Sexuality-related harms, such as unwanted pregnancies, sexually transmitted diseases (STD), and regretted sexual intercourses are among the most frequent threats of adolescents’ health and mental well-being (Bearak et al., 2020; Taylor & Wi, 2019; Wight et al., 2000). Sexual health education programs aim to promote healthy behaviors and to prevent these risks (Centers for Disease Control & Prevention, 2020). Definitions of sex, however, may be different among participants of these programs. This should be considered by health education professionals when planning and providing sexual education sessions and when evaluating research data.

The main focus of studies published on sexual definitions is whether a certain act is considered sex or not (Horowitz & Bedford, 2017; Mehta et al., 2011) or whether a certain type of sexual behavior leads to “loss of virginity” or breaking of “abstinence” (Barnett et al., 2017; Byers et al., 2009). So far, no widely accepted tool exists to assess the issue, most of the research groups either developed their own questionnaire or used a version of the Sexual Definitions Survey (SDS) (Sanders & Reinisch, 1999). Studies using instruments based on the SDS present to respondents a spectrum of different intimate behaviors usually starting from kissing and ending with penile-vaginal intercourse (Horowitz & Bedford, 2017; Horowitz & Spicer, 2013). Most studies published to date primarily focus on English-speaking young adults (e.g., college students). Studies conducted in the USA indicate that 99.5% and 81% of university students consider penile-vaginal and penile-anal intercourse as sex, respectively (Sanders & Reinisch, 1999), while one-third of them also considered oral sex as sex. Even fewer university students perceived the touching of genitals or nipples and kissing (2%) as sex. Other studies conducted in Canada, UK, and USA corroborate these results (Barnett et al., 2017; Byers et al., 2009; Gute et al., 2008; Pitts & Rahman, 2001; Sanders et al., 2010) clearly showing the “hierarchy” of sexual behaviors described by Horowitz and Spicer (2013). The ambiguity over oral and manual stimulation of other’s genitals was also identified by two other studies in which scaled choices were used instead of dichotomous answers (Horowitz & Spicer, 2013; Sewell & Strassberg, 2015). In these studies, penile-vaginal intercourse was considered “definitely sex” by 98% of the participants, while kissing was mostly believed as “definitely not sex.” As for oral-genital contact or touching of genitals, non-extreme answers were observed more often, ranging from 46 to 72%.

To enlighten the ambiguity of sexual definitions, Byers et al. (2009) divided intimate behaviors into three groups depending on the number of genitals involved (Byers et al., 2009). Bidirectional intercourse happens when both participants use their genital organs, e.g., vaginal and anal penetration. Unidirectional intimate behavior includes all types of intimate settings where one participant is involved with their genitals, while the partner uses non-genital organs (e.g., hand, mouth). The third category is non-genital contact that includes all other intimate behaviors, such as touching of nipples and deep kissing. Unlike the other two categories, semantic variability is often detected for unidirectional intimate behaviors, but certain internal consistencies were observed as well; for instance, participants who consider oral stimulation as sex are more likely to perceive manual stimulation as sex as well (Horowitz & Bedford, 2017).

Divergence of sexual definitions seems to be influenced by gender, age, and previous personal sexual experience (Barnett et al., 2017; Bersamin et al., 2007; Gute et al., 2008; Pitts & Rahman, 2001; Richters & Song, 1999; Sanders & Reinisch, 1999). For instance, male responders tended to regard more activities as sex compared to female counterparts (Barnett et al., 2017; Gute et al., 2008; Pitts & Rahman, 2001; Richters & Song, 1999), but this association was not necessarily observed by other studies (Randall & Byers, 2003; Sewell & Strassberg, 2015). Another observed tendency is that younger individuals usually claim less frequently that manual or oral stimulation is equivalent to sex (Pitts & Rahman, 2001; Richters & Song, 1999). It must be noted that Sanders et al. (2010) found a U-shaped relationship, meaning that both the younger and older individuals are less likely to believe that certain behaviors are equivalent to sex compared to age groups in between (Sanders et al., 2010). Previous sexual experience may also influence sexual definitions of some non-penetrative intimate behaviors (Bersamin et al., 2007; Sanders & Reinisch, 1999), but findings are inconclusive (Randall & Byers, 2003; Trotter & Alderson, 2007). Bersamin et al. (2007) found that the impact of previous sexual experience is superior to age and gender regarding loss of virginity.

A recent study underlined the possible role of social desirability—the tendency to give culturally appropriate responses in surveys. Among participants admitting to have taken part in oral sex without experience of vaginal or anal penetration, those who answered “No” to having had sex scored higher on the Marlowe-Crowne social desirability scale than those who answered “Yes” (Den Haese & King, 2022). The effect of other psychometric measures seems to be limited in the scientific literature regarding sexual definitions, although anxiety and body image have been widely investigated as predictors of sexual health behavior before (Akers et al., 2009; Coyle et al., 2019; Grogan, 2016). Anxiety is characterized by excessive worrying, hypervigilance, and physical symptoms originating from the heightened activation of the sympathetic nervous system (Zsido et al., 2020). Anxiety may influence sexual heath behavior, as those with higher anxiety level exhibit risky sexual behaviors more often (Coyle et al., 2019). Body image is a complex phenomenon referring to a person’s perceptions, thoughts, and feelings about his or her own body (Grogan, 2016). Negative body image may be related to many unfavorable outcomes, such as mental disorders and risky health behaviors. The role of these psychometric factors on diverse sexual definitions was not subject to previous investigations.

All types of intimate behaviors carry the risk of regret and emotional harm or the transmission of STDs. Schools have an important role in the prevention of these risks and improvement in attitudes towards sex by offering proper sexual education (Eitle & Thorsen, 2018; Kirby, 2002). These programs should also encompass the differences in sexual definitions. If a sexual health educator mentions only the term “sex” without elaborating its meaning, adolescents who do not consider unidirectional intimate behavior as “sex,” may not be aware of the risky nature of these kinds of interactions. To the best of our knowledge, no study has focused exclusively on adolescents except Bersamin et al. (2007) who investigated the terms “virginity” and “abstinence”. Furthermore, there is no evidence of non-English-speaking subjects in the literature yet and not much is known about how psychometric factors influence sexual definitions either. Thus, the present study aims to explore the following:

  1. 1.

    Assess sexual definitions among adolescents in a Northern-Hungarian settlement. Our hypothesis was that the previously described hierarchy of intimate behaviors would be reproduced in our sample. Furthermore, we hypothesized that our newly introduced items: “fondling each other’s body through clothes” and “stroking genitals through clothes” would be placed between kissing and manual stimulation of the genitals in the hierarchy.

  2. 2.

    Assess possible predictors of sexual definitions of the most ambiguous, unidirectional intimate behaviors and introduce the investigation of anxiety and body image in this field. Our hypothesis was that older participants, males, and sexually experienced adolescents will consider these types of behaviors less frequently as sex, while adolescents with higher anxiety and less favorable body image may regard these acts as sex more often.

Methods

Participants and Procedure

The present study used data from a survey-based longitudinal research project. The primary goal of the project was to measure the effectiveness of a health education intervention program (Balassagyarmat Health Education Program (BEP)) in secondary schools of Balassagyarmat (a settlement in Northern Hungary). The BEP aimed to improve health literacy in a highly deprived region of Hungary and focused on various topics, like sexual health, substance use, nutrition, physical activity, basic life support, and infection control. All five secondary schools of Balassagyarmat (detailed characteristics of all school are described in Appendix) were included in the program.

In this analysis, we used pooled cross-sectional pre-intervention data of all 9th-grade classes of two consecutive academic years of the project from 2019 and 2020. Those who did not take part in the intervention were considered eligible for inclusion in the current research. The head of each school gave permission for the study, and parental approval was sought with an opt-out procedure. Participation of students was voluntary. Pupils completed an online questionnaire individually in 20–45 min during school hours in a classroom setting. All questions were mandatory to answer. Data was collected under the supervision of research assistants without the presence of teachers. Ethical approval was obtained from the Institutional Review Board of Semmelweis University (SE TUKEB: 276/2017).

Measures

Sexual Definitions

A series of questions listing seven behaviors similar to the ones used in SDS (Sanders & Reinisch, 1999) were arranged to define what respondents consider as sex. We introduced two new items (Items 2 and 3) for a more detailed exploration of sexual definitions, as the first sexual experiences of adolescents may follow smaller incremental steps from kissing to sexual intercourse in our opinion. The English version of the item read: “Please indicate which of the following count as sex according to your opinion.” (1—“kissing”; 2, “fondling each other’s body through clothes”; 3, “stroking genitals through clothes”; 4, “stroking genitals directly/naked”; 5, “stimulation of genitals with mouth”; 6, “penetration as union of genitals”; 7, “anal penetration as intercourse in the rectum”), and pupils could mark “yes” or “no” for each behavior.

Predictor Variables

We examined the associations of five factors with sexual definitions of unidirectional intimate behaviors: age, gender, previous sexual experience, body image, and trait anxiety. To evaluate sexual experience of participants, two questions were developed based on the Health Behavior in School-aged Children study (HBSC, 2014) (Arnold et al., 2016): (1) “Have you ever had any kind of sexual encounter?” and (2) “Have you had penetrative sexual intercourse?” Based on the answer, a 3-level variable was created: (1, no/no) never had any kind of sexual experience, (2, yes/no) had only non-penetrative sex, and (3, yes/yes) had penetrative intercourse. Body image was measured by the body image subscale of the Body Investment Scale (BIS) (Orbach & Mikulincer, 1998) which is widely used for assessing body image feelings and attitudes in order to predict self-harm and other risky behavior especially among adolescents. It consists of 6 statements (e.g., “I am satisfied with my appearance.”) rated on a 5-point scale (ranging from 0 “ I do not agree at all” to 4 “Strongly agree”). A higher total score means more positive feelings about the body (Orbach & Mikulincer, 1998). The Hungarian adaptation of the subscale (Arnold et al., 2016) was used and it showed excellent internal consistency in our sample (Cronbach’s alpha = 0.91). Trait anxiety was assessed with one of the most frequently used tools, the State-Trait Anxiety Inventory for Children (STAI-C) Trait scale (Spielberger et al., 1973). This measure comprises 20 items describing behaviors (e.g., “I worry too much.”), and respondents are asked to indicate the frequency of occurrence of these behaviors on a 3-point scale (1, “hardly ever”; 2, “sometimes”; or 3, “often”). The ratings are summed up, and higher scores represent more proneness to anxiety. The Hungarian adaptation of the scale (Sipos & Sipos, 1979) was used and it showed good reliability in our sample (α = 0.91).

Statistical Analysis

We performed descriptive analysis on sexual definitions and background variables (frequency with percentage or mean with standard deviation). Multilevel logistic regression models (MLM) were used to assess the association of background variables with adolescents’ definitions on “stroking genitals directly/naked” (Model a) and “stimulation of genitals with mouth” (Model b) as outcome variables. MLM was chosen to deal with the hierarchical data structure of our study design resulting from the clustering of participants in classes within schools. The necessity of using the MLM was confirmed by calculating the design effect (DEa = 5.28 and DEb = 2.98). Any values above 2 suggest that clustering should not be ignored (Muthén & Satorra, 1995). Only school was used as a contextual variable because the group size of classes was too small to include. Variables of school characteristics (e.g., school type) were not included because all five secondary schools provide unique atmosphere (see Appendix) which cannot be compared to each other. The analysis was conducted in consecutive steps based on Sommet and Morselli (2017): variables were first centered and then, the proportion of the variability attributable to school-membership (intra-class correlation) was assessed with an intercept-only model. We then attested that estimating for random slopes would not result in a better fit to the data. Therefore, our final models employed random intercepts and fixed slopes by including all background variables, and odds ratios with 95% confidence intervals were calculated. To assess the total variance explained for both models, the \({R}_{\mathrm{MVP}}^{2}\) measure was calculated based on multilevel variance partitioning as suggested by LaHuis et al. (2014). Statistical analyses were conducted in R (R-3.6.3.). Significance was set at p < 0.05.

Results

Of the 854 9th-grade pupils, 817 were found to be eligible for the study as they did not participate in the intervention. The final sample consisted of 612 participants (27 did not agree to participate, 161 were not present on the day of data collection, six did not complete the majority of the questions, and 11 students were excluded due to age limit [mean age + 3SD]), resulting in a response rate of 74.9%. The mean age was 16.1 years (min 15; max 18) and the proportion of females was 46.2% (n = 283). No significant differences were found between the samples from the two different academic years in participation rate, gender, and age.

Significant Gender Difference

Kissing and fondling each other’s body through clothes were interpreted as “sex” by only one-tenth of the participants, while more than one-third of our sample (36.8%) thought of stroking genitals through clothes as sex. The majority of participants thought the same for stroking genitals directly/naked, stimulation of genitals with mouth, and anal penetration. Agreement increased from stroking genitals directly/naked through stimulation of genitals with mouth to anal penetration. Penetration as a union of genitals was considered sex by almost every pupil (98.5%). Male students were more likely to consider behaviors in the middle of the hierarchy as “sex” compared to females. Results are summarized in Table 1, while further descriptive data on background variables are shown in Table 2.

Table 1 Descriptives on sexual definitions: the number and proportion of adolescents considering the listed intimate behaviors as sex
Full size table
Table 2 Descriptives on background variables
Full size table

Predictors of Sexual Definitions

The intra-class correlations in the intercept-only models were found to be ρ = 0.12 in Model a and ρ = 0.05 in Model b, so the contextual variable (school) explained 12% of the total variance for stroking genitals directly/naked and 5% for stimulation of genitals with mouth.

Age and gender were found to be significant predictors in both multilevel models. Model a (shown in Table 3) showed an OR of 0.75 (95% CI 0.59–0.96) for age, meaning that the likelihood of older participants to consider stroking genitals directly/naked as “sex” is 0.75 times lower compared to their 1-year younger peers. The OR = 0.40 ( 95% CI 0.25–0.62) for female gender means that girls were 0.4 times less likely to interpret stroking genitals directly/naked as “sex” compared to boys. In Model b (shown in Table 4), the ORs were 0.66 (95% CI 0.49–0.87) and 0.40 (95% CI 0.23–0.66) for age and gender, respectively. Sexual experience, trait anxiety, and body image were not shown to be significant predicators in either of the models. The total variance explained was \({R}_{\mathrm{MVP}}^{2}\) = 0.21 in Model a and 0.23 in Model b.

Table 3 Model a: considering stroking genitals directly/naked as “sex”
Full size table
Table 4 Model b: considering stimulation of genitals with mouth as “sex”
Full size table

Discussion

Our results show that different forms of intimate behaviors are considered “sex” in similar rates to previous findings, even though these studies were conducted on different age groups (Barnett et al., 2017; Gute et al., 2008; Pitts & Rahman, 2001; Richters & Song, 1999; Sanders & Reinisch, 1999; Sanders et al., 2010). Bidirectional genital contact was defined as “sex” by the vast majority of our responders, whereas most participant do not consider kissing and fondling each other’s body through clothes as “sex.” On the other hand, agreement was lower for unidirectional genital contacts. The previously described hierarchy of sexual behaviors by Horowitz and Spicer (2013) appeared in our population as well. Our two new items, “fondling each other’s body through clothes” and “stroking genitals through clothes,” were positioned between “kissing” and “stroking genitals directly/naked” in the hierarchy. Interestingly, our responders considered “stroking genitals through clothes” as “sex” at a surprisingly higher rate than expected. This could be explained the fact that a large majority of our respondents did not have any sexual experience.

Schools play an important role in promoting a safe sexual life, and studies show that school-based sexual education programs are effective (Eitle & Thorsen, 2018; Garzón-Orjuela et al., 2021; Lameiras-Fernández et al., 2021). Their effectiveness, however, could be further improved by specifying the definition of sex, as this may show individual differences, resulting in the misinterpretation of sexual education messages by some. Our study investigated the association of age, gender, sexual experience, and two psychometric measures with unidirectional sexual definitions. As we hypothesized before, older participants were less likely to label oral and genital contact as sex. This is in line with the results of Bersamin et al. (2007) namely, that older adolescents are more likely to consider someone still a virgin even after engaging in genital touching. This was not corroborated by Pitts and Rahman and Richter and Song. They, however, examined young adults and not adolescents (Pitts & Rahman, 2001; Richters & Song, 1999).

Regarding gender, our results confirm the findings that males are more likely to consider unidirectional intimate behaviors as sex (Pitts & Rahman, 2001) even if this is not necessarily replicated in other studies (Randall & Byers, 2003; Sewell & Strassberg, 2015). In our opinion, this incongruency may be solved by the phenomenon of “sexual double standard” (Crawford & Popp, 2003), which is already present in the adolescent population (Kreager & Staff, 2009; Soller & Haynie, 2017). This concept derives from the different social norms expected from men and women and states that it is more acceptable for men to have sex with multiple partners or to engage in sex at an earlier age compared to women, who are often stigmatized if exhibiting the same behavior (Crawford & Popp, 2003). This can lead to different perceptions of the same behavior.

In our target group, sexual experience does not seem to play a role in sexual definitions, which is in concordance with previous studies (Hans & Kimberly, 2011; Randall & Byers, 2003; Trotter & Alderson, 2007). However, it is interesting that when adolescents are asked to define what remaining virgin means, subjects with prior sexual experience of a certain type of intimate behavior (except for vaginal intercourse) are more likely to think that someone is still a virgin after engaging in that specific behavior compared to those who do not have any experience of that behavior (Bersamin et al., 2007). The authors explained it as a result of adolescents rationalizing their behavior to maintain virginity status (Bersamin et al., 2007). However, the desirability of maintaining or losing virginity is dependent on social norms (Carpenter, 2001; Moussaoui et al., 2022), just like the definition of sex. Unfortunately, we do not have information on this in our target group.

As a novel approach in this field, we introduced two psychometric factors as possible predictors of sexual definition: trait anxiety and body image. According to our results, these factors do not influence sexual definitions, which is somewhat surprising, as higher anxiety has been linked to higher fear of intimacy and lower rate of sexual satisfaction (Carcedo et al., 2020; Montesi et al., 2013), and this may influence the perception of the different forms of intimate behaviors. As for body image, previous literature emphasizes that people with negative body image may avoid sexual activity (La Rocque & Cioe, 2011) and also tend to take more sexual risks than those with favorable body image (Littleton et al., 2005). We could not find a significant correlation between body image and sexual definitions, which can be attributed to the fact that the BIS score of our sample is relatively positive, and few people were affected by negative body image. Qualitative studies focusing on groups with negative body image may elaborate this possible correlation.

Strengths and Limitations

A major strength of our study is that it focused on adolescents, the primary target of sexual education programs. Moreover, this is one of the first studies to examine sexual definitions in a non-English-speaking community. A subsequent strength is that our study was carried out in a socially deprived population, who may be more predisposed to the unwanted consequences of sexual encounters, and thus, it is especially important to understand their views and understanding on sex. Another strength is that we examined the effect of psychometric factors like anxiety and body image on sexual definitions, a relatively understudied relationship in the field.

One of the limitations of our study is that our sample included secondary school students of a single settlement. Moreover, we were not able to exclude the confounding effect of social desirability (Den Haese & King, 2022), ethnicity, or sexual orientation.

Conclusions

The definition of sexuality has greatly expanded over the years; therefore, a more detailed sexual education encompassing a wider variety of behaviors is required. The main message of sexual health promotion is that participants should be mindful when engaging in any form of sexual conduct, as this may be associated with both emotionally joyful and harmful experiences (Wesche et al., 2017). To decrease the risk of misunderstanding, it is essential to know that simple keywords such as “sex” may have different meanings for different people.

In our study, we focused on scenarios where partners touch each other’s genitals orally or manually. Even though these situations do not increase the risk of unintended pregnancy, they may still be linked to STD transmission, emotional harm, and regret. We have found two factors that coincided with an increased chance for these types of intimate behaviors being considered “sex”: male gender and younger age. This reinforces that the diversity of sexual definitions should be considered when designing health promotion programs. Therefore, we advise professionals to consider the following:

  • Be mindful of semantic sensitivity and heterogeneity in sexual health education.

  • Be and make people aware of the risk of regretted sexual experience of any kind of intimate behavior besides unwanted pregnancy and STD transmission.

  • Further investigate the factors affecting the effectiveness of sexual health education.

In our opinion, future studies should focus more on adolescents, as most individuals become sexually active in this period of their life and sexual education often targets adolescents. Studies should also focus more on elaborating the effect of gender and age in this specific age group. Additionally, the role of anxiety and body image issues should also be further examined, as these factors affect a growing number of students and may influence the perception of their sexual behaviors.