Psychological factors associated with adolescent sexual behavior : the role of self-esteem and self-efficacy

Received on: Oct. 22, 2019. Approved on: Feb. 24, 2021. Published on: Agust. 10, 2021. Abstract: Sexually transmitted infections are increasing in Brazilian adolescents and youth. The aim of this study was to investigate the association of self-esteem, self-efficacy, and other psychosocial variables with condom use behavior and sexual debut in a sample of adolescents from Santa Maria-RS. Data was collected with 452 adolescents (57% girls), mean age of 15.9 years (SD = 1.4), using the Brazilian Youth Questionnaire – Phase II. Condom use behavior was not associated with self-esteem or self-efficacy in the logistic regression model, contrary to the main hypothesis. General self-efficacy was positively associated with sexual debut, while religiosity was negatively associated with this outcome. Family support was associated with older age at sexual debut. The findings support the importance of analyzing different sexual behaviors separately, as they may have distinct predictors. Studies should use specific measures of self-efficacy when studying sexual behavior and measure how participants value different sexual behaviors.

Palabras clave: condones, adolescencia, autoeficacia, autoestima, religiosidad Adolescence, which spans from puberty until the onset of adulthood, is characterized by the adolescent's differentiation from their family and greater approximation with their peers (Huang et al., 2012). It is common for adolescents to perform exploratory behaviors in the new environments that they are experiencing, searching for novel feelings or peer approval (Cruzeiro et al., 2010).
In this setting, the performance of risk behaviors is a concern for families and institutions. Risk behaviors are those that increase the probability of adverse outcomes, such as injuries, mental illness or sexually transmitted infections (STIs). The use of psychoactive substances, antisocial behavior or having unprotected sex are some examples of risk behaviors that may occur during adolescence (Zappe & Dell'Aglio, 2016b). However, risk behavior performance differs between adolescents. Studies have shown that psychological constructs such as self-esteem and self-efficacy may help to understand why some adolescents are more inclined to the performance of risk behaviors than others (Bermúdez et al., 2019;Danielson et al., 2014;Louwet et al., 2018;Xu et al., 2017).  (Cruzeiro et al., 2010).
The inconsistent use of condoms by Brazilian adolescents appears to be associated with several social and demographic variables, such as sex, parent education, income and skin color (Woolley & Macinko, 2019). The findings are inconclusive for sex differences, with studies showing similar rates of condom use between boys and girls (Neto & Cerqueira-Santos, 2012;Woolley & Macinko, 2019), or small effects (Cruzeiro et al., 2010). However, sex may mediate the effect of other demographic variables. For boys, having black or mixed skin color and being in the lower-income bracket accounted for a 25% decrease chance to have used a condom during their last intercourse. As for girls, a similar effect was only found for the difference in condom use between mixed and white girls (Woolley & Macinko, 2019).
While condom use is the main preventive behavior against STIs and unplanned pregnancies, it is also valuable to understand what drives adolescents to have their first sexual experiences, as early sexual debut is often linked to smaller rates of condom use (Alves et al., 2015;Gräf et al., 2020;Moreira et al., 2018;Paiva et al., 2020).
During adolescence, the changing dynamics of family and community relationships, as well as the construction of a self-identity may play a part in increasing the occurrence of sexual risk behaviors. Factors such as religiosity, family and community support, and dealing with stressful environments have been linked with sexual risk behaviors (Estrada et al., 2017;Paiva et al., 2020;Zappe & Dell'Aglio, 2016a;2016b). Religiosity has been shown to correlate with sexual debut, but had no association with a general sexual risk behavior measurement (Cerqueira-Santos & Koller, 2016). In a sample of 925 adolescents, specific religions were associated with delayed sexual debut, while adolescents that reported no affiliation also reported earlier sexual debut.
As for condom use, attending religious service was associated with less frequent condom use at sexual debut, among sexually active adolescents (Paiva et al., 2020). Adolescents that report better relationships with their family, with openness to discuss issues have lesser chances of engaging in sexual risk behaviors, and the same applies to community support. Being exposed to stress generating events, such as violence and or parental divorce, has also been linked to increased sexual risk behavior in adolescents, as well as with other risk behaviors (Zappe & Dell'Aglio, 2016a;2016b). In fact, risk behaviors in adolescence tend to occur in addition to each other, such that drug use and sexual risk behavior are also entwined in this developmental phase (Huang et al., 2012;Zappe et al., 2018;Zappe & Dell'Aglio, 2016b). Even in instances in which drug usage is not directly associated with unprotected sex, reporting sexual intercourse under the influence of drugs has been associated with condomless sex in adolescents (Nehl et al., 2016).
Being that adolescence is a period of development of one's identity, it is likely that valuing oneself more should prevent the performance of harmful behaviors. Therefore, it would be expected for adolescents with high self-esteem to be less prone to performing risk behaviors, as they would value their health more than their peers would. However, it is unclear if there is a specific link between self-esteem and condom use behavior. In a study with young South African women, higher self-esteem correlated with a better sense of future and relationship control, and with a lower probability of HIV diagnosis (Louw et al., 2018). Still, self-esteem was not associated with reported condom use behavior.
A similar finding was reported in a sample of 159 Brazilian adolescents, in which self-esteem did not account for changes in condom use behavior.
It did associate with other sexual behaviors, as adolescents with higher self-esteem reported fewer instances of sexual intercourse under the influence of drugs (Neto & Cerqueira-Santos, 2012). A latent-class analysis with 701 adolescent girls split the sample between high-risk and lowrisk clusters, based on previous STI diagnoses, the number of sexual partners, history of having sex under the influence of drugs, and condom use behavior (Danielson et al., 2014). The high-risk group presented with lower self-esteem when compared to the low-risk group, as well as with lower sex refusal self-efficacy and more fear of condom negotiation. The high-risk group also reported more instances of sexual violence and other trauma, which the authors claim may also explain lower levels of self-esteem. Self-esteem was also positively associated with condom use in a sample of 919 refugee adolescent girls in Ethiopia, despite no connection between selfesteem and sexual violence in the multivariate analysis (Bermúdez et al., 2019). Overall, the links between self-esteem and sexual health outcomes vary substantially between studies. Meta-analytic findings show that there is a consistent, small effect of self-esteem on positive sexual health outcomes. However, there also seems to be some publication bias, with negative or null findings more likely to be found on unpublished research (Sakaluk et al., 2020). These results highlight the need for a better understanding of whether selfesteem is related to the performance of sexual risk behavior during adolescence, and how does it affect condom use behavior in particular.
One possible explanation for the discrepancy of the effects of self-esteem in predicting condom use behavior is that condom use is not a behavior that is completely under an individual's control.
Namely, in order to use a condom during sexual intercourse, all sexual partners must agree to it.
While being assertive during condom negotiation is a challenge that persists after adolescence, it can be even more harmful in this period due to the novelty of sexual relations and pressure to conform to peer norms. Based on this rationale, the construct of self-efficacy could explain why adolescents may have a hard time using condoms even if they would like to be protected. Selfefficacy measures how individuals perceive their abilities to succeed in a task even in the face of adversities. Transposing this idea to condom use, self-efficacy would be the perceived ability to always use condoms even when there are obstacles to the behavior, such as having an unwilling partner or having to postpone the intercourse in order to acquire condoms first.
The evidence for sex-specific self-efficacy shows that it is positively correlated with consistent condom use. Palacios (2019) reports that sexual self-efficacy was a significant predictor of sexual behavior in a sample of 1012 Mexican adolescents. Xu et al. (2017) propose that condom use selfefficacy is crucial to understanding condom use behavior, given that it modulates the effect of other variables, such as perceived costs and benefits of using condoms, in a study with 1012 Bahamian students. The study measured condom use in the last six months with three levels: "Never used", "Sometimes used", "Always used". It was found adolescents with higher self-efficacy were more likely to report having always used condoms, while others had a similar distribution between "Sometimes" and "Always" categories (Xu et al., 2017). Sousa et al. (2017), in a study to adapt the condom use self-efficacy scale with Brazilian adolescents and young adults, found that condom use self-efficacy was positively associated with condom use behavior, even for people in stable relationships. These findings suggest that self-efficacy is a possible pathway for changing condom use behavior in adolescence.
This article aimed to investigate the association between psychological (self-esteem, general self-efficacy, religiosity, future expectations) and social context variables (perceived family support, perceived community support) with adolescent sexual behavior. Specifically, we examined the association of these factors with the occurrence of sexual debut, the age at sexual debut, and the frequency of condom use in the sample, controlling for social and demographic variables and other risk behaviors, such as drug use behavior. It is hypothesized that higher self-efficacy and self-esteem will be associated with more frequent use of condoms in the last year and older age at sexual debut when controlling for demographic variables, family and community support, religiosity, life stress and drug use in the past month. It is also expected that, controlling for other predictors, higher religiosity and family support will be associated with later sexual debut, while drug use will be associated with earlier sexual debut.

Participants
The data were collected with students from public schools of [Identifying information retracted].
The 15 institutions were randomly selected with a block draw. Data were collected with 535 students. Inclusion criteria were as follows: being between 13 and 19 years old, having answered in a correct manner the questions on (1) sex, (2) age, (4) race/color and (42) sexual debut. After the exclusion of incomplete or older cases, the final database was made up of 452 participants, 257 girls and 195 boys. The mean age was 15.9 years (SD = 1.4). Self-declared white adolescents were the majority of the sample (68.6%), followed by mixed (21.7%) and black (6.9%). Most participants were single at the time of data collection (88.7%). Social and demographic data from questions 1 (sex), 2 (age) and 4 (race/color) will be used in the analysis. Questions pertaining to income will not be used due to a high frequency of missing values.

Measures
The independent variables sexual debut, age at sexual debut, and frequency of condom use in the last year were coded from the responses to items 42 and 45. For question 45, which measures condom use frequency in the last year with four levels: "Never", "A few times", "Often, but not always", and "Always", responses were dichotomized so that "Always" (1) was considered consistent use and the other responses were considered inconsistent use (0). Unless otherwise stated, all the following scales had a response scale ranging from 1 to 5. The Rosenberg Self-Esteem Scale (Sbicigo et al., 2010) was the item 74 of the questionnaire.
Made up of ten items, had α = .87 in this sample.
This scale contains inverted items, such as "Sometimes, I feel useless". The General Self-Efficacy Scale (Sbicigo et al., 2012)

Data analysis
The RStudio software v. 1.1463, R version 3.6.1, was used to manage the data and run the descriptive and inferential analysis. To test the hypotheses, two logistic regression models were analyzed, with the following dependent variables: sexual debut and condom use frequency in the last year. To analyze the predictors of sexual debut age, a linear multiple regression was run.
For the models of sexual debut age and condom use, only the data from those adolescents that reported sexual intercourse was used (n = 237).
In the regression model for sexual debut, the first step aggregated age, sex, race/color, life stressors and substance use as predictors.
Psychological variables (self-efficacy, self-esteem, religiosity, future expectations, perceived family and community support) were inserted in step 2, simultaneously. The same procedure was adopted in the analysis for the age of sexual debut, with the exclusion of age as a predictor.
The model for condom use in the last year used age, sex, race/color, substance use, life stressors and age of sexual debut as predictors in the first step. Again, psychological variables were inserted simultaneously in the second step. All analyses used an alpha of .05. Table 1 presents descriptive data of the sample and between-group comparisons based on sexual debut and condom use during the previous year.

Results
It shows the sample was evenly split between adolescents that have had sex or not and also between those that have used condoms consistently or not during the previous year.  The logistic regression analysis of sexual debut showed effects of age and drug use in the first step. Older adolescents were 70% more likely to have had sex, and those that had more instances of drug use in the last month had over three times the chance to have had sex. With the addition of psychological predictors in step 2, age was still a significant predictor, while drug use ceased to be. Self-efficacy was positively associated with sexually active adolescents, while participants that were more religious were 44% less likely to report having had sexual intercourse. The model is represented in Table 3. Notes. N = 452, a Nagelkerke R² = .274, b Nagelkerke R² = .329, * p < .05 The linear regression on age at sexual debut resulted in one significant predictor in the first step. Adolescents that self-declared mixed were more likely to be younger than white adolescents were at sexual debut. This effect continued in the second step, in addition to an effect from perceived family support. Adolescents that reported better relationships with their families were more likely to be older at sexual debut. This model is represented in Table 4. Notes. N = 230, a R² (adjusted) = .079 (.050), b R² (adjusted) = .124 (.071), * p < .05

Discussion
The main goal of this study was to investigate the association of psychological and social context variables with condom use behavior in adolescents. The results did not support the main hypothesis that self-efficacy and self-esteem many other findings that suggest that earlier sexual debut places adolescents at higher sexual risk (Alves et al., 2015;Gräf et al., 2020;Moreira et al. 2018;Paiva et al., 2020). One possible explanation is that Brazilian adolescents experience a culture with prevalent sexist norms that favor performance of sexual risk behaviors as a marker for masculinity or a token of trust and love (Garcia & Souza, 2010).
When adolescents have unprotected sex for the first time, they may be less afraid of negative consequences of unprotected sex if they "get away" with it. An analysis of condom use behavior in adolescents from Canoas-RS showed that sexual experience was correlated with more positive attitudes towards condomless sex (Chinazzo et al., 2014). This would also explain why self-efficacy alone does not prevent risk sexual behavior. If adolescents perceive no benefit in the use of condoms, they will not use them regardless of how well they believe they would be able to. such as condom use self-efficacy (Palacios, 2019;Sousa et al., 2017;Xu et al., 2017) or sex-refusal self-efficacy (Danielson et al., 2014). Since causality cannot be inferred from the design of this study, it is not possible to distinguish if adolescents in this sample feel more capable after having their sexual debut, as proposed by Ramiro et al., (2013), or if they have their debut before their peers in part because of their higher self-efficacy.
These hypotheses highlight an important caveat of using general self-efficacy as a predictor of