Introduction

Arnett (2004) posited that the time between late adolescence and young adulthood constitutes a distinct developmental period marked by novelty, instability, and identity exploration. During this timeframe, individuals are expected to gain knowledge about their life direction by experientially testing the feasibility of their dreams against the constraints of adult life. Exploration, choices, and changes concerning salient contexts in emerging adult development (e.g., love, career, personal identity) can generate a chain of life events marked by considerable emotional upheaval (Arnett 2004; Rodgers and Tennison 2009). Consequently, it is important for emerging adults to expand the strategies by which they regulate their emotions and steady themselves. Failure to identify and implement a diverse range of emotional modulation tactics may increase the susceptibility to engage in risky behaviors (Arnett 2005). Specifically, emerging adults who employ immature and rigid emotion modulation tactics are likely to perceive risky behaviors as a viable means to reduce negative affect and generate positive emotional experiences (Ben-Zur and Zeidner 2009).

Voluntary participation in risky activities that increase probability of harm is a source of great concern (Cooper 2002; Fromme et al. 1997). To date, research indicates that 40% of college students admit to binge drinking (Centers for Disease Control and Prevention 2009) and 60% of students report using marijuana by their sophomore year (Arria et al. 2008). Each year, nearly 100,000 U.S. students indicate that at some point they were too drunk to know if they had consented to sex (Hingson et al. 2002). Of further concern, students who engage in frequent high-risk behaviors are more likely to develop mental health difficulties, contract an STD, drop out of college, and/or attempt suicide (Andersson et al. 2009; Schaffer et al. 2008; Seal and Agostinelli 1996).

Gender differences exist in the frequency of self-reported risky behaviors among college-attending emerging adults (Hirschberger et al. 2002). College men not only participate in risky behaviors more frequently than college women, but they also have more peers who participate in these types of activities (Rolison and Scherman 2003). College men also perceive greater benefits and fewer consequences associated with risky activities (Hirschberger et al. 2002). Conversely, college women attribute more negative consequences to risky behaviors (Rolison and Scherman 2003). College men have been found to engage in higher rates of physical aggression, binge drinking, and casual sex; they also report more permissive attitudes toward sex than college women (Caetano and Cunradi 2002; Cubbins and Tanfer 2000; Nelson et al. 2008). However, gender differences in the rates of engaging in some of these behaviors may be diminishing (Rolison and Scherman 2003).

Consequently, there is a need to examine current gender differences in the rates and perceived likelihood of college students engaging in risk-taking while identifying gender-specific and gender invariant risk factors for engaging in these behaviors. Identifying vulnerability factors that may contribute to the onset and/or exacerbation of these harmful habits is also necessary (Cooper 2002; Fromme et al. 1997). To date, few studies have examined the influence of debilitative cognitive factors on college students’ anticipated engagement in risky behaviors, yet, these dysfunctional cognitions may constitute a fruitful intervention point for a variety of problematic behaviors. The current study sought to examine gender-specific associations between maladaptive schemas and different risk-behavior subtypes as outlined by Fromme et al. (1997). Table 1 depicts the operational definitions for these risk-behavior subtypes.

Table 1 Classification and operational definitions for risky behavior subtypes (Fromme et al. 1997)
Full size table

Debilitative emotional patterns and embedded cognitive themes can be examined in the context of “early maladaptive schemas” which develop as the result of chronic unmet emotional needs in childhood and adolescence (Young et al. 2003). According to Young and colleagues, early maladaptive schemas originate from the interaction between a child’s innate temperament and the dynamics of the nuclear family. Selected temperaments (e.g., anxious, shy, passive) leave children vulnerable to experience difficult life circumstances and may increase the probability of a misfit between parent and child (Belsky and Pluess 2009). For instance, emotionally labile children often elicit invalidating, harsh, and critical responses from parents/guardians. Resulting memories, emotions, and bodily sensations from conflictual or abusive interactions with primary caregivers can then form a cognitive blueprint (early maladaptive schemas) by which children learn about and interact with the world.

Early maladaptive schemas regulate how individuals view themselves as well as their interpersonal relationships with others. Young et al. (2003) hypothesize that maladaptive schemas are internalized at a very young age, subsequently making effective coping strategies in adulthood difficult to establish. Schema theorists organize early maladaptive schemas into five different domains: disconnection and rejection; impaired autonomy and performance; impaired limits; other-directedness; and overvigilance and inhibition. Table 2 depicts operational definitions and underlying cognitive themes for each of these schema domains. These schemas are thought to describe the motivations and cognitive processes that underlie choices to engage in risky behaviors.

Table 2 Classification and operational definitions for schema domains (Young et al. 2003)
Full size table

At the theoretical level, Beck et al. (1993) hypothesize that ineffective schematic functioning, marked by faulty information processing, constitutes a vulnerability to harmful activities. Studies specifically targeting the role of maladaptive schema in the engagement of risk behaviors are rare; however, schema theory (Young et al. 2003) is consistent with other models of risk-taking behavior. For example, according to Crick and Dodge’s (1994) social-information processing model of delinquency, externalized behaviors are generated through a rapid series of biased cognitive processes. Specifically, individuals who report engaging in delinquent acts (e.g., substance abuse, law-breaking) are likely to demonstrate disruptions in healthy schematic functioning as indicated by impairments in encoding relevant cues about context and emotion, misinterpreting other people’s intentions, and failing to select problem-solving strategies to perceived interpersonal challenges (Dodge et al. 2013). It is argued that these cognitive difficulties contribute to impulsive decision making and positive expectations regarding risk behavior, two concepts known to promote risk-taking (Sultan et al. 2002). As theorized, distortions in schema functioning have been shown to predict deviance and conduct problems across development (Dodge et al. 2008).

Crick and Dodge’s (1994) social-information processing model offers a pathway by which schema processing can be linked to delinquency, but it was not developed to explain susceptibility to a variety of types of risk-taking behavior in emerging adults. Emerging adulthood is characterized by substantial shifts in autonomy and responsibility that necessitate the cultivation of a diverse range of effective problem-solving and coping strategies (Johnson et al. 2010). As a result, impairments in specific schema processes that interfere with the identification and implementation of coping resources may be uniquely predictive of risk-taking in this population. Maladaptive schemas are considered toxic because they trigger memories, thoughts, and sensations that perpetuate feelings of defectiveness, inadequacy, and/or incompetence in managing adversity. When schemas are activated, individuals will go to great lengths to escape emotional upheaval, possibly engaging in a wide range of risky behaviors (Young et al. 2003). For example, substance use may be seen as a suitable means of self-soothing, increasing self-control, and combating negative mood states resulting from schema activation (Ball 2007; Beck et al. 1993).

Evidence supporting the proposed associations between maladaptive schemas and risky behaviors is emerging, particularly in clinical samples. Young women and men seeking substance abuse treatment endorsed higher reports of maladaptive schema traits compared to non-clinical samples of young adults (Shorey et al. 2013, 2014). Maladaptive schemas were also found to have significant relationships with gambling problems in a sample of alcohol-dependent men seeking treatment (Shorey et al. 2012a). Of further relevance to the current study, the rate at which young adults suffering from substance abuse difficulties report traits associated with maladaptive schemas varied by gender (Shorey et al. 2012b), emphasizing the importance of understanding gender differences in the associations between particular maladaptive schema and a variety of types of risk-taking behavior.

While theorists have constructed cognitive vulnerability models for a wide spectrum of emotional disorders including depression (Alloy and Riskind 2006), there is scarcity of cognitive models that explain the development and exacerbation of other types of risk behaviors. Empirical evidence linking particular cognitive deficits to specific risk behaviors has also been lacking. Given findings of gender-specific expressions of both particular risk behaviors and maladaptive schemas, a separate focus on men and women is warranted. Consequently, the primary purpose of the current study was to investigate gender-specific associations among multiple indices of early maladaptive schemas and a variety of risky behaviors in a sample of college-attending emerging adults.

Consistent with research indicating that behavioral intentions to engage in risk behaviors were related to self-reported risk-taking behavior 6 months later (Combs-Lane and Smith 2002), the current research focuses on behavioral intentions to engage in risky behaviors rather than self-reports of past transgressions or the perceived risks/benefits associated with particular behaviors. Given that different risky behaviors are rooted in unique sociocultural processes and elicit unique responses from others (Klibert et al. 2011), the current study measures college students’ anticipated involvement in five different types of risky behaviors already shown to be associated with development in emerging adulthood (academic/work, sexual, illicit drug use, heavy drinking, and illegal/aggressive; Fromme et al. 1997), which represents a strength relative to previous studies. Five domains of maladaptive schemas were assessed in the current study (disconnection and rejection, impaired autonomy, impaired limits, other-directedness, and overvigilance and inhibition) in order to facilitate a more fine grained analysis of maladaptive cognitions associated with a particular type of risky behavior.

Hypotheses

We hypothesized men would report greater anticipated engagement in all types of risky behaviors than women. We expected that the combination of the five early maladaptive schema domains (disconnection and rejection, impaired autonomy, impaired limits, other-directedness, and overvigilance and inhibition) would account for variance in anticipated engagement in each of the five distinct domains of risky behavior (problematic academic and work, risky sexual, illicit drug use, heavy drinking, and illegal and aggressive behaviors). As an exploratory measure, we examined whether or not different schema domains would individually predict variance in risk-taking behaviors for women versus men. Given the current literature, we could not explicate which schema domains would predict variance in risk-behavior subtypes across gender. However, given gender variation in the manifestation and expression of risk-taking behavior, we expected different schema domains would serve as individual predictors for risk-behavior subtypes for women versus men.

Materials and Methods

Participants

Participants were 484 undergraduate students taking introductory psychology classes at a mid-sized southeastern university. Of the 484 participants, 341 (70.5%) were women and 143 (29.5%) were men. The preponderance of women is consistent with current trends in sampling procedures for college-attending adults in southeastern areas of the U.S. (Klibert et al. 2011). Most participants identified as White/European American or Black/African American. A complete breakdown of the demographic profile of the sample is depicted in Table 3. These demographics are similar to Introductory Psychology participation pool demographics in the southeastern region of the U.S. Participation was voluntary and the students were compensated with class credit.

Table 3 Demographics, means, standard deviations, and minimum and maximum scores for early maladaptive schemas and risky behaviors for college women and men
Full size table

Measures

Young Schema Questionnaire-Short Form 3rd Revision

The Young Schema Questionnaire-Short Form-3rd Revision (YSQ-S3; Young 2005) is a 90-item self-report measure of five early maladaptive schema domains, each containing a number of individually clustered cognitive themes (see Table 2). The YSQ-S3 uses a 6-point Likert-type rating scale. Participants respond to items based on the previous 12 months. The possible range of scores for each schema domain is as follows: disconnection and rejection (25–150), impaired autonomy (20–120), impaired limits (10–60), other-directedness (15–90), and overvigilance and inhibition (20–120). Higher overall scores suggest greater adherence to early maladaptive schemas. In a previous study, Stumblingbear et al. (2007) reported that all five schema domain scores of the YSQ-S3 had adequate to excellent internal consistency, with alpha coefficients ranging from 0.75 to 0.92. Additionally, the five schema domain scores have been found to have excellent convergent validity as evidenced by moderate correlations with trait measures of depression, anger, and anxiety in emerging adults (Stumblingbear et al. 2007). Likewise, the coefficient alphas for the five schemas in the current study were adequate to excellent: disconnection and rejection (α = 0.94), impaired autonomy (α = 0.89), impaired limits (α = 0.79), other-directedness (α = 0.82), and overvigilance and inhibition (α = 0.88).

Cognitive Appraisal of Risky Events

The Cognitive Appraisal of Risky Events (CARE; Fromme et al. 1997) questionnaire is a 30-item self-report instrument designed to assess people’ beliefs about the consequences of being involved in risky activities and how often they expect to be involved in these activities. For the purposes of this study, only the Expected Involvement (EI) subscales of the CARE were used. The EI domain measures individuals’ expected participation in a broad range of risky activities across the subsequent 6 months. The six subscales of this measure assess illicit drug use (“Mixing drugs and alcohol.”), aggressive and illegal behaviors (“Getting into a fight or argument.”), risky sexual activities (“Sex without protection against pregnancy.”), heavy drinking (“Playing drinking games.”), high-risk sports (“Rock or mountain climbing.”), and problematic academic and work behaviors (“Leaving tasks or assignments for the last minute.”). The CARE prompts participants to rate the likelihood of their involvement in these activities on a 7-point Likert scale from one (not at all likely) to seven (extremely likely). Higher scores indicate a greater probability of engaging in risky activities in the near future. The EI subscales have been found to have good internal consistency with scores ranging from 0.78 to 0.85 (Fromme et al. 1997), with the exception of the extreme sports subscale (α = 0.64). The EI subscales have also been found to have excellent construct validity as evidenced by moderate correlations with measures of impulsivity and sensation seeking (Fromme et al. 1997). Only five of the six risky behavior subscales were utilized in the current study. The extreme sports subscale was excluded due to an inadequate reliability score (α = 0.62). The coefficient alphas for the five other risky behavior subscales were good: problematic academic and work behaviors (AWB, α = 0.85), risky sexual behaviors (SEX, α = 0.79), illicit drug behaviors (DRUG, α = 0.84), heavy drinking behaviors (DRINK, α = 0.87), and illegal and aggressive behaviors (AGG/ILL, α = 0.86).

Procedure

IRB approval was obtained prior to the commencement of this study and ethical procedures were followed throughout. Using an online platform, data were collected anonymously from students enrolled in undergraduate psychology courses. All participants freely gave their informed consent to participate as a volunteer in this study. Most students completed the survey in 30–45 min. All participants were thoroughly debriefed at the conclusion of this study. As a part of the debriefing process, students were made aware of free and low cost mental and physical health resources in the area.

Results

Means, standard deviations, and minimum/maximum scores for each risk-behavior subtype and schema domain score are displayed in Table 3. Using P–P plots and z-transformation procedures, we evaluated the distribution of scores for all variables. With the exception of the overvigilance and inhibition schema domain score, a majority of the variables were non-normally distributed (positively skewed). The results suggest that college-attending adults anticipated engaging in few risky behaviors and report few cognitive styles consistent with maladaptive schemas. However, statistical procedures using an F-ratio in large samples are generally robust in terms of minimizing Type I errors rates resulting from non-normal distributions (Stevens 2002). Therefore, despite the skewness of our data, our findings should be interpretable as meaningful contributions to the literature.

A multivariate analysis of variance (MANOVA) was conducted to examine gender differences in the behavioral intention to engage in each of the five risky behaviors. The MANOVA revealed a significant overall effect for gender, F(1, 482) = 19.23, p < .01, η2 = 0.17, across the five risky behavior subscale scores. Follow-up ANOVA’s revealed statistically significant gender differences in expected engagement in problematic academic and work behaviors, F(1, 482) = 4.14, p < .05, η2 = 0.01; risky sexual behaviors, F(1, 482) = 92.94, p < .01, η2 = 0.17; illicit drug use behaviors, F(1, 482) = 11.36, p < .01, η2 = 0.02; heavy drinking behaviors, F(1, 482) = 16.31, p < .01, η2 = 0.03; and illegal and aggressive behaviors, F(1, 482) = 24.07, p < .01, η2 = 0.05. As expected, men anticipated greater involvement in problematic academic and work behaviors, risky sexual behaviors, illicit drug use, heavy drinking, and illegal and aggressive behaviors than did women.

Given statistically significant gender differences in the five risky behavior subscales, separate correlations were conducted to examine the interrelationships among risky behavior types and maladaptive schema domains for women versus men. The majority of maladaptive schemas were positively related to anticipated engagement in all five domains of risk behaviors for both women and men; however, a few gender-specific patterns were noted. Specifically, all five schemas were significantly correlated with women’s binge drinking behavior, whereas none of the schemas were significantly correlated with men’s binge drinking behavior. Similarly, compared to men, there was a more consistent pattern of association between maladaptive schemas and women’s expected engagement in academically risky behavior. Bivariate correlations for women and men are displayed in Table 4.

Table 4 Correlations among anticipated engagement in risky behaviors and early maladaptive schemas for college women and college men
Full size table

Separate multiple regression analyses were conducted for women and men to examine the extent to which maladaptive schemas contributed variance to anticipated engagement in each of the five types of risky behaviors (see Tables 5, 6). The five schema domain scores accounted for a significantly greater amount of variance in anticipated engagement in academic risk behaviors for women, R2 = 0.20, p < .01; F(5, 335) = 16.21, p < .01 compared to men, R2 = 0.11, p < .01; F(5, 137) = 3.21, p < .01. However, an examination of the beta weights revealed impaired limits schemas and overvigilance and inhibition schemas predicted unique variance in the models for anticipated engagement in academic risk behaviors for both women and men.

Table 5 Regression analysis using schemas to predict risky behaviors for women
Full size table
Table 6 Regression analysis using schema to predict risky behaviors for men
Full size table

In contrast, the five schemas accounted for a substantially greater amount of variance in anticipated engagement in sexual risk behaviors for men, R2 = 0.24, p < .01; F(5, 137) = 8.48, p < .01, compared to women, R2 = 0.09, p < .01; F(5, 335) = 6.57, p < .01. Moreover, a different pattern of schemas emerged as significant predictors for anticipated engagement in sexual risky behaviors for women versus men. Specifically, disconnection and rejection schemas and impaired limits schemas were significant individual predictors for women, whereas only disconnection and rejection schemas were significant predictors for men.

Consistent with previously noted correlations, the five schema domain scores accounted for 9% of the variance in anticipated engagement in heavy drinking behaviors for women, F(5, 335) = 6.66, p < .01, but did not significantly account for variance in anticipated engagement in heavy drinking behaviors for men. In the model for women, disconnection and rejection schemas and impaired limits schemas were significant predictors for anticipated engagement in heavy drinking behaviors. An examination of beta weights for anticipated engagement illicit drug use revealed that impaired autonomy schemas were significant individual predictors for women. None of the schema domains significantly contributed to variance in anticipated engagement illicit drug use for men.

Finally, the five schema domain scores accounted for a larger amount of variance in anticipated engagement in illegal and aggressive behaviors for women, R2 = 0.20, p < .01; F(5, 335) = 7.30, p < .01, compared to men, R2 = 0.09, p < .01; F(5, 137) = 2.61, p < .01. Both disconnection and rejection and impaired limits schemas were significant predictors for women. None of the schema domains contributed to the variance in this risk-behavior index for men.

Discussion

One aim of the current study was to examine gender differences in anticipated engagement in five types of risky behaviors. Our results indicated that men anticipated greater involvement in all five of the measured risky behaviors compared to women. Although these findings are consistent with previous research involving emerging adults (i.e., Caetano and Cunradi 2002; Nelson et al. 2008), gender differences in anticipated engagement in problematic academic, illicit drug use, heavy drinking, and illegal/aggressive behaviors had very small effect sizes (η2 = 0.01–0.05), which suggests that gender norms may be shifting such that college women and men are becoming more similar in their views regarding engagement in risky behaviors.

In contrast, gender differences in self-reports of anticipated engagement in risky sexual behavior remained large (men’s mean reports were approximately twice that of women’s). This finding suggests that there is a gender difference in emerging adults’ views of future risky sexual encounters; men appear to anticipate engaging in more unprotected, casual, and multiple-partner sexual relationships compared to women. The results are consistent with Cubbins and Tanfer’s (2000) findings and are likely related to social phenomena regarding men’s perceptions of sex. For example, when contemplating whether to engage in casual sexual relationships, men may overemphasize pleasurable consequences while downplaying potential costs to their emotional and physical health. Another possibility is that women may be less willing to report anticipated involvement in risky sexual behaviors compared to men, despite engaging in these risky sexual behaviors at similar rates, which is consistent with numerous theories related to gender-specific social desirability concerns (Perlini and Boychuk 2006). Future research is needed to disentangle various possibilities.

A second aim of the current study was to examine the relationships among the five types of early maladaptive schemas and women and men’s anticipated engagement in the five different types of risky behaviors. As predicted on the basis of schema theory, univariate results for both women and men indicated that the majority of the maladaptive schemas were positively associated with all five indices of anticipated risky behaviors. However, a few gender-specific results did emerge. Women’s anticipated engagement in risky academic behaviors was correlated with all five schemas, whereas a less consistent pattern was observed for men. Similarly, all five schemas were associated with women’s anticipated engagement in heavy drinking, whereas none of the schemas were significantly correlated with men’s potential future heavy drinking. Further study of cognitions underlying college men’s decisions to drink heavily is needed.

In order to determine which schemas accounted for unique variance within the five domains of anticipated engagement in risky behaviors, regression models were conducted separately for college women and men. Several important gender-specific findings emerged. First, the amount of variance accounted for by maladaptive schemas differed considerably between women and men. The combination of the five schemas was associated with women’s reports of anticipated engagement in problematic academic or work (20% for women vs. 11% for men) and illegal and aggressive (20% for women vs. 9% for men) behaviors. Conversely, the schema domains accounted for a substantial amount of variance in men’s reports of anticipated engagement in future risky sexual behaviors as compared to women (24% for men vs. 9% for women). Furthermore, the schemas significantly accounted for 9% of women’s anticipated engagement in heavy drinking behavior but failed to significantly account for variance in college men’s anticipated engagement in these same behaviors. Different cognitive processes appear to contribute in varying amounts to college women and men’s anticipated risky behavior.

A more thorough examination of the regression models highlights a handful of notable gender-specific trends concerning individual schema domains. First, higher levels of the impaired limits schema accounted for unique variance in four of the five risky behaviors types (academic/work risk, sexual risk, heavy drinking, and illegal/aggressive behaviors) for women. The finding suggests that women who expect to engage in future risky behaviors are likely to have difficulties maintaining sufficient self-control and self-discipline; they may also struggle to overcome low levels of frustration. Impulsivity and low frustration tolerance components of the impaired limits schemas may be particularly useful in clarifying the decision to engage in risky behaviors. Screening for impulsivity and low frustration tolerance beliefs may help to identify young women who are at risk for engaging in problem academic, sexual, heavy drinking, and illegal/aggressive behaviors. Likewise, prevention and intervention efforts aimed at women with low frustration tolerance and high impulsivity may be worthwhile. Consistent with Schema Therapy, clinicians can implement experiential strategies (e.g., imagery) to mitigate the activation of impaired limits schemas by employing behavioral-pattern breaking tactics (e.g., setting up contingencies) to disrupt coping efforts known to reinforce impaired limits schema perpetuation and model more adaptive means of obtaining unmet emotional needs associated with impaired limits schemas.

Additional gender discrepant findings emerged in the current study. Disconnection and rejection schemas were significant predictors for women’s anticipated engagement in heavy drinking and aggressive/illegal behaviors. These schemas were not retained in the regression models for men. This finding extends the current literature by highlighting cognitive processes associated with instability and unavailability of care and support as robust predictors in college women’s decisions to engage in alcohol abuse and illegal transgressions. Clinical assessment of disconnection and rejection schemas may be a fruitful target when attempting to identify and prevent women from engaging in these dangerous and destructive behaviors, which is consistent with social-cognitive models of risk prevention (Dodge et al. 2013). Specific thoughts, like “Someone is always going to betray me” and “No one takes the time to consider my feelings,” may contribute to inaccurate responses to threatening stimuli in a way that increases the likelihood to engage in antisocial and substance use behaviors. As a result, it may be important for clinicians to help women who report higher levels of disconnection and rejection schemas to reconstruct social cues in ways that promote health-focused behaviors as a primary prevention strategy.

Gender-congruent findings are also worth noting. Specifically, disconnection and rejection schemas accounted for significance variance in self-reports of future engagement in risky sexual behaviors for both women and men. Thus, women and men who feel personally defective and socially disconnected may exhibit a greater number of reckless and unsafe sexual behaviors. Individuals high in disconnection and rejection schemas are likely to engage maladaptive coping efforts when their core emotional needs are frustrated (Young et al. 2003). These individuals may consider engagement in risky sexual practices (e.g., paying for sex, seeking out multiple partners) as a means to increase positive personal contact, affection, and validation. This position is consistent with a number findings in the current literature (e.g., Bancroft and Vukadinovic 2004). However, it is important that future research examine the coping tendencies of individuals high in disconnection and rejection schemas to obtain a firmer understanding of women and men’s choice to engage in risky sexual behaviors.

Finally, similar findings emerged for women and men with regard to risky academic and work behaviors, with impaired limits and overvigilance and inhibition schemas serving as unique predictors in the model. However, results did not completely conform to our expectations. An inverse predicted relationship suggested higher levels of overvigilance and inhibition schemas were predictive of lower levels of anticipated engagement in risky academic and work behaviors. This finding seems paradoxical, especially considering the bivariate associations between these two constructs was positive. Upon further examination, this result appears to be an example of a suppressor effect (Paulhus et al. 2004), where the strength and direction of a relationship is altered after the inclusion of other correlated variables in the model. After accounting for the shared variance among the five schema variables, the residual effects for overvigilance and inhibition schemas explained a very small, yet statistically significant, proportion of variance in anticipated engagement in risky academic and work behaviors for women and men. However, considering the ambiguity in describing the nature of these residual effects and the fact they accounted for minimal variance, it is likely that overvigilance and inhibition schemas hold little practical significance as predictors for risky academic and work behaviors.

Limitations of the current study are worth noting. Most notably, data were collected using a cross-sectional design, which precludes causal associations between maladaptive schemas and risky behaviors. Future studies should implement prospective designs that consider the interplay between schemas and risky behaviors across time. Participants in the current sample anticipated relatively low levels of engagement in risky behaviors. Future research could consider the influence of early maladaptive schemas in the risk-related decision processes of high-risk individuals as a means of validating the generalizability and stability of our findings. The current study did not measure actual engagement in risk behaviors, which may limit interpretations regarding the association between schema processing and risk-taking. In order to clarify the contribution of cognitive processes in predicting poorer behavioral health outcomes, future research may need to evaluate the associations between schema processing and a history of risk-taking. Finally, given that data regarding anticipated engagement in risky behaviors were obtained via self-report, shared method variance may explain some of the obtained associations. Measuring students’ expectancies for the future may have also activated additional concerns about social desirability. Women may have been particularly reluctant to admit to the possibility of engaging in future risky sexual activities. As such, future research should re-examine these relationships while controlling for social desirability factors.

Despite limitations, the findings from the current study offer some support for the position that certain types of early maladaptive schemas may influence decisions to engage in future risky behaviors for emerging women and men. The results highlight the importance of considering each type of risky behavior independently and emphasize the importance of considering gender differences in emerging adults’ decisions to engage in risky behaviors. Prevention and intervention activities are likely to be more effective if they consider the interactions among gender and schema processes in the decision to engage in specific risk behaviors. In summary, the findings presented in the current study provide preliminary evidence for the associations between dysfunctional cognitive processes and expected engagement in risk-taking and highlight gender-specific insights that may guide risk prevention and intervention efforts for college-attending adults.