Risk-taking among adolescents: Associations with social and affective factors
Abstract
The research investigated the associations of social and affective factors with risk-taking in male and female adolescents. A sample of 269 Israeli adolescents completed questionnaires measuring frequency of involvement in risk-taking behaviours, relationships with parents, orientation towards peer group, depressive mood, and aggressive behaviour. Correlations and multiple regression analyses showed that risk behaviour among male adolescents was mainly related to orientation towards peer group, while for female adolescents relationships with parents was the prominent factor in risk behaviour. The parental factor also contributed to the depressive mood of both genders in the sample. However, depressive mood showed only a weak association with risk taking. These results underscore the differential associations of relationships with parents vs. peers among adolescent boys and girls, respectively, in regard to risk taking.
Introduction
Risk-taking among adolescents
The period of adolescence (ages 11–21) is a time of multiple transitions, namely, the transition to puberty and transitions involving parent–child relationships, school, peers, and cognitive and emotional abilities. This period is also characterized by an increase in risk-taking behaviours, e.g. those linked to careless driving, substance use, unprotected sexual behaviour, eating disorders, delinquency, homicidal and suicidal behaviours, and dangerous sports (e.g., Muuss & Porton, 1998). These behaviours are defined as risky since they are usually volitional, their outcomes are uncertain (Furby & Beyth-Marom, 1992; Igra & Irwin, 1996), and they entail negative consequences (Irwin, 1990).
Increasingly larger segments of young people today seem to adopt risk-taking behaviours (DiClement, Hansen, & Ponton, 1996). The National Longitudinal Study of Adolescent Health in the United States (Zweig, Duberstein-Lindberg, & Alexander-McGinley, 2001) found that the most frequent risk-taking or risk-associated behaviours among high school adolescents were the consumption of alcohol (66% of the males and 65% of the females), smoking cigarettes (64% and 64%, respectively), sexual intercourse (49% and 48%, respectively), binge drinking (38% and 31%, respectively), using marijuana (35% and 31%, respectively), and being in a physical fight (41% and 20%, respectively). Less frequent were behaviours such as using other illicit drugs besides marijuana, and having serious thoughts about and attempts at suicide (13% and 17%, respectively). A WHO longitudinal cross-national study in 2001/2002 investigating health behaviour in school-aged children in 35 countries in the America and Europe showed that by age 15, 23% of young people smoked cigarettes, 29% drank alcohol on a weekly basis, 22% tried cannabis, and 45% of all age groups (11, 13 and 15 yr old) reported at least one injury in the previous year (Currie et al., 2004). In Israel (Harel, Kany, & Rahav, 1997), 24% of a broad sample of adolescents reported that they had been involved in risk-taking behaviour once or twice over the preceding year in order to feel “how it is” or to prove they could do so; and 26% of the boys and 11% of the girls reported they had been involved in risk-taking behaviour several or many times during the same period. The data indicate that overall, risk-taking behaviours begin at an early age, increase over the adolescent years, and are more common among boys than girls. The present study focuses on the associations of social relationships and negative affects with risk taking among adolescents.
Social factors
Social/environmental theories emphasize the influence of parents, peers, teachers, community and culture on risk taking during adolescence, which is conceptualized as a period of growing autonomy and emerging individuation from the family (Igra & Irwin, 1996), yet in parallel sustained reliance on parents and other significant adults (Laible, Carlo, & Raffaelli, 2000). In this period adolescents undergo changes in roles and status that redefine their place in society and may lead to confrontations and conflicts with parents (Coleman, 1992; Holmbeck, Paikoff, & Brooks-Gunn, 1995; Paikoff & Brooks-Gunn, 1991; Paikoff, Carlton-Ford, & Brooks-Gunn, 1993). However, good parenting, which includes frequent communication, regular daily involvement, monitoring and related skills, and instilling appropriate coping skills, may assist the adolescent in avoiding involvement in deviant behaviour and enable him/her to deal with frustration, control anger, and apply other self-management strategies. Furthermore, the perceived importance of the parents is related to lower levels of susceptibility to antisocial peer pressure and thus to less involvement in risk-taking behaviour (Sim, 2000).
In parallel to the relationship with parents, the relationship with the peer group functions both as a source of social support and as a source of temptation and endangerment. Peer pressure by the adolescent social group tends to encourage involvement in dangerous acts, with socialization in the peer group leading the individual towards risk-taking behaviour (Diblasio, 1986). The adolescent becomes involved in risk-taking behaviour because he/she seeks to participate in experiences that appear relevant to the group identity and to attain accomplishments that establish a self-identity within the group (Lightfoot, 1992). Indeed, perceived susceptibility to peer pressure, along with having a friend who drinks alcoholic beverages, have been found to be the two most consistent predictors of substance abuse among adolescents (Flannery, Vazsonyi, Torquati, & Fridrich, 1998).
In sum, considerable research supports the finding that both peer group (e.g. Caffray & Schneider, 2000; Glendinning & Inglis, 1999; MacNeil, Stewart, & Kaufman, 2000; Paetsch & Bertrand, 1997; Romer et al., 1994; Santor, Messervey, & Kusumakar, 2000) and parents (Field, Lang, Yando, & Bendell, 1995; Gonzalez et al., 1994; Resnick et al., 1997; Sim, 2000) have a strong influence on adolescent behaviour, including risk taking. Hansell and Mechanic (1990) suggest that adolescents’ risk behaviours are affected by parents and by peer value systems, but that these two factors are not compatible. Parents usually value the wellbeing of their children, and are concerned about the implications of health-related behaviours in the long term. Peers usually emphasize the present and have less concern about the future and thus about risk-behaviour consequences. Hence, positive relationships with parents may be expected to be shown as relating to less risk-taking, while orientation towards peers may be expected to be shown as directed toward more risk-taking among adolescents.
Affective factors
Positive relationships with parents were shown to be associated with less depression and aggression among adolescents (Field et al., 1995; Hurd, Wooding, & Noller, 1999; Laible et al., 2000; Lasko et al., 1996). It is claimed that adolescent risk taking is partly based on the impetus to overcome or eliminate negative feelings resulting from poor relationships with parents by gaining the social rewards inherent in participating with others in risky behaviours (Caffray & Schneider, 2000). Specifically, feeling depressed is a negative emotion that adolescents may seek to change by risk-taking. Data suggest that depression during adolescence is highly prevalent, with 20–35% of male adolescents and 25–40% of female adolescents reporting depressed mood (Peterson, Compas, Brooks-Gunn, & Stemmler, 1993). Depressive moods, in turn, are related to risky behaviours such as smoking (Coogan et al., 1998) or sexual risk-taking (Bonomo et al., 2001; Shrier, Harris, Sternberg, & Beardslee, 2001).
Aggressive behaviour, another common phenomenon during adolescence (Ellickson, Saner, & McGuigan, 1997), is defined as a behaviour that involves intent to harm another person. It is related to anger and hostility and can be manifested in verbal or physical abuse (Felson, Liska, South, & McNulty, 1994). Aggression, and the interpersonal violence that usually results from it, contribute to injuries, illnesses and deaths (Griffin, Scheier, Botvin, Diaz, & Miller, 1999). In a research sample of adolescents in the US, 43% reported that they had threatened or hit someone during the preceding year, 14% had assaulted others, and 13% had carried a weapon (Ellickson et al., 1997). In Israel, 50.3% of the boys and 22.7% of the girls reported that they were involved in bullying, harassing or fighting once or twice over the preceding year (Harel et al., 1997). Aggression and violence among adolescents usually coexist with emotional and behavioural problems, including risk-taking behaviour (Ellickson et al., 1997; Haynie et al., 2001; Silver, Field, Sanders, & Diego, 2000) such as substance use (Flannery et al., 1998).
In sum, one of the factors involved in adolescent engagement in risk behaviour may be the need for regulation of emotion and affect, which in themselves result from relationships with parents and peers.
Research model and hypotheses
The present research focuses on social and affective factors as the main variables associated with risk taking among adolescents. It was assumed that social factors, i.e. parental and peer factors, will affect risk taking in adolescents both directly and through the mediation of affective factors, i.e. depressive mood or aggression, as depicted in Fig. 1.
Analytic research model.
- 1. Positive relationships between adolescents and their parents will be related to less risk taking.
- 2. A stronger orientation of adolescents towards their peer group will be related to more risk taking.
- 3. Higher levels of depressive mood or aggressive behaviour will be related to more risk taking.
- 4. Positive relationships between adolescents and their parents will be related to less depressive mood and less aggression.
- 5. A stronger orientation of adolescents towards their peer group will be related to more aggression and more depressive mood.
A path analytic model was also tested, with the expectation that parental and peer factors are related directly to risk taking and indirectly through their associations with depressive mood and aggression. The path model was tested separately for boys and girls, in light of prior findings showing different effects of social and affective factors as a function of gender (Linver & Silverberg, 1997; Marshal & Chassin, 2000; Rodgers, 1999).
Method
Sample
The sample consisted of 269 Israeli adolescents in 10th (49.4%) and 11th (50.6%) grade, ages 16–18 yr (
, sd=0.60), 59.4% girls and 40.6% boys. All were Jewish and Hebrew-speaking. Most (83.6%) were born in Israel. Most (88.1%) came from intact families with 
(sd=0.84, range=1–6) children. About half the parents were born in Israel (50.4% of the fathers, 56.4% of the mothers) and the rest in Europe (35.1% and 27.1%, respectively), Africa (11.2% and 12.8%, respectively) or elsewhere. About half the participants reported that their parents had higher education (50.2% of the fathers, 53.1% of the mothers), and most of the rest reported a high school education (46.7% and 45.8%, respectively). They assessed their economic status as very good (19.7%), good (40.2%) or average (33.7%). Most were not religiously observant (78.9%).
Instruments
The following inventories were used:
1. Modified risk involvement questionnaire. This questionnaire on risk taking is based on a scale by Siegel et al. (1994; 19 items) and a modified version of it (Shapiro, Siegel, Scovill, & Hays, 1998; 26 items). The original items depict risk behaviours in the areas of driving, health, drugs and lawbreaking. The assessment of risk behaviour is obtained by self-report on frequency of involvement in risk-taking behaviours during the preceding year (range: 0=“never” to 8=“daily or more”). The test–retest reliability for the scale was 0.86; internal reliability was 0.72. The translation into Hebrew and the adaptation of the scale to the Israeli milieu is described in Ben-Zur and Reshef-Kfir (2003). The final Hebrew version contained 27 items—17 from the Shapiro et al. inventory and 10 modified or new ones. The rating scale was simplified to 0=“never,” 4=“daily or more.” The internal reliability of the Hebrew version was high 
, and its test–retest correlation was 0.76, suggesting good reliability of the data over time. An additional short questionnaire of 4 items was included to measure social daring (Harel et al., 1997; e.g. “Do you sometimes take risks to impress your friends?”) showing a correlation of 0.58 
with the risk involvement scale, suggesting a good validity for the scale as well. The internal reliability of the questionnaire was high (
; see Table 1). Following an inspection of the risk-data distribution, which showed a high skewness due to several very high scores, the 2, 3, 4 scores were recoded as 1. The transformed mean was 0.25 (sd=0.14, range=0.00–0.93, 
).
| Variable | Mean | sd | Range | Alpha |
|---|---|---|---|---|
| Risk-taking behaviour | 0.47 | 0.39 | 0.00–3.26 | 0.86 |
| Adolescent–parent relationships | 4.56 | 0.95 | 1.43–6.00 | 0.97 |
| Orientation towards the peer group | 2.62 | 0.45 | 1.50–4.19 | 0.83 |
| Depression | 0.29 | 0.19 | 0.03–1.00 | 0.88 |
| Aggression | 0.55 | 0.15 | 0.19–0.93 | 0.81 |
| Social desirability | 0.46 | 0.23 | 0.00–1.00 | 0.51 |
2. Adolescent–parent relationships scale. This assessment of parent–child relationships by adolescents consists of 30 items relating to each parent, selected from the Relationships with Father/Mother Questionnaire (Mayseless & Hai, 1998; Mayseless, Wiseman, & Hai, 1998). Of seven subscales in the original questionnaire (emotional closeness, communication, mutuality, open confrontation, coolness and rejection, autonomy, and supervision)—all with high reliability levels and high intercorrelations—10 items were chosen from the emotional closeness scale (e.g. “She tends to hug me occasionally”), 10 from the communication scale (e.g. “He always listens to my ideas and opinions”), and 10 from the mutuality scale. The items were rated on a 6-point scale (1=“not true at all,” 6=“very true”). The three 10-item sets were highly correlated for fathers and for mothers and were therefore averaged in one index. The means of the resulting 30-item scales were found to be above scale midpoint (i.e. 3), with high internal reliabilities for the fathers and mothers (
). The adolescent-father and adolescent-mother relationships scores were also highly correlated (
, 
) and were therefore averaged and used as a combined score in the analyses that followed.1
3. Orientation toward the peer group questionnaire. This questionnaire is based on a scale by Sharabany (1980) and a modified version of it (Rosenthal, 1980). It consists of 36 items divided into 18 pairs according to the following dimensions: (1) dependence, (2) conformity to group pressure, (3) group as a source of help, (4) group as a source of power, (5) individual as representative of the group, (6) group as a source of support, (7) conformity to group values, (8) anxiety over group approval, (9) primacy of the group over the individual, (10) group symbolism, (11) group discipline, (12) conditional acceptance, (13) group as a closed system, (14) group possessiveness, (15) anti-adult orientation, (16) non-conditional acceptance, (17) feelings of diffusion of responsibility and (18) primacy of the individual over the group. Each dimension was composed of two consecutive items (dimension 1=items 1, 2; dimension 2=items 3, 4, etc.). The assessment of orientation toward peer group was obtained by the respondent's agreement/disagreement with the statements along a Likert-type scale (1=“do not agree at all” to 5=“very much agree”). Responses to items that described a weak orientation towards the peer group were reversed. Rosenthal (1980) reported an internal reliability for this scale ranging from 0.31 to 0.81 for each of the 18 dimensions, and tested validity with another questionnaire that measured intimacy (Sharabany, 1974). After summing the 18 pairs of items, the internal reliability for the entire scale found in the present study was 0.83. Hence, the general score was used, representing the level of the adolescent's orientation towards the peer group.
4. Depressive adjective checklist (DACL). This questionnaire on depressive mood is based on a scale developed by Lubin (1967) and a modified version translated into Hebrew (Lomeranz, Lubin, Medini, & Eyal, 1981). It consists of 34 items that describe moods: 22 items relating to a bad mood (e.g. “unhappy,” “sad”) and 12 items relating to a good mood (e.g. “healthy,” “lucky”). The questionnaire is filled in by using the trait version, instructing the respondent to mark only items that describe his/her mood. Scoring is done by summing the bad-mood items marked plus the good-mood items unmarked (range 0–34). The present study used the average score for the 34 items with a high score indicating a more depressed mood. The internal reliability found for the Hebrew version (Lomeranz et al., 1981) was 0.85, and for the present study 0.88.
5. Aggression scale. This questionnaire, based on a scale by Buss and Durkee (1957), differentiates between eight types of hostility and aggression: (1) Assault—physical violence against others. (2) Indirect hostility—both roundabout and undirected aggression. (3) Irritability—a readiness to explode with negative affect at the slightest provocation. (4) Negativism—oppositional behaviour, usually directed against authority. (5) Resentment—jealousy and hatred of others. (6) Suspicion—projection of hostility onto others. (7). Verbal hostility—negative affect expressed in both the style and content of speech. (8) Guilt—feelings of being bad or having done wrong or suffering pangs of conscience. The original eight subscales included 75 items that indicated two types of hostility (subscales 5, 6), five types of aggression (subscales 1, 2, 3, 4, 7), and one type of guilt (subscale 8). The assessment of aggressive behaviour was obtained by self-report on involvement or lack of involvement in aggressive behaviours. The internal reliability of the original questionnaire (Buss & Durkee, 1957) was 0.72. Only four subscales were included in the present research: assault (8 items; 
), indirect hostility (6 items; 
), irritability (17 items; 
), and verbal hostility (13 items; 
). A general “aggression” scale was used instead of the specific subscales, due to their medium reliability levels. The internal reliability of the final questionnaire was 0.81.
6. Social desirability scale (SDS). This eight-item measure, based on Crowne and Marlowe's social desirability questionnaire (1964), was used in its Hebrew format (Ben-Zur, 2002; 
). The SDS represents the mean of the eight items (marked as true or untrue) after reversing four items in a social desirability direction. The scale showed medium reliability (
) in the present study, but since it was positively correlated with aggression, it was used as a control variable in the path analysis.
7. Demographic data. Items consisted of gender, age, grade in school (i.e. 10th, 11th), parents’ marital status, number of siblings, and assessment of economic status.
Procedure
Data collection took place in 2001 in a comprehensive regional school located in northern Israel. The school draws students from a heterogeneous representative population, who are not at risk in any special way. The student body contains a range of scholastic levels from full matriculation certificate (highest level) to certificate of completion of 12 years of study (lowest level). Each grade in the school has 14 classes. The participants were pupils in four 10th grade classes and four 11th grade classes chosen randomly from each grade.
The inventory was completed voluntarily during class time by all students present. It was prepared in four sets that differed in order of questionnaires based on a Latin square design. Each of the four sets was distributed in equal proportions in each class and for each gender and all were filled in anonymously.
Results
Pearson correlations were used to examine the associations between the indicators of relationships with parents and orientation towards peers, depressive mood, aggressive behaviour and risk-taking behaviour (Table 2). All effects were tested using a conservative 
level.
| Risk-taking behaviour | Adolescent–parent relationships | Orientation towards the peer group | Depression | Aggression | |
|---|---|---|---|---|---|
| Adolescent–parent relationships | −0.26*** | ||||
| Orientation towards the peer group | 0.22** | 0.11 | |||
| Depression | 0.21** | −0.51** | −0.01 | ||
| Aggression | 0.26*** | −0.17 | 0.24** | 0.26** | |
| Social desirability | −0.10 | 0.13 | −0.15* | −0.12 | −0.50*** |
- *
; **
; ***
.
As can be seen in Table 2, more positive relationships between the adolescents and their parents were associated with less risk taking, while a stronger orientation towards the peer group was associated with more risk taking. Higher levels of depressive mood or aggressive behaviour were also associated with more risk taking. Furthermore, positive relationships between the adolescents and their parents were associated with less depressive mood, while a stronger orientation towards the peer group was associated with more aggression. These results conform generally to the research hypotheses.
Multiple regressions were then conducted to assess the specific contribution of each variable to risk-taking behaviour (Table 3). The regressions were also used to create a path analysis in which relationships with parents and orientation towards peer group were assumed to affect depressive mood and aggression, which in turn were assumed to affect risk-taking (Fig. 2). Prior to these analyses, the demographic variables were tested for their associations with risk-taking. Age was the only variable to show a significant association. Gender, perceived economic status and social desirability were included as control variables.
| Dependent variable | Mediator/independent variables | ||||
|---|---|---|---|---|---|
| Risk-taking | Depression | Aggression | Peers | Parents | |
| Depression | 0.03 | — | — | — | — |
| Aggression | 0.19* | — | — | — | — |
| Peers | 0.19* | 0.10 | 0.17* | — | — |
| Parents | −0.20* | −0.48*** | −0.12 | — | — |
| Gender | −0.06 | 0.12 | −0.04 | −0.27*** | 0.09 |
| Age | 0.16* | 0.03 | −0.05 | 0.01 | −0.09 |
| Economic status | 0.06 | 0.18** | 0.01 | −0.10 | −0.23** |
| Social desirability | 0.06 | −0.07 | −0.46** | −0.10 | 0.11 |
| R2 | 0.19** | 0.31*** | 0.30*** | 0.11*** | 0.08** |
- *
; **
; ***
.
Path analysis describing the structural relations between the research variables—general.
As can be seen in Table 3, the resulting R2 for the dependent variable was 0.19 (F[8,242]=6.53, 
). In conformity with Hypotheses 1 and 2, positive relationships between adolescents and their parents had a significant negative direct effect on risk-taking (
, 
), while orientation towards peer group had a significant positive direct effect (
, 
). High levels of aggression also had a significant positive direct effect on risk-taking (
, 
), supporting part of Hypothesis 3 in this analysis. However, depressive mood had no direct effect on risk-taking behaviour in this analysis, contrary to part of Hypothesis 3.
As can be observed in Table 3 and Fig. 2, relationships with parents were highly related to the adolescent's depressive mood (
, 
) but not to aggression, while orientation towards peer group was related to aggression (
, 
) but not to depressive mood. These results confirm parts of Hypotheses 4 and 5. Additionally, the age of the adolescent had a significant effect (
, 
), with risk-taking increasing with age.
To further our understanding of these findings, the path model (Fig. 2) was tested separately for boys and girls (Figs. 3 and 4). The resulting data showed that the girls’ risk-taking behaviour was significantly affected by relationships with parents, while the boys’ risk-taking behaviour was significantly affected by orientation towards the peer group. Additionally, relationships with parents had an effect on boys’ depression and on girls’ depression and aggression. Hypothesis 1, therefore, was confirmed for girls, while Hypothesis 2 was confirmed for boys. The associations of affective factors with risk-taking, however, were found to be weak and non-significant in the separate analyses. Hypothesis 3, therefore, was not confirmed in this analysis. The associations of relationships with parents with depressive mood were observed for boys and girls, while the peer orientation-aggression association was not replicated, suggesting that this association is weaker. Thus, only part of Hypothesis 4 was confirmed in the separate analysis, and Hypothesis 5 was not confirmed in this analysis.
Path analysis describing the structural relations between the research variables—boys.
Path analysis describing the structural relations between the research variables—girls.
Discussion
The present research investigated the associations between involvement in risk-taking behaviour by adolescents, and social and affective factors.
Social factors
Parental influence and orientation towards the peer group correlated with risk taking in a reverse pattern: good relationships with parents were related to low involvement in risk-taking behaviour among adolescents, while stronger relationships with peers were related to high involvement in risk taking. These findings conform to other studies about parental influence (Field et al., 1995; Gonzalez et al., 1994; Resnick et al., 1997; Sim, 2000) and peer influence (Caffray & Schneider, 2000; Flannery et al., 1998; Glendinning & Inglis, 1999; MacNeil et al., 2000; Paetsch & Bertrand, 1997; Romer et al., 1994; Santor et al., 2000) on adolescent risk taking. However, while these associations seemed to conform to the hypotheses, the regression analyses indicated differentiation by gender: risk taking by the boys was mainly related to strong orientation towards the peer group, while risk taking by the girls was related to positive relationships with parents.
Research has shown that parental warmth and support facilitate a positive relationship between interpersonal communication, parental monitoring, and low sexual risk-taking behaviour among adolescent offspring (Meschke, Bartholomae, & Zentall, 2000). This finding, along with other empirical evidence presented above, and the findings of the present study, support the argument that the extent of parents’ influence on risk taking by their children is dependant mostly on the quality of their mutual relationships. Positive and stable relationships may be assumed to engender beneficial outcomes for both sides: parents and children have good feelings and the environment at home is pleasant; the children can rely on parental moral support and guidance; the parents act as role models and as authentic examples of values and norms; the parents can monitor the children and not feel isolated from them; and the children have a high regard for their parents’ opinions and tend to get their approval. Presumably, such relationships foster low risk-taking behaviour among children.
One of the strongest effects on risk-taking in the present study was found to be the orientation towards the peer group. Besides being subjected to peer pressure and encouragement, individuals absorb two types of norms from the peer group (Sheeran, Abraham, & Orbell, 1999). The first, known as “subjective norms”, relates to a person's perception of social pressure to perform in a certain way, and his/her beliefs about what the peer group thinks he/she should do. The second, known as “descriptive norms”, relates to a person's perception of his/her significant other's attitudes and behaviours in each life domain, which provide information that he/she may use in deciding how to behave. It may be assumed that adolescents who have a strong orientation towards the peer group will tend to follow their peers’ expectations and will be affected by the group's attitudes and acts. These adolescents might be more involved in risk-taking behaviour, as compared to adolescents with a weak orientation towards the peer group. Furthermore, for these adolescents, risk taking will constitute an act of identification with the group, loyalty towards its members, and conformity with its values.
The effect of orientation towards the peer group on adolescent risk-taking supports the literature that holds that the peer group is the most significant source of social support for the individual during adolescence (Furman & Buhrmester, 1992). Reliance on the peer group becomes increasingly vital during this period as a result of the impetus to gain autonomy from parents and because of adolescent concerns, which are easier to handle with friends than with parents (e.g. sexual issues). According to one view (Harris, 1998), experiences with peers constitute the primary social factor responsible for molding the individual's personality and his/her socialization in the culture.
This view is borne out in extant research about the differential effects of parents and peers on adolescent risk behaviour as a function of gender. Rodgers (1999), investigating male and female adolescents in grades 9–12 who reported having had sexual intercourse, found that less parental psychological control (i.e. more autonomy) decreased the odds of sexual risk-taking for female adolescents only. This finding was interpreted as indicating greater psychological maturity on the part of the girls, leading to less risky sexual decisions. In a similar vein, Linver and Silverberg (1997) found that gender moderated the associations between 8th-grade adolescents’ psychosocial maturity and their relationships with their mothers, with maternal warmth and psychological autonomy more strongly related to psychological maturity among girls than among boys. Marshal and Chassin (2000), studying the combined effects of parents and peers on alcohol use among adolescents aged 10.5–15.5 yr, found that support by both parents, or consistency of discipline, moderated the association between the adolescent's alcohol use and affiliation with drug-use peers, depending on gender. For girls, parental influence led to low risk even under high peer influence, while for boys, parental influence did not mitigate against high risk under high peer influence.
One possible explanation for these data, and for the present research findings, relates to the perceptions by adolescents of their relationships with their parents (Linver & Silverberg, 1997; Marshal & Chassin, 2000). Girls may perceive warm and supportive relationships with their parents as conforming with the traditional roles of gender, and as a result may tend to spend more time with their parents and incorporate their parents’ values more fully, leading, in turn, to less risk taking (e.g. Hansell & Mechanic, 1990). Boys, in contrast, may interpret the same parental behaviour as threats to their autonomy and sense of control. Such perceptions may lead them to greater attachment to the peer group and its risky orientation and values. The greater peer influence on adolescent boys, therefore, may be explained by their stronger need for independence from parents, and their reliance on their peers as the dominant source of support, while girls tend to retain their reliance on their parents in parallel to ties with peers.
Affective factors
The affective factors of depression showed weaker associations with the adolescents’ risk-taking behaviour, as compared with the social factors. A significant relationship between depression and risk taking was observed only on the correlational level but not in the regression analyses. These results support the literature which emphasizes the stronger influence of interpersonal rather than intrapersonal factors in predicting adolescent behaviour (Flannery et al., 1998). By contrast, low depression and low aggression were related to positive relationships with parents. This conforms to other research indicating that adolescents who reported greater intimacy with their parents showed higher self-esteem, happiness with self, and less depression (Field et al., 1995; Lasko et al., 1996), and were more positively adjusted and resilient (Herman-Stahl & Peterson, 1996). Regard for parents by adolescents mediated the relationship between parental monitoring and susceptibility to antisocial peer pressure, and positive relationships with parents were related to adolescents’ higher life satisfaction (Sim, 2000).
Advantages, limitations and recommendations
The present study tested a large sample of adolescents on a measure of risk taking that included a large range of risk behaviours. The research also tested social and affective factors in the same context, allowing for an assessment of their unique effects on adolescent risk taking. A short measure of social desirability was included in the analyses, so that the results represent outcomes that are free of the effects of this tendency.
Conceivably, since the study took place at school during class time, with teachers present, the environment may have engendered feelings of threat, stress and discomfort, even though a commitment to anonymity was made to the respondents. Additionally, since the measure of risk behaviours is based on self-report, the extent to which it reflects actual behaviour is not known.
Future research may profit by sampling adolescents in varied socioeconomic levels and sectors (e.g. new immigrants, youth at risk) to provide a more comprehensive perspective for examining the impact of social and affective factors on risk-taking behaviour among adolescents. Furthermore, testing the associations of different emotional constructs, such as anger or sadness, in the context of social factors affecting risk taking may enhance the understanding of the processes involved in the influence of social and affective factors on risk taking. Lastly, the findings suggest that future research should focus on gender differences in risk-taking behaviours and in their associations with social and affective factors.
The intense relationships between the individual and the peer group during adolescence point to the importance of planning prevention and intervention programs regarding risky behaviours on a group level, in addition to individual counseling. Interventions might be more effective if programs are conducted within cliques instead of in school classes. The results of the study also suggest that interventions might achieve their goals successfully if they are conducted for each gender separately, emphasizing the differential influential agents among boys and girls—namely, peer group and parents, respectively.
