Risk-Taking and Delinquent Behaviors Among Youth with and without Intellectual Disabilities

Abstract

Introduction:

Youth with intellectual disabilities (ID) demonstrate higher rates of disruptive behavior disorders (DBDs) than youth with typical development (TD). DBDs such as oppositional defiant disorder (ODD) predict higher rates of delinquency during adolescence. Yet, few studies have examined risk-taking and delinquency among youth with ID.

Methods:

We used a self-report measure to determine whether 13-year-old youth with ID (n= 23) reported higher rates of risk-taking and delinquent behavior than their TD peers (n=77). We also examined whether or not youth had a previous diagnosis of ODD.

Results:

Our results suggest that youth with ID reported fewer rule-breaking and risk-taking behaviors than their TD peers. In contrast, youth with a previous diagnosis of ODD reported more of these behaviors.

Conclusion:

Our results appear discrepant from previous studies, which find higher rates of risk-taking and delinquency among youth with ID. As such, we discuss the factors that may explain our discrepant results, including our definition and assessment of ID, and the age of our participants.

In the process of developing autonomy from their parents and establishing a separate identity for themselves, adolescents often engage in some degree of rule breaking, experimentation with substances, and other risky behaviors (Kann, et al., 2016; Shedler & Block, 1990; Young et al., 2002). Although not uncommon, adolescents who engage in these behaviors are at risk for a variety of negative outcomes, ranging from fines to serious consequences such as incarceration or death (Miller, Malone, Dodge & Conduct Problems Prevention Research Group, 2010; Moffitt, Caspi, Rutter & Silva, 2001; Odgers et al., 2008). Accordingly, engaging in risk-taking and/or delinquent behavior during adolescence is a concern for law enforcers, researchers, and parents.

Within the literature, risk-taking and delinquency represent two theoretically distinct concepts. The former captures a variety of potentially harmful behaviors that are not uncommon to adolescence and may be socially accepted to some degree (Igra & Irwin, 1996); for example, drinking alcohol, experimenting with various drugs, and engaging in unprotected sex. In contrast, delinquency is defined as “acts, the detection of which is thought to result in punishment of the person committing them by agents of the larger society” (Hirschi, 2002) or “conduct that is out of accord with accepted behavior or the law” (Merriam-Webster, 2019). Thus, while often used in a legal context to identify those who have committed a crime, delinquency as a psychological term manifests a continuum of rule-breaking and antisocial behaviors. While theoretically distinct, there appears to be some overlap between risk-taking and delinquency, and some behaviors (e.g. driving under the influence) fall into both categories.

Risk-taking Behavior in Adolescence

Research supports a social neuroscience perspective on risk-taking in adolescence, in which behaviors such as experimentation with substances are seen as a function of changes to the brain’s dopaminergic system, resulting in greater reward-seeking behaviors particularly in the presence of peers. As the adolescent becomes an adult, changes in the cognitive control system allow for greater self-regulation, resulting in a decline in risk-taking behaviors (for a review, see: Casey, Getz & Galvan, 2007; Steinberg, 2008; Steinberg, 2010). This process is understood as part of normative adolescent brain development; however, previous research also finds individual differences in neural activation, suggesting that some adolescents may be more prone to risk-taking than others (Galvan, Hare, Voss, Glover & Casey, 2007).

Delinquent Behaviors in Adolescence

While some adolescents experiment with occasional antisocial or delinquent behaviors, others may develop diagnosable conduct disorder (CD) or be declared a juvenile delinquent. As defined in the Diagnostic and Statistical Manual of Mental Disorders – Fifth Edition (DSM V; APA, 2013, page 469), a person must engage in at least 3 of 15 specific antisocial behaviors within a year to be diagnosed with CD. A juvenile delinquent (often considered a legal term), on the other hand, may be labeled as such based on one illegal act.

Research on antisocial behaviors in adolescence finds support for at least two trajectories, particularly among male offenders (Moffitt, 2018; Jennings, Rocque, Hahn-Fox, Piquero & Farrington, 2016). One group, termed adolescence-limited, appear to engage in antisocial behaviors during adolescence, but desist once they reach adulthood. The other group, termed life-course-persistent, demonstrate conduct problems in early childhood and, as adolescents, continue to engage in antisocial behaviors throughout their lifetime. Among a number of characteristics, early conduct problems, neuropsychological deficits and harsh, inconsistent or negligent parenting seem to differentiate these two groups (for a review, see Moffitt, 2018). However, adolescents in either group may be diagnosed with Conduct Disorder or declared a juvenile delinquent based on the criteria defined above.

Age of Onset for Risk-taking & Delinquency

There is some indication that pubertal onset is associated with increased risk-taking and delinquent behavior (Costello, Sung, Worthman & Angold, 2007; Downing & Bellis, 2009). For example, one study found that pubertal onset, and specifically changes in testosterone, may predict brain activation in response to rewards – thus, driving the increase in risk-taking seen in adolescence (Braams, van Duijevnvoorde, Peper & Crone, 2015). Accordingly, there is concern that as the average age of pubertal onset has decreased, so has the age of onset for behaviors such as drinking alcohol, experimenting with substances, and engaging in sexual contact (Bellis, Downing & Ashton, 2006). Thus, while there is ample variation in pubertal onset, research suggests an overall trend towards initiating risk-taking and delinquent behavior in early adolescence – making this a crucial period of study.

Risk-Taking & Delinquency for Children with Developmental Delays

Researchers have found considerable support for differences in cognitive functioning between children with and without significant conduct problems; youth who commit antisocial and delinquent behaviors can have mean IQs a full standard deviation below average (Fergusson, Horwood & Ridder, 2005; Ge, Donnellan & Wenk, 2001; Kratzer & Hodgins, 1999; Lynam, Moffitt & Stouthamer-Loeber, 1993). Moreover, investigators have identified executive functioning and verbal reasoning as two areas of interest, and deficits in these domains may characterize the neuropsychological profile of early-onset antisocial behavior (Lynam & Henry, 2001; Morgan & Lilienfeld, 2000; Nordvall, Jonsson & Neely, 2017; Sorge, Skilling & Toplak, 2015).

Considerably less research has focused on risk-taking/delinquent behavior among individuals with intellectual disabilities (ID), although there is some indication that a significant portion of the prison population may have ID (estimates range between 7–10%; Hayes, Shackell, Mottram & Lancaster, 2007; Hellenbach, Karatzias & Brown, 2017; Murphy, Gardner & Freeman, 2017; Young et al., 2018). There is also some research suggesting that adolescents with intellectual disabilities demonstrate higher rates of risk-taking and antisocial behavior than peers with typical development (Dickson, Emerson & Hatton, 2005; Emerson & Halpin, 2013; Emerson & Turnbull, 2005; Frison, Wallander & Browne, 1998). However, it is unclear if this effect is driven by other factors that are often related to antisocial behavior (e.g. household and neighborhood disadvantage; Emerson & Halpin, 2013; Emerson & Turnbull, 2005), and/or if adolescents with intellectual disabilities are more likely to be caught than their typically developing peers.

Dickson et al. (2005) used interviews to measure self-reported antisocial behavior among 98 11–15-year-olds with ID (and a comparison sample of typically developing adolescents). Findings illustrated the higher rate of antisocial behavior (specifically, bullied/threatened others, stolen valuable things, used weapons, started a fire, destroyed property, stolen in the street, trouble with police) in this population relative to typically developing peers. Moreover, the authors found that antisocial behavior was associated with male gender, older age, lower levels of maternal education, living in poverty, family dysfunction, and having a comorbid mental health disorder. Of note, Dickson et al. examined 13 specific antisocial behaviors such as those described above. Moreover, the authors used a fairly inclusive definition of ID, including adolescents with ‘learning difficulties’ whose parents had concerns about their language development.

Emerson and Halpin (2013) used data from the Longitudinal Study of Young People in England (LSYPE) and the English Department for Education Data on Special Education Needs (SEN) to identify a sample of children with mild or moderate intellectual disability (based on their classification as Moderate Learning Difficulty within the SEN). The authors conducted interviews with the adolescent participants (ages 13–15; total sample – 12,907; 532 with mild or moderate ID), and their parents. They found that parents of adolescents with mild or moderate intellectual disability reported more frequent contact with police, and adolescents self-reported more fighting, shoplifting and graffiti than their peers. However, consistent with other studies, the authors found that these effects were largely mediated by other factors (e.g. male gender, single-parent household, household and neighborhood disadvantage). Moreover, the authors acknowledged that their classification method for intellectual disability was based on educational status, and not a direct assessment of intellectual functioning.

Emerson & Turnbull (2005) explored the prevalence of smoking and alcohol use among 95 adolescents (ages 11–15) with intellectual disabilities, and 4069 adolescents with typical development. Findings suggested that adolescent with ID self-reported higher rates of smoking relative to their peers, and lower rates of monthly alcohol use. However, the increased rate of smoking appeared to be an artifact of the higher rates of poverty among the study’s population of adolescents with ID. Again, the authors noted that their classification of ID was based on parent-reported “learning difficulties” and/or school placement (in a setting for children with “learning difficulties”). Finally, Bexkens, Van der Molen, Collot d’Escury-Koenigs & Huizenga (2014b) and Bexkens, Ruzzano, Collot d’Escury-Koenigs, Van der Molen & Huizenga (2014a) found evidence for cognitive control deficits among adolescents with intellectual disabilities, positing that this creates a larger discrepancy between reward sensitivity and cognitive control, and contributes to increased risk-taking among adolescents with ID.

Risk-Taking for Children with ID and Comorbid Diagnoses

It is well documented that children with ID demonstrate higher rates of externalizing behavior problems and disruptive behavior disorders (DBDs) than youth with typical development (Baker, Blacher, Crnic & Edelbrock, 2002; Baker, Neece, Fenning, Crnic & Blacher, 2010; Einfeld & Tonge, 1996; Whitaker & Read, 2006). Moreover, there appears to be significant overlap between externalizing behavior problems, disruptive behavior disorders, and delinquency. Children and adolescents with externalizing behavior problems appear to engage in more antisocial behaviors, and there is considerable research suggesting that aggression and noncompliance during childhood is a significant predictor of delinquency in adolescence (Caspi & Moffit, 1995; Hay, Meldrum, Widdowson & Piquero, 2017; Lillehoj, Trudeau, Spoth & Mason, 2005; Loeber, 1982; Piquero, Carriaga, Diamond, Kazemian & Farrington, 2012; Sampson & Laub, 1993; Skinner, et al., 2017). Accordingly, it is important to consider the role of comorbid DBDs during childhood in predicting risk-taking and delinquent behaviors in adolescence, particularly among individuals with ID.

Oppositional defiant disorder (ODD) is a disruptive behavior disorder and a common form of comorbid psychopathology for children with intellectual disabilities (Baker et al., 2010; Christensen & Baker, 2013; Dekker & Koot, 2003; Emerson & Hatton, 2007). The DSM-V has defined ODD as “a pattern of angry/irritable mood, argumentative/defiant behavior, or vindictiveness lasting at least six months as evidenced by four symptoms from any of three categories, and exhibited during interactions with at least one individual who is not a sibling” (DSM V; APA, 2013, page 462). Beyond the intuitive connection between defiance and delinquency, ODD is often viewed as a developmental precursor to conduct disorder. Research suggests that many adolescents with conduct disorder met criteria for ODD previously, although more recent studies find that this is predominantly true for males (Burke, Loeber & Birmaher, 2002; Lahey, et al., 1994; Loeber, Keenan, Lahey & Green, 1993; Rowe, Costello, Angold, Copeland & Maughan, 2010). In this light, it is important to consider what risk-taking and delinquent behaviors children with ODD (and particularly those with comorbid ID) might engage in once they reach adolescence.

A recent study by Bexkens et al. (2018) examined risk-taking among adolescents with mild to borderline intellectual disabilities (MBID) and/or behavior disorders (BD; specifically, ODD, CD, Attention-Deficit/Hyperactivity Disorder and/or Disruptive Behavior Disorder NOS). The authors used the Balloon Analogue Risk-Task (BART) to examine the behavior of adolescents in one of four groups (MBID only, BD only, MBID + BD, typically developing), with or without the presence of unknown peers. Findings suggest that adolescents with MBID engaged in greater risk-taking only in the presence of peers; comorbid BD did not appear to increase the propensity for risk-taking. Of note, however, this study focused only on risk-taking, and did not include measures of antisocial or delinquent behavior.

The Current Study

The present study addressed three questions. On a self-report measure of risk-taking and delinquency: (1) Do adolescents with intellectual disability (ID) differ from adolescents with typical cognitive development? (2) Do adolescents with a childhood diagnosis of oppositional defiant disorder (ODD) differ from typically developing adolescents? (3) Does having comorbid diagnoses of ID and ODD increase the rate of risk-taking and delinquency relative to either ID or ODD alone?

This study expands on previous research in several important ways. First, the current study defined ID (IQ < 70 and corresponding adaptive behavior deficits; borderline intellectual functioning – IQ <85 and corresponding adaptive behavior deficits) based on standardized instruments (e.g. Stanford-Binet IV and Vineland Adaptive Behavior Scales II), rather than classifying participants based on parent report of “learning difficulties” or classroom placement. Second, we examined a wide range of both delinquent and risk-taking behaviors, thereby simultaneously exploring these distinct yet overlapping concepts. Third, in addition to traditionally delinquent or antisocial behaviors, such as truancy and theft, we examined typical rule-breaking behaviors to see if these also differ by status group and comorbid ODD. Finally, as the data for this study were drawn from a larger longitudinal investigation of children with and without intellectual disabilities, we explored the relationship between an ODD diagnosis in childhood and risk-taking/delinquent behaviors in early adolescence. In so doing, we examined how two early risk factors (intellectual functioning and comorbid ODD) impact psychosocial functioning in adolescence.

Method

Participants

Participants were selected from 238 families recruited to participate in a longitudinal study (“The Collaborative Family Study”) conducted at the University of California, Los Angeles (UCLA), the University of California, Riverside (UCR), and Pennsylvania State University. Families of typically developing children as well as families of children with developmental delays residing in the Southern California and Central Pennsylvania areas were recruited through community agencies serving persons with developmental disabilities, and the corresponding local preschools and daycare centers. Selection criteria for enrollment for typically developing (TD) children included (a) an enrollment age between 30 and 39 months, (b) a score of 85 or above on the Bayley Scales of Infant Development – II (BSID II; Bayley, 1969) at the intake assessment, and (c) no history of prematurity or diagnosis of a developmental disability. Criteria for children with developmental delays (DD) included (a) an enrollment age between 30 and 39 months, (b) a score between 30 and 84 on the BSID II, (c) being ambulatory and (d) not diagnosed with autism.

Participants and their families completed a battery of questionnaires and lab tasks at child ages 3, 4, 5, 6, 7, 8, 9 and 13. As is expected in longitudinal research, there was some attrition over time (Gustavson, von Soest, Karevold & Roysamb, 2012). The data reported in this study are from 100 of the original 238 families (TD n= 77; DD n=23) who met specific inclusion criteria (i.e. data from age 5 and age 13, as well as at least 2 of the 4 assessment points in between).

Although the BSID-II was used to recruit participants with and without developmental delays, scores from the BSID-II have been shown to have limited validity in predicting later cognitive abilities, particularly among high-risk populations (e.g. low-birth-weight infants; Hack et al., 2005; Luttikhuizen dos Santos, de Kievet, Konigs, van Elburg & Oosterlaan, 2013). As such, children were classified based on their scores on the Stanford-Binet IV and Vineland Adaptive Behavior Scales II (VABS) at child age 5, the first year considered in the present study. Participants were considered to have an intellectual disability if they received an IQ of 70 or below on the Stanford-Binet IV and a score of 85 or below on the Vineland Scales of Adaptive Behavior II (VABS). Participants were considered to have borderline intellectual functioning if they scored between 71 and 84 on the Stanford-Binet IV and had scores on the VABS below 85. Due to the small sample size, we combined the children with borderline intellectual functioning and those with intellectual disabilities into one group (termed Developmental Delays, DD). Combining these groups is supported by prior research suggesting that children with borderline intellectual functioning experience rates of emotional and behavior problems that are similar to children with ID (Fenning, Baker, Baker & Crnic, 2007; Emerson, Einfeld & Stancliffe, 2011). Finally, the inclusion criterion for the typically developing group was an IQ on the Stanford-Binet IV above 85.

The children and their families were then followed through age 13. The Wechsler Intelligence Scale for Children (WISC-IV) and Vineland Adaptive Behavior Scales II were administered at ages 9 and 13 to confirm status grouping.¹

The center visit at age 13 (the focus of the present study), involved both the adolescent and his/her participating caregiver (typically, the mother). At the start of the assessment, two graduate student examiners met with the dyad to complete informed consent. Limits of confidentiality were discussed, and both the adolescent and his/her caregiver were informed that only information pertaining to child abuse, elder/dependent adult abuse, harm to others or harm to self would be shared, if necessary, to maintain safety. Likewise, the dyad was informed that the caregiver would not have access to the adolescent’s responses, except under the aforementioned conditions. The dyad then completed two joint activities – a discussion of a mutually-identified conflict between them, and a problem-solving task, each lasting approximately 5 minutes. The dyad then separated. The caregiver completed the Family Information Form (a measure of basic demographics), a semi-structured interview regarding their child’s friendships, sibling relationships, perception of school, and experiences of bullying, the Vineland Adaptive Behavior Scales II, and the Diagnostic Interview Schedule for Children IV. Meanwhile, the adolescent completed the Wechsler Intelligence Scale for Children –IV, neuropsychological tests of spatial memory, response inhibition, and set-shifting, a 5-minute self-regulation task, an interview about their friendships, sibling relationships, perception of school and experiences of bullying, and a questionnaire regarding risk-taking and delinquent behaviors. This questionnaire was placed intentionally at the end of the assessment to allow the adolescent to build rapport with the examiner, and so that the examiner would have a good understanding of what supports (if any) the adolescent needed to complete this task.

Table 1 shows demographic characteristics at child age 13 by developmental status group (TD and DD). Child gender did not differ by status group and approximately 55% of the participants in each group were male. However, race/ethnicity did differ across the groups, with a higher percentage of Caucasians in the TD group than in the DD group. Family income and mothers’ education were also significantly related to group status. The TD group had more families with an annual income of $50,000 or higher than the DD group, and mothers in the TD group completed significantly more years of school than mothers in the DD group. Previous studies have found a strong association between measures of socioeconomic status and delinquency (Emerson & Halpin, 2013; Emerson & Turnbull, 2005); however, none of these variables related significantly to the dependent variables in the present study. Accordingly, the demographic variables were not covaried in any analysis.

Table 1.

Demographics by Delay Status at Child Age 13

Demographic	Typically Developing (n=77)	Developmental Delay (n=23)	X2 or t(df)
Child
WISC-IV IQ Mean (SD)	109.1 (12.1) Range: 80–134	65.0 (14.1) Range: 46–98	t = 14.45 (95)***
Vineland Scales of Adaptive Behavior Mean (SD)	96.4 (9.5) Range: 76–119	76.7 (9.5) Range: 61–100	t = 8.70 (98)***
Gender (% Male)	54.5%	56.5%	X2 = 0.03 (N = 100)
Race/Ethnicity (% Caucasian- Non-Hispanic)	63.6%	43.4%	X2 = 2.98 (N=100)(*)
Mother and Family
Income (% $50,000 +)	79.2%	60.1%	X2 = 3.18 (N=100)(*)
Mother’s Years of Schooling Mean (SD)	16.2 (2.4)	14.4 (2.2)	t = 3.24 (96)*

^(*)p < .10,

^*p < .05,

^**p < .01,

^***p < .001

Measures

Stanford-Binet IV (SB-IV; (Thorndike, Hagen, & Sattler, 1986)).

The SB-IV is a widely used instrument designed to assess the cognitive abilities of individuals 2 to 23 years in age. The SB-IV yields a composite IQ score with a normative mean of 100 and a standard deviation of 15. The SB-IV has good internal consistency and test-retest reliability (.95-.99 and .90 respectively for the composite score; Thorndike, et al., 1986; Sandoval & Irvin, 1988; Youngstrom, Glutting & Watkins, 2003). It is well suited to the evaluation of children with delays because starting points for the subtests can be adapted to meet the child’s developmental level. The composite score of the SB-IV was used as a measure of overall cognitive abilities at child age 5 as part of the group classification (described above).

Wechsler Intelligence Scale for Children, Fourth Edition (WISC-IV; Wechsler, 2003):

Scores from the Vocabulary, Matrix Reasoning, and Arithmetic subtests of the WISC-IV were used to determine an estimated IQ range and confirm status group classification at child ages 9 and 13. The selection of these three subtests was based on the high correlation between IQ estimates from these subtests and IQ scores based on the full WISC-IV administration, r = .91 (Sattler & Dumont, 2004).

Vineland Scales of Adaptive Behavior-II (VABS; Sparrow, Cicchetti, & Balla, 2005).

The VABS is a semi-structured interview conducted with a parent or other caregiver that assesses the child’s adaptive behavior in multiple domains. Scores from the communication, daily living skills and socialization subscales comprise the Adaptive Behavior Composite. This measure is shown to have high test-retest reliability (range of .81 to .88), as well as good internal consistency (alphas of .83 to .94; Sparrow, et al., 2005). The VABS was administered at ages 5, 9, and 13, and the composite score was used to group participants into the DD and TD groups.

Diagnostic Interview Schedule for Children IV (DISC-IV; Costello, Edelbrock, & Costello, 1985).

The DISC-IV is a highly structured diagnostic interview administered to the parent or caregiver that covers DSM-IV-TR criteria for all of the major mental illnesses observed in children and adolescents. Respondents are asked about the presence or absence of symptoms within the diagnostic categories, and disorder-specific algorithms are used to derive diagnoses from responses to individual items. The DISC-IV has undergone extensive testing, refinement and revision (Fisher, Shaffer, Piacentini & Lapkin, 1993; Shaffer, et al., 1993; Schwab-Stone, Fisher, Piacentini & Shaffer, 1993) and has achieved acceptable levels of reliability and validity (test-retest reliability of .48 to .86, depending on diagnosis - Shaffer, Fisher, Lucas, Dulcan & Schwab-Stone, 2000; k coefficients of .40 to .80 depending on diagnosis - Jensen et al., 1995). The DISC-IV was administered at ages 5, 6, 7, 8, 9, and 13 as a measure of child psychopathology. Standard administration was followed, except that mothers were given a brief description of each area (e.g. anxiety, fear, behavior problems) and asked to select those areas relevant to their child (with a minimum of one area). This procedure reduces administration time, increases reliability, and is judged to be more engaging for the respondents than the standard procedure of administering all areas in a fixed order (Edelbrock, Crnic & Bohnert, 1999).

Adolescent Risk-Taking and Delinquency Questionnaire.

This questionnaire was created for the current study and administered at youth age 13, as a measure of participation in risky and/or delinquent activities. It includes 31 items about the youth’s participation in a range of risk-taking and delinquent behaviors (plus two sample items). Items were based on questions from the Self-Report Delinquency Questionnaire (SRD; Elliot, & Huizinga, 1983), the Adolescent Risk-Taking Questionnaire (Gullone, Moore, Moss & Boyd, 2000), and other unpublished risk-taking and delinquency measures (Kaplan & Robbins, 1983; Leung & Lau, 1989). Given the younger age of the sample in the current study, we included items addressing a variety of rule-breaking, risk-taking, and delinquent behaviors that range in severity from “I have passed notes, instant messaged or texted friends while in class” to “I have used other drugs (cocaine, heroin, LSD, K, opium, etc.).” Participants responded to each item on a 4-point Likert scale (“never,” “once or twice,” “3–4 times,” “a lot”; represented numerically as 1–4).

The questionnaire was converted into an online survey using SurveyMonkey.com and presented on a laptop to reduce discomfort and encourage honest responses. A graduate student examiner was present for the administration and helped the adolescent complete two sample items (e.g. “I have seen a movie in a movie theatre.”). If the adolescent was able to read each item and appeared to understand how to complete the questionnaire, the examiner then moved to the opposite corner of the room to protect the adolescent’s privacy and offered to answer questions if/as needed. If the adolescent was unable to read or appeared to have any difficulty completing the questionnaire, the examiner sat next to him/her and provided continuous support. Despite intuitive concerns about the use of self-report with this population, previous research supports the use of self-report Likert scales with adolescents and adults with borderline to mild intellectual disability, and pre-tests (e.g. via the use of sample items) have been shown to increase response rates and validity (Hartley & Maclean, 2006).

Prior to analyzing the questions of interest, the Adolescent Risk-Taking and Delinquency questionnaire was divided into three subscales²: Rule Breaking, Risk-Taking, and Antisocial/Delinquent. As shown in Table 2, the Rule Breaking subscale consisted of those items assessing normative adolescent non-compliance with rules, including talking to other students during class, and using curse words/bad language. The Risk-Taking subscale focused on those items assessing true risk-taking (i.e. seeking a tangible or intangible reward with some risk of harm to self or other negative consequences), with items such as: skipping school/class, using substances, and riding in a car with an unlicensed or intoxicated driver. Finally, the Antisocial/Delinquent subscale focused on intentional violations of others’ rights or property and lying/cheating/deceiving. Items included stealing from other people/stores and lying about one’s age to purchase something.

Table 2.

Items by Subscale and Cronbach’s Alpha

Items	Rule Breaking	Risk-Taking	Antisocial/Delinquent
Talked to Others During Class	X
Eaten Food/Chewed Gum in Class	X
Worn Clothing Parents Disapproved of Outside of the Home	X
Used Bad/Curse Words	X
Loud or Rowdy in Public	X
Read Other Materials Besides Classwork in Class	X
Watched TV or Played Videogames that Parents Did Not Approve	X
Kissed Someone, Hooked Up, or Had Sex	X
Passed Notes, Texted or Instant Messaged in Class	X
Lied to Parents about Destination/Location	X
Downloaded Music without Paying	X
Not Worn a Helmet while Bike Riding, Skateboarding, or Rollerblading	X
Made a Bet/Gambled		X
Gotten into a Car with a Stranger/Hitchhiked		X
Run Away From Home		X
Watched/Read Pornography		X
Skipped Class or School Without an Excuse		X
In a Gang or Initiated into a Gang		X
Smoked a Cigarette, Used Drugs, or Drank Alcohol		X
Ridden in a Car with an Unlicensed/Intoxicated Driver		X
Ridden a Bike, Skateboard or Rollerblades off a Staircase or Wall		X
Have Done Tricks on a Bike, Skateboard or Rollerblades		X
Stolen from Another Person/Store			X
Lied about My Age to Get in Someplace/Buy Something			X
T-ped/Messed Up Someone’s Car or House			X
Snuck Out Past Curfew			X
Chatted with A Stranger Online			X
Written or Painted on Walls, Desks, or Other Surfaces			X
Threatened or Attacked Another Person			X
Cheated on a Test or Homework			X
Helped Someone Else Cheat			X
Cronbach’s Alpha (α)	.79	.74	.73
Cronbach’s Alpha (α) – Total Score - .87*

^*All items were included in the total score.

The internal consistency of these subscales was then examined, and Cronbach’s alphas ranged from .73 (Antisocial/Delinquent subscale) to .79 (Rule Breaking subscale). Accordingly, the internal consistency of each of the scales appears acceptable (Nunnally & Berstein, 1994), and items within the subscales can generally be considered to measure the same construct. The distributions of these subscales were then examined via quantile-quantile (Q-Q) plots to determine whether the variables were normally distributed (and thus, whether the assumptions of the planned two-way ANOVAs were met). As the distributions of the scores did not appear normal, non-parametric tests were used to confirm findings as appropriate.

Results

Given the relationship between risk-taking/delinquency and socioeconomic status seen in previous studies, we examined this in our sample. However, as mentioned, we did not find a significant relationship between risk-taking/delinquent behaviors and race, family income or mother’s education. As such, we did not covary these variables in our subsequent analyses.

The differences in risk-taking and delinquent behaviors between adolescents with or without developmental delays were examined. A two-way ANOVA was used to examine the main effects of disability status and previous diagnosis of ODD as well as the interaction between these two variables. Table 3 shows results from the ANOVA. Of note, distributions of the subscales suggest that these variables are not normally distributed and thus, the two-way ANOVAs reported might not be the most appropriate analyses. Accordingly, Mann-Whitney rank sum tests were used to confirm the main effects indicated in the tables. Results from the Mann-Whitney rank sum tests were almost identical to those found via the two-way ANOVAS. Therefore, the ANOVA results are presented for ease of interpretation, and the one discrepancy is described below.

Table 3.

Risk-Taking & Delinquent Behaviors (Raw Score) by Status and ODD Diagnosis – ANOVA Results

	Means (St. Dev)				F Values (Partial Eta2)
	TD (n=77)	DD (n=23)	No ODD (n=50)	ODD (n=45)	Status	ODD	Status x ODD
Rule Breaking Subscale (12 Items)	22.21 (6.07)	17.86 (5.07)	19.94 (5.10)	22.60 (6.86)	11.06** (0.11)	5.63* (0.06)	0.22 (<0.01)
Risk-Taking Subscale (10 Items)	11.99 (2.76)	10.77 (1.07)	11.50 (1.94)	11.93 (3.04)	4.23* (0.04)	0.39 (<0.01)	0.17 (<0.01)
Antisocial/ Delinquent Subscale (9 Items)	11.52 (3.09)	10.64 (2.68)	10.84 (2.18)	11.84 (3.69)	1.77 (0.02)	1.56 (0.02)	0.16 (<0.01)
Total Score (31 Items)	48.04 (11.09)	41.41 (7.26)	44.28 (7.89)	48.98 (12.74)	8.30** (0.08)	4.68* (0.05)	0.19 (<0.01)

^*p < .05

^**p < .01

In the DD group, 12 participants met criteria for ODD (52.2%) at least once between ages 5–9.

In the TD group, 33 participants met criteria for ODD (42.8%) at least once between ages 5–9.

On the Rule Breaking subscale, there was a significant main effect of having a developmental delay; youth with DD reported fewer rule-breaking behaviors than youth with typical cognitive development. Likewise, there was a significant main effect of having met criteria for ODD on at least one occasion between ages 5 and 9; youth meeting diagnostic criteria for ODD in childhood reported more rule-breaking behaviors.³ There was no interaction between disability status and prior diagnosis of ODD.

On the Risk-Taking subscale, there was a significant main effect of having a developmental delay; youth with DD reported fewer risk-taking behaviors. However, there was neither a main effect of prior ODD nor any interaction between the two predictors. Finally, for the Antisocial/Delinquent subscale, neither main effect nor the interaction was significant.

The role of gender was examined with two-way ANOVAs (run with gender and status group). Table 4 shows results from this analysis. There was a main effect of gender on Risk-Taking only, such that adolescent boys demonstrated more of these behaviors than adolescent girls. However, there were no significant interactions between status group and gender for any of the subscales.

Table 4.

Risk-Taking & Delinquent Behaviors (Raw Scores) by Status and Gender – ANOVA Results

	Means (St. Dev)				F Values (Partial Eta2)
	TD (n=77)	DD (n=23)	Male (n=55)	Female (n=45)	Status	Gender	Status x Gender
Rule Breaking Subscale (12 Items)	21.81 (6.16)	18.09 (5.07)	21.18 (6.18)	20.67 (6.08)	6.70* (0.07)	0.11 (<0.01)	0.01 (<0.01)
Risk-Taking Subscale (10 Items)	11.94 (2.71)	10.74 (1.05)	12.38 (2.99)	10.78 (1.17)	4.42* (0.04)	5.75* (0.06)	0.98 (0.01)
Antisocial/ Delinquent Subscale (9 Items)	11.49 (3.03)	10.74 (2.67)	11.49 (3.07)	11.11 (2.82)	1.19 (0.01)	0.43 (<0.01)	0.04 (<0.01)
Total Score (31 Items)	47.57 (10.99)	41.69 (7.23)	47.27 (11.18)	44.93 (9.62)	5.69* (0.06)	0.83 (0.01)	0.02 (<0.01)

^*p < .05

^**p < .01

In the DD group, 13 participants were male (56.5%).

In the TD group, 42 participants were male (54.5%).

Discussion

Overall, we found that 13-year-old adolescents with developmental delays reported fewer risk-taking behaviors than their typically developing peers. This was the case for normative rule-breaking/non-compliant behavior (Rule Breaking subscale) as well as true risk-taking behaviors (Risk-Taking subscale). However, adolescents with DD did not differ from typically developing peers with regard to the Antisocial/Delinquent subscale, and both status groups reported relatively low frequencies of these behaviors.

This finding is contrary to previous research, which suggests that adolescents with intellectual disabilities report engaging in more antisocial and delinquent behavior than their typically developing peers (Dickson et al., 2005; Emerson & Halpin, 2013). One possibility is that the age of the participants in our sample may have contributed to this finding. While the aforementioned studies also included younger adolescents, their participants ranged in age from 11–15 (Dickson et al., 2005; Emerson & Halpin, 2013). At age 13, all of our participants were at the start of adolescence. As such, many of these youth may not yet have had opportunities to engage in risk-taking and delinquent behaviors, given higher parental supervision and reduced freedoms at this age relative to later adolescence. This may be particularly true for adolescents with DD, as they usually receive even closer supervision than their typically developing peers, effectively preventing them from engaging in risk-taking behaviors.

The school setting may also contribute to the fewer rule-breaking behaviors reported by adolescents with DD. Particularly if adolescents with DD are in self-contained classrooms or non-public schools, they may have closer supervision than is typically provided in a regular education middle school. At the same time, behavioral expectations may also differ between special and general education settings. Thus, what is considered rule breaking in a general education classroom (e.g. talking to a peer) may not be seen as such in a special education setting. Unfortunately, our sample was too small to examine this effect; however, future researchers are encouraged to examine the role that classroom placement has on risk-taking and delinquency.

Finally, as mentioned previously, Dickson et al. (2005) used a broad definition of intellectual disability (or mental retardation, as it was typically termed at that time), and included adolescents with “learning problems,” who may represent a somewhat different population than the adolescents included in the present study. Likewise, Emerson & Halpin (2013) used educational classifications, and did not directly assess the intellectual functioning of their participants. In contrast, the present study used standardized assessment tools to classify participants and re-examined their classification across time. As a result, our sample is likely more representative of youth with ID than some of the studies described previously. This may also explain why, unlike previous studies, we did not find a relationship between risk-taking/delinquent behavior and measures of socioeconomic status.

On the one hand, cognitive and executive functioning deficits may place adolescents with DD at risk for engaging in more delinquent behaviors, by virtue of not fully understanding the consequences of their actions, and also being susceptible to the influence of peers engaged in these same behaviors. On the other hand, if we consider this population from a developmental perspective, we might expect that adolescents with DD are less likely to engage in delinquent behaviors. If they are functioning at a younger age socially, we would expect them to engage in fewer risk-taking and delinquent behaviors, consistent with our expectations for a school-age child. Findings from Dickson et al. (2005) and Emerson & Halpin (2013) are consistent with the former theory, while our results are consistent with the latter. To complicate the issue further, it may be that there is a division among the population of children with intellectual and developmental delays, such that those with higher IQs and more “learning problems” engage in more risk-taking and delinquency, while those with lower intellectual and adaptive functioning show fewer of these behaviors. Future studies might consider exploring this possibility, as our sample was too small to do so.

With regard to oppositional defiant disorder (ODD), adolescents who met criteria in early childhood (i.e. between ages 5–9) reported higher rates of rule breaking and non-compliance (Rule Breaking subscale) at age 13. However, they did not report engaging in higher rates of risk-taking behavior (Risk-Taking subscale) nor did they report higher rates of antisocial or delinquent behavior (Antisocial/Delinquent subscale). With regard to risk-taking, our findings are consistent with Bexkens, et al. (2018); having a behavior disorder such as ODD does not appear to increase risk-taking. In contrast, we find that having a prior diagnosis of ODD does impact rule-breaking behaviors and noncompliance. This makes sense within a developmental context; children with a history of ODD during middle childhood demonstrate higher rates of normative non-compliance than their peers in later adolescence. This likely reflects a temperamentally based tendency towards non-compliance, even if these adolescents do not go on to demonstrate more antisocial or delinquent behaviors.

While our results demonstrated main effects of ODD diagnosis and status group, we found no interaction between these two variables on risk-taking and delinquency. It is important to consider our sample size in this context. Specifically, it is likely that our numbers are too small to detect an interaction, and our finding of no interaction between ODD diagnosis and status group may be spurious as a result. Interestingly, our findings support previous research (Christensen & Baker, 2013), which suggests that ODD does not manifest differently for adolescents with and without DD. Nonetheless, the reader is encouraged to interpret the lack of interaction with considerable caution. As indicated below, we hope that future investigators will explore this further, with a larger sample.

When gender differences were examined, males reported higher rates of risk-taking behaviors than females. This finding is intuitive given that many of the behaviors included in this subscale involve physical activities that are more often performed by adolescent males as opposed to adolescent females (e.g. skateboarding). Moreover, previous research has consistently found differences in risk-taking and delinquency based on gender, with males demonstrating more of these behaviors (Loeber, Capaldi & Costello, 2013; Mennis & Mason, 2012; Rhodes & Fischer, 1993). Of note, there was no interaction between gender and status, suggesting that males with DD and those with typical development both engage in risk-taking behaviors more than females.

Limitations

There are a few limitations to the current study that require consideration when interpreting the results. First, the primary measure used in the analyses is self-report, and thus influenced by the adolescent’s understanding of the questions as well as his/her/their reporting biases. As discussed in Methods, this measure was administered with these possible limitations in mind. Accordingly, while there may be some limitations related to social desirability biases, these were reduced as much as possible by the procedure. Nonetheless, future research may focus on gathering collateral information to support the adolescents’ self-report. Future researchers may elect to include parent, teacher and/or peer reports of risk-taking and delinquency, keeping in mind that some of these reporters (e.g. parents) may have varying levels of awareness of such behaviors (e.g. parents of typically developing adolescents may have limited awareness of such behaviors unless/until the adolescent is caught, while parents of children with DD may monitor their adolescent’s behavior more carefully and thereby, have greater knowledge of any risk-taking or delinquent behavior).

Another limitation relates to the sample size included in the study and in the DD group in particular. The generalizability of our results must be considered in light of this, and hopefully future studies will explore risk-taking and delinquent behavior among larger samples of youth with ID. With regard to the Adolescent Risk-Taking and Delinquency Questionnaire, we hope that future studies will continue to explore the structure of this measure. We employed multiple methods to identify the subscales used in the current study (i.e. exploratory factor analyses, external/face validity). However, it would be ideal to examine the structure of this questionnaire with a much larger sample to better understand the relationships between different risk-taking and delinquent behaviors.

Finally, it is possible that the relationship between prior ODD diagnosis and rule-breaking behavior is the result of concurrent ODD symptoms (at age 13). In our sample, this does not appear to be the case, as very few participants met ODD criteria at age 13. Nonetheless, future research might consider the relationship between concurrent ODD or conduct disorder and risk-taking/delinquent behaviors, as our sample was too small to do so.

Conclusions, Implications and Future Directions

Overall, our results suggest that adolescents with intellectual disabilities (or borderline intellectual functioning) report fewer rule-breaking and risk-taking behaviors than their typically developing peers. Adolescents with a history of ODD report increased rule-breaking behaviors.

Our findings have a number of implications. First, if adolescents with intellectual disabilities engage in fewer age-appropriate risk-taking behaviors, this may place them at risk for being ostracized by peers. To this end, research suggests that adolescents with intellectual disabilities are bullied more often than typically developing teens (Christensen et al., 2013). In turn, being victimized by peers may place adolescents with intellectual disabilities at risk for internalizing disorders such as depression and anxiety. Social skills interventions may consider addressing this issue, encouraging youth with intellectual disabilities to identify ways of fitting in with peers without engaging in risk-taking and delinquent behaviors.

Second, our research suggests that youth with ODD between ages 5–9 continue to report higher rates of non-compliance at age 13. At present, there exist a number of empirically supported treatments for children with disruptive behavior disorders (e.g. Incredible Years, Triple-P Parenting Program, Parent-Child Interaction Therapy). While clinicians may be inclined to attribute most behavior problems to a child’s intellectual disability (called diagnostic overshadowing), research suggests that youth with intellectual disabilities meet criteria for ODD at a higher rate than their typically developing peers (Christensen & Baker, 2013). Moreover, children with and without intellectual disabilities do not appear to meet criteria for ODD in fundamentally different ways. Accordingly, it is important that clinicians refer youth with diagnoses of ODD to appropriate empirically supported treatments, regardless of whether these children also meet criteria for an intellectual or developmental disability. For children who meet ODD criteria in early to middle childhood, treatment may reduce the likelihood of their continuing to engage in non-compliant behaviors over time (Eyberg & Boggs, 1989; Sanders, 1999; Webster-Stratton & Reid, 2003). For older youth, treatments such as Multisystemic Therapy and Functional Family Therapy may be necessary to prevent increasingly delinquent behaviors, and/or the onset of conduct disorder (for a review, see McCart & Sheidow, 2016).

As the participants included in the current study were 13 years of age and thus at the start of adolescence, it is important for future research to address whether the results of the current study persist during late adolescence. It is possible that as adolescents age and gain greater freedoms, those with intellectual disabilities may start to have the same opportunities to engage in risk-taking behaviors as their typically developing peers. Accordingly, it is important to address whether adolescents with intellectual disabilities begin to engage in risk-taking and delinquent behaviors at the same, greater, or lesser rates than their typically developing peers in later adolescence.

Finally, we want to highlight the importance of using standardized measures to assess intellectual disability, and this may be one of the reasons our findings differed from previous studies. In order to build a literature that accurately describes a vulnerable population such as youth with ID, it is necessary that we as researchers use similarly rigorous methods for identifying and describing our participants. Thus, it is our hope that future studies will employ similar standardized assessments when further exploring the risk-taking and delinquent behaviors of youth with ID.