Interactive association of maternal education and peer relationship with oppositional defiant disorder: an observational study

About one in five children and youths met criteria for a mental disorder but less than one-third had received a mental health professional service in a Canadian study [1]. To strengthen prevention and early intervention efforts by depicting familial and peer influences are essential to meet the mental health needs of youths. Of the mental disorders in youths, oppositional defiant disorder (ODD) represents enduring patterns of argumentative, disobedient attitude toward authority figures, characterized with irritable mood and vindictiveness. ODD is one of the most commonly encountered clinical disorders in children and adolescents. The clinician is most often alerted when problems with oppositionality, vindictiveness, negativistic and hostile behavior which create a significant disturbance in social, academic, or occupational functioning [2]. Data from Avon Longitudinal Study of Parents and Children have shown individuals with ODD were likely to develop anxiety, depression or conduct disorder in the long term follow-up [3]. The Cambridge study on delinquency also found that ODD persons had higher rate of hospitalization or being registered disabled in their middle age [4]. As ODD is among the most common reasons referred to mental health professionals, primary care physicians are positioned to assess youth at risk of ODD and refer their families to community programs [5].

Both familial and peer factors had been found to associate with the ODD. Familial factors, such as parent emotional regulation and cohesion, are related to child ODD symptoms and understanding child behavioral problems within the family context is important [6]. Harsh inconsistent discipline, low warmth and involvement, and high criticism parenting styles were considered risky for the ODD [7]. The parenting style might correlate with children’s externalizing behaviors [8]. Parents’ own antisocial behavior also played a role on the development of children’s oppositional problems [9]. During childhood and adolescence, the oppositional child may be rejected by non-deviant peers and tends to attract other deviant ones. The affiliation with deviant peers were more likely to reinforce mutual antisocial behaviors [10]. The poor social skills and lack of prosocial behaviors were common among children with ODD [11]. These familial and peer factors were found to influence children’s defiance through parenting practices and social skill learning at peer interactions. Also, social learning theories have shown that individual who witnesses more deviant behaviors from their parents or peers would model the advantage of these behaviors, thus promoting aggression and oppositional behaviors [12].

A total of 981 subjects between 18 and 25 years of age were enrolled from 14 schools. Social demographic characteristics associated with ODD are shown in Table 1. In the univariate analyses, male (OR = 2.43), study in night division providing classes in the evening (OR = 2.12), living in a single-parent family (OR = 1.75), poor academic performance (OR = 1.95), high personal income (OR = 2.25), going to high risk environment of illicit substance use (OR = 5.61), and peer with conduct behavior (OR = 1.77,) or illicit substance use (OR = 2.78) were significantly correlated with the possible case of ODD. Comparing to the low educational level, higher maternal education was less likely to develop ODD (OR = 0.64).

Table 1also shows the adjusted odds ratio of demographic, familial and peer factors on the risk of ODD. Six variables remained significant at the multivariate logistic regression, including male, night division, poor academic performance, high risk environment, peer with illicit substance use and high maternal education level. High maternal education persisted to exert independent protective effect on the development of ODD (aOR = 0.65).

At a further inspection, we stratified the familial factor, maternal educational level, to analyze the differential impact of peer factors on the ODD. In the Table 2, two factors exerted distinct impact on the presence of ODD. Peer with illicit substance use was more likely to associate with ODD in the low maternal education group, but study in the night division made the significant effect in the high level counterpart. We in particular calculated the 2-way interactive effect of maternal education and peer with illicit substance use on the development of ODD, and O.R. = 4.96 (2.96, 8.31) was found.

Our study investigated the associated risk factors for ODD and 6 factors were found to exert independent association, including gender, campus problems, peer interaction and familial variables. We specifically inspected the interactive impact of peer and familial factors on the presence of ODD. The high maternal education level provided a significant protective effect on the male college student at facing the substance enabler in peers.

The current study showed young adult male (age 18–25) was more likely to have ODD than female, which is consistent with previous epidemiology surveys. The gender is constantly a significant predictor for oppositional and further antisocial behavior [21]. However, the prevalence rate of ODD in boys outnumber girls before adolescent, the male predominance existing but declining among adolescent and adult samples. This sex difference of ODD is possibly affected by sociocultural factors, apparent in Western society but not so significant for non-Western culture [22].

As to the influencing factors in family, it is suggested that highly educated mother may use the more sensitive parenting style instead of controlling way [23, 24]. These mothers are likely to know about the importance of keeping consistent and warm parenting style, which thus reduces the externalizing problems in youths [7, 25]. Moreover, the better maternal working memory or cognitive control of emotion along with decreased harsh reactive parenting was found to lessen the oppositional reactivity at interaction with children [26]. Our finding that mother with higher educational level was a protective factor for the development of ODD is compatible with those findings.

In contrast, low educated mothers were more neglectful and this parenting style itself increased the risk of developing ODD in children [27]. Our survey showed that the concurrence of low maternal education level and peer with illicit substance use exerted the highest risk for having the ODD. It is possible that the low educated mother uses dysfunctional parenting style and is less likely to monitor her child’s peer relationship and whereabouts. As the child, particularly boy, had delinquent peers around, he was at risk to increase delinquent behavior and developing oppositional behavior. The illicit substance use in adolescents may predict future conduct problems [28]. Youths with deviant peers would increase their rates of antisocial behaviors after joining a gang, which risk reduces if leaving the gang [29]. The maternal support had not only direct effect on reducing the delinquency but also acting indirectly on deviant peer affiliation [30].

The exposure to high risk environment of illicit substance use persisted highly associated with the ODD despite the protective effect of high maternal education level. A student studying in the night division was more likely to have ODD when mother well educated than with a low educated mother. Studying in night division might be the result of low academic achievement at senior high school or during university entrance exam, particularly for students in higher socioeconomic family. The previous study has showed that the low IQ and low school achievement are important predictors of ODD [31]. Students with low performance may encounter unrewarding experiences through the school lives and turn out to seek support from deviant peers. As deviant youths get together, they possibly reinforce mutual delinquent behaviors [10]. The interaction also increases the risk of developing ODD. At another direction, ODD is commonly a disorder of childhood. Its symptoms significantly interfere with social or interpersonal relationships at primary and middle schools. The persistence of impairment associated with ODD into young adulthood may make the patient enter a night division of college despite the socioeconomic advantage. This phenomenon also calls for a reconsideration of ODD as a disorder limited to childhood [32].

Despite those intriguing findings, our survey is limited by its cross-sectional design and temporal relationship among risk factors and ODD cannot be well defined. Therefore, we could not confirm the causal direction of risk factors and the disorder. Secondly, the definition of ODD is according to the self –administered questionnaire instead of interview, thus not a formal diagnosis. Though the questionnaire is made from diagnostic criteria, potential bias might come from the different interpretation of the symptom questions between participants and experts. However, the present assessment of ODD symptoms offered the advantage of variability in a non-clinical sample. Moreover, the sample was all from college students, not the community population. The selection bias exists and generalizability is restricted. We made every possible effort to address the potential sources of bias. First, we recruited participants from both general and vocational colleges and universities, from both public and private schools, from urban and rural areas, from day and night divisions, and from more than half counties nationwide. Second, we provide leaflets, invitation letter, and oral explanation to increase the understanding of self-report questionnaire for participants. We also address that all the answers were processed anonymously and aimed for medical or educational concern not for any legal concern or persecution purpose to increase the truthfulness of the self-reports. Third, we handled the data carefully and rechecked with expert penal from time to time to verify the findings comparable to clinical impressions and campus real world conditions. Forth, we used multivariate regression and calculate the receiver operating characteristic curve (ROC curve) to see the sensitivity and specificity of our model. The interaction ROC curve were 0.79, indicating the acceptable discrimination of our model.

This study found that family and peer factors interacted to exert the highest effect for developing ODD. It gives implications of importance at developing relevant programs to prevent college students from poor long-term outcome. In addition to enhance supportive parenting skills, positive campus experiences might help the student resisting the risky peer influence [33, 34]. The concern of the generalizability (external validity) is a construct that attempts to answer whether we can use the results of the current study in college students other than those enrolled in the study. The participants’ profiles, such as demographics and social-economic status (SES), were similar to general college students in Taiwan. However, it needs cautions to penalize the findings to other counties with significantly non-comparable SES or without similar healthcare and educational resources. In order to increase the generalizability, further study may need more rigorous design, such as collect data from the general population, enroll ethnical minorities, conduct an international collaborative study, perform longitudinal trend analysis to increase the external validity.