ≤ .05 except where indicated by a. Table options Hierarchical Regressions Predicting Aggression and Maltreatment The hierarchical regressions shown in Table 3 through 5 test whether chronic neglect and its subtypes significantly predict aggression and delinquency, controlling for the contribution of other types of maltreatment. We initially examined whether the demographic variables—race and gender, controlling for study site—were significant predictors in a separate step; none were, so we omitted this step from the presentation of the data. Table 3 provides the results of the overall measure of neglect, which significantly predicted both aggression and delinquency controlling for study center and demographics. However, when other forms of maltreatment (chronic physical and sexual abuse, chronic emotional maltreatment) were added in step 2, chronic neglect did not remain a significant predictor of delinquency. The coefficient for aggression was also substantially diminished, suggesting shared variance (co-occurrence) with other forms of maltreatment. Table 3. Hierarchical regressions with chronic neglect predicting delinquency and aggression. Aggression Delinquency Step 1 Step 2 Step 1 Step 2 Demographics Blacka −0.059 −0.045 −0.058 −0.040 White −0.057 −0.059 −0.065 −0.067 Genderb −0.008 −0.010 −0.062 −0.063 Maltreatment Chronic neglect 0.135*** 0.087* 0.118*** 0.064 Chronic physical 0.128** 0.168*** Chronic sexual 0.084* 0.087* Chronic emotional −0.022 −0.043 ΔR2 0.051*** 0.020*** 0.036*** 0.028*** R2 0.051 0.071 0.036 0.064 F value 5.831*** 6.027*** 4.036*** 5.417*** Note: All coefficients are standardized betas. Results control for study centers. * p ≤ .05. ** p ≤ .01. *** p ≤ .001. a Referent is all other races. b Males = 1, Females = 2. Table options When examining the subtypes of neglect, chronic FTP was also significantly predictive of aggression and delinquency net of demographics, and this significance was retained controlling for other forms of maltreatment, although the coefficients again were reduced (Table 4). Chronic LOS, however, was not a significant predictor of either delinquency or aggression, even before controlling for other forms of maltreatment (Table 5). Based on this finding, further testing of moderation and mediation was not conducted for chronic LOS. It is important to note that the coefficients for chronic physical abuse were in every case the strongest predictors of aggression and delinquency. Table 4. Hierarchical regressions with chronic failure to provide (FTP) predicting delinquency and aggression. Aggression Delinquency Step 1 Step 2 Step 1 Step 2 Demographics Blacka −0.057 −0.045 −0.055 −0.040 White −0.055 −0.058 −0.063 −0.066 Genderb −0.014 −0.016 −0.067* −0.066* Maltreatment Chronic FTP 0.153*** 0.128*** 0.116** 0.086* Chronic physical 0.126** 0.167*** Chronic sexual 0.092** 0.093** Chronic emotional −0.032 −0.049 ΔR2 0.056*** 0.022*** 0.036** 0.031*** R2 0.056 0.066 0.036 0.067 F value 6.474*** 6.672*** 4.108*** 5.670*** Note: FTP = failure to provide. All coefficients are standardized betas. Step 1 is shared with regression results in Table 3. Results control for study centers. * p ≤ .05. ** p ≤ .01. *** p ≤ .001. a Referent is all other races. b Males = 1, Females = 2. Table options Table 5. Hierarchical regressions with chronic lack of supervision (LOS) predicting delinquency and. Aggression Delinquency Step 1 Step 2 Step 1 Step 2 Demographics Blacka −0.048 −0.036 −0.050 −0.034 White −0.056 −0.057 −0.063 −0.065 Genderb −0.007 −0.010 −0.060 −0.062 Maltreatment Chronic LOS 0.059 −0.010 0.068 0.000 Chronic physical 0.147*** 0.180*** Chronic sexual 0.085* 0.088* Chronic emotional 0.010 −0.023 ΔR2 0.040*** 0.026*** 0.029*** 0.032*** R2 0.040 0.066 0.029 0.062 F value 4.559*** 5.638*** 3.298*** 5.207*** Note: LOS = lack of supervision. All coefficients are standardized betas. Step 1 is shared with regression results in Table 3. Results control for study centers. * p ≤ .05. ** p ≤ .01. *** p ≤ .001. a Referent is all other races. b Males = 1, Females = 2. Table options Tests of the Moderating Effects of Gender Based on the divergent bivariate correlation results for boys and girls, the moderating effects of gender were tested using a multiplicative term in linear regression, controlling for study center and race (see Table 6). The findings were mixed. Gender significantly moderated the association between chronic neglect and delinquency, but not aggression. However, gender was a significant moderator of the effects of chronic FTP for both aggression and delinquency. A sample graph of these moderating effects is provided in Fig. 1 (Dawson, 2014). From this it is clear that the contribution of chronic neglect to aggression/delinquency is higher for males than for females, although there is a significant association for both genders. This pattern was seen for all significant moderating effects found. Table 6. Test of moderating effects of gender on chronic neglect and failure to provide (FTP). Aggression Delinquency Chronic neglect Gender 0.039 0.004 Chronic neglect 0.297** 0.343*** Chronic neglect * gender −0.176 −0.245* R2 0.054 0.041 F 5.491*** 4.174*** Chronic FTP Gender 0.041 −0.009* Chronic FTP 0.367*** 0.340** Chronic FTP * gender −0.232* −0.244* R2 0.061 0.042 F 6.292 4.221*** Note: FTP = failure to provide. Regressions control for center and race. * p ≤ .05. ** p ≤ .01. *** p ≤ .001. Table options Sample interaction plot of gender with chronic neglect predicting delinquency. Fig. 1. Sample interaction plot of gender with chronic neglect predicting delinquency. Figure options Tests of Mediation by Social Problems Finally, we tested whether social problems significantly mediated the relationship between chronic neglect and aggression/delinquency (see Table 7). These tests were conducted using path analysis in Mplus, using gender as a grouping variable, controlling for center and race, following the classic model proposed by Baron and Kenny (1986) using bootstrapping techniques (Shrout & Bolger, 2002). Table 7 provides the coefficients for the dependent variable (either aggression or delinquency) modeled on chronic neglect (or chronic FTP) directly, as well as mediated through social problems for boys and girls. Significance testing is not provided for difference based on group (boys versus girls). The fit statistics are also presented; these ranged from satisfactory to borderline. Our intention was not to test the model fit for an overarching theory, however, thus we interpret the model coefficients with caution. Table 7. Tests of social problems as a mediator using gender as a grouping variable. DV = Aggression DV = Delinquency Males Females Males Females Chronic neglect DV on chronic neglect 0.128* 0.041 0.118* 0.038 DV on social problems 0.397*** 0.0519*** 0.279*** 0.400*** Social problems on chronic neglect 0.157*** 0.107* 0.162*** 0.105* R2 of DV 23.3% 28.9% 15.0% 17.4% RMSEA 0.049 0.051 CFI 0.937 0.890 Chronic FTP DV on chronic FTP 0.163** 0.080 0.121* 0.072 DV on social problems 0.391*** 0.519*** 0.277*** 0.400*** Social problems on chronic FTP 0.159*** 0.051 0.160*** 0.042 R2 of DV 24.1% 29.1% 15.0% 17.7% RMSEA 0.053 0.056 CFI 0.929 0.870 Note: DV = dependent variable, FTP = failure to provide. All coefficients are standardized. * p ≤ .05. ** p ≤ .01. *** p ≤ .001. Table options For males, social problems was a significant predictor and mediator of the link between both overall neglect and FTP for both aggression and delinquency. The coefficients may be compared to those in Step 1 of Table 3 and Table 4; all are diminished, suggesting partial mediation (Fairchild & MacKinnon, 2009). For females, mediation was complete for all tests; the pathways between both chronic neglect and chronic FTP and aggression/delinquency were no longer significant when social problems was added to the model.