LSA

During the last decade in Norway, the mean age at sexual initiation has dropped one year for females (to 16.7 years) and 6 months for males (to 18 years), and the number of early initiators has risen (Pedersen & Samuelsen, 2003). A wide array of studies describe the characteristics of adolescents with early sexual initiation. Early onset of sexual intercourse has been found to be associated with increased risk of sexually transmitted diseases, a high number of sexual patterns, and unwanted pregnancies and abortions (Hayes, 1987; O'Donnell, O'Donnell, & Stueve, 2001; Zabin, Kantner, & Zelnik, 1979), having experienced sexual abuse (Edgardh, 2000), having used drugs and marijuana (Fergusson, Horwood, & Lynskey, 1994; Lanctot & Smith, 2001; Orr, Beiter, & Ingersoll, 1971), early alcohol or smoking initiation (Andersson-Ellstrom, Forssman, & Milsom, 1996; Tyden, 1996), delinquency (Orr et al., 1991) and eating disorders (Kaltiala-Heino, Rimpela, Rissa nen, & Rantanen, 2001). An early onset of noncoital sexual interaction has also been shown to be related to antisocial behavior and substance abuse (Jakobsen, Rise, Aas, & Anderssen, 1997).

Research on early sexual initiation has also shown that having an older boyfriend or girlfriend and peer norms about sexual behavior are strongly associated with sexual experience in young adolescents (Kinsman, Romer, & Schwarz, 1998; Marin, Coyle, Gomez, Carvajal, & Kirby, 2000). In particular, females with an early sexual debut more often have partners considerably older than themselves (Edgardh, 2000; Leitenberg & Saltzman, 2000). Having much older male partners has been shown to be associated with a higher degree of problem behaviors in females who have had their first intercourse very early or early in adolescence (Leitenberg & Saltzman, 2000). The same trend is evidenced with respect to American teenage pregnancies: Men who have finished their schooling father two-thirds of the infants born to school-age mothers and are on the average 4.2 years older than the senior high mothers and 6.8 years older than the junior high mothers (Males & Chew, 1996). Thus, these studies converge in identifying late adolescent or young adult males as sexual partners for early adolescent females as being potentially detrimental to some females at this age.

Research on early sexual intercourse often focuses primarily on the characteristics of the early debut adolescents and on the negative consequences of their sexual debut. As a consequence, much prevention aimed at postponing sexual initiation targets young girls and tends to focus on individualized explanations of sexuality and engagement in risky sexual behavior, while the much older partners tend to be invisible or not perceived as relevant. However, the extensive involvement of young adult males in both school-age motherhood and its precursors represents a significant factor deserving greater attention and discussion (Males & Chew, 1996). Preventive efforts aimed at protecting pre-pubertal or pubertal adolescents from negative consequences of an early sexual debut should also address the sexual attitudes and willingness of older adolescents to have sex with their much younger peers.

Norms concerning sexual behavior among young people is institutionalized through the age of legal consent, but may vary in different segments of the youth population independently of this judicial demarcation. A hypothesis of a high prevalence of positive attitudes toward engaging in sexual relationships with very young adolescents follows the factual lowering of the age of sexual initiation. In general, only a small proportion of American adolescents regard first intercourse as being appropriate behavior for young adolescents between 12 and 14 years (Rosenthal & Smith, 1997). Nevertheless, the presence of erotic fantasies about younger adolescents, or the will to seize the opportunity to have sex with a younger adolescent may be widespread among older adolescents. A substantial quantity of research literature describes the effect of perceptions of peers' sexual behaviors and attitudes on young people's own sexual practices. Older adolescents with positive attitudes toward having sex with younger adolescents could be understood to establish norms and to be bearers of general social attitudes. This may in turn affect the sexual practice of others through peer influence.

To our knowledge there are two reported studies in the international research literature that have attempted to investigate the hypothetical likelihood of young adults having sex with underage youths based on nonclinical samples. In these studies, 7% of male undergraduate students (Briere & Runtz, 1989) and 4.4% of university students (Briere, Henschel, & Smiljanich, 1992) reported some hypothetical likelihood Of "having sex with a child" (LSC). Briere and Runtz (1989) concluded: "male sexual response to children is relatively common in our society" (Ibid p. 71). Endorsing some likelihood of having sex with a child was associated with lenient attitudes toward child sexual abuse (ATSA) (Briere et al., 1992), having had early negative sexual experiences themselves, large number of sexual partners, and masturbation to pornography (Briere & Runtz, 1989).

The term "child" was not defined according to age in Briere et al.'s study. How old is a "child" in the mind of the respondents when it comes to sexual activity? It is a weakness in the Briere surveys that the concept of "child" is not specified, leaving it to the respondents to define the age of the "child" sexual partner. Normally, a "child" is perceived as a pre-pubertal individual. In a sexual context, however, the individual could be perceived as a "child" until reaching the age of legal consent (16 years).

Thus, our first research question is whether the older adolescent's assessment of the hypothetical likelihood of having sex with a much younger sexual partner is affected by the age of the younger partner in question. The age span between 12 and 15 is related not only to development in physical and emotional maturity, but also to the nature of the social relations and situations encountered by a young adolescent. Dating and parties where large amounts of alcohol are consumed become more frequent as the younger adolescent grows older, and early onset of alcohol use has been shown to be a risk factor for early sexual debut among young people (Hingson, Heeren, Winter, & Wechsler, 2003). These factors may influence the way respondents reflect on the hypothetical likelihood that he/she is asked to assess.

We hypothesize that an older adolescent's assessment of the likelihood of having sex with a younger adolescent would vary both with the age of the sexual partner and the age difference between the respondent and the sexual partner. Reserving the term "children" for pre-adolescents, the conclusion that male sexual responses to children are relatively common on these grounds may be premature. By specifying different age groups, we want to assess the hypothetical normative sexual interests in preadolescents separately from a corresponding interest toward adolescents older than 12 years.

Three Approaches

In a context of adolescent sexuality, expressing a hypothetical likelihood of having sex with a much younger adolescent of a specific age should not be interpreted as having an urge to sexually molest adolescents or children. Rather it is an indication of positive attitudes toward younger adolescents as possible sexual partners. Such positive attitudes represent nonconforming sexual timetable norms in a relationship where the two partners differ in age and most likely in power, status, and maturity. Sexualized images of younger adolescents abound in advertising and pornography, and the Lolita figure is well known in literature and film. The interpretation of these positive attitudes toward younger adolescents as sexual partners could be approached through these three aspects: norm-breaking, social relations, and sexual nonconformity.

The first approach, related to norm. breaking, places nonconforming sexual attitudes within a domain of general antisocial behavior in adolescence. The syndrome of problem behavior described by Jessor and Jessor (1977) includes criminality and norm-breaking behavior as well as alcohol and drug abuse and early sexual initiation. Positive attitudes toward having sex with a 13- to 14-year-old may be associated with other forms of norm-breaking behavior, or even be norm-confirming behavior in anti-social subcultures among young people.

The second approach may be that of social inadequacy, exemplified by the socially inadequate male, perhaps more immature than his peers, wishing for a sexual relationship with "any" girl (boy) or a girl (boy) with lower self-esteem and self-assertiveness than himself. In relationships between a younger adolescent and an adult partner where there is a significant age difference, there is reason to assume a corresponding power differential. A sexual relationship between a low-status or socially inadequate late adolescent boy and a much younger girl may not imply the same disparity in power. A boy with emotional loneliness and low self-esteem may therefore seek, or be less likely to reject, sexual contact with a younger, more vulnerable partner (Barbaree, Marshall, & McCormick, 1998). In line with this approach, we would hypothesize that a high degree of loneliness and limited ability to form close relationships characterizes older adolescents with positive attitudes toward sex with younger adolescents.

A third way of interpreting the phenomenon of positive attitudes in older adolescents toward sex with younger adolescents is to see it as a reflection of other nonconforming sexual activity. A high number of sexual partners, a profound interest in pornography or child pornography or even involvement in social networks where violent/child pornography is used and perceived positively may be one aspect of this, while previous experiences of forcing someone to have sex or of prostitution may be another.

Thus, the second research question is whether older adolescents with positive attitudes toward sex with a younger adolescent are characterized by signs of social inadequacy, problem behavior, and high levels of sexual activity and interest. These approaches are not to be viewed as being mutually exclusive, but rather potentially complementary.

Aim of the Study

The aim of the present study was to describe the distribution of (1) self-reported hypothetical likelihood of having sex with a child (LSC), and (2) self-reported hypothetical likelihood of having sex with youths of different ages (LSA < 12, LSA 13-14, LSA 15) if they were sure that no one would know and that there would be no punishment, in a school-based cluster sample of 18- to 19-year-old Norwegian urban adolescents in Oslo. What is the prevalence of older adolescents of this age who indicate such a hypothetical likelihood? Are there differences in prevalence between males and females? Do the older adolescents reconsider their perception of the likelihood of having sex with a younger adolescent when the age of the sexual partner is changed? We identified a group of late adolescent males with a likelihood of having sex with a partner in the ambiguous age span of puberty (13-14 years). This group is described in more detail by exploring the relationship between LSA 13-14 and a number of possible relevant characteristics.

METHOD

Sample

All students in the third grade (final year) in a stratified sample of seven nonvocational senior high schools in Oslo were asked to complete a questionnaire during class. In order to represent differences in socioeconomic status among students in nonvocational senior high schools in Oslo, the seven schools were from different areas in Oslo. It was thus a school-based cluster sample, where each of the schools constituted one cluster. In general, over 88% of the cohort attends senior high school in Oslo, and 52% receive nonvocational training (Statistics Norway, 1999). Permission to include students attending vocational high schools was not granted from the central school authorities on the grounds that the students are scattered in various industries and enterprises all over Oslo during their final year. Thus, the respondents included in the sample are mainly students planning to continue their education. This may have introduced a bias into the study, since students planning a higher education are generally more privileged than those who take vocational training.

The total sample consisted of 909 students registered at the beginning of the school year (mid-August 1997). Of these, 713 completed the questionnaire, yielding a response rate of 78%. To ensure that all students had reached the age of majority (18 years), data were collected in the spring of 1998. Above this age students are free to take part in surveys without informed consent from their parents. All of the students present in the classroom completed the questionnaire and returned it. The missing students were mainly those who had dropped out during the school year, in addition to a few who were absent on the day of the survey. Only one respondent was underage and was withdrawn from the study.

Procedure

The questionnaire and a description of the procedures were submitted to and authorized by the Norwegian Data Inspectorate and the central schooling authorities in Oslo. The questionnaire was handed out during class, and the students were seated at individual desks. The seating arrangements allowed sufficient distance between students to provide them with the necessary privacy and anonymity. Each student completed the questionnaire by himself/herself and placed in an unmarked envelope and then sealed it. The teacher was responsible for handing out and collecting the envelopes. On no occasion were lists of students' names handled by any member of the project group. All students were informed that participation was voluntary. In an enclosed letter the students were informed that the research group included a trained clinical psychologist who could offer consultation to respondents if required.

Instruments

In addition to standard background measures, the survey included four questions about the self-reported hypothetical likelihood of having sex with a child or younger adolescent defined by different age values. Several measures covering problem behavior, sexual experiences, and interests and psychosocial adjustment were also included. In order to ensure correct translation, all international scales and measures were translated into Norwegian and then back into English by a third person.

Hypothetical likelihood of having sex with a child (LSC) or younger adolescent (LSA). Respondents were asked to rate on a 5-point scale (5 = not likely to 1 = very likely) how likely they would be to have "sex with a child" if they could be assured that no one would know and that they would in no way be punished (Briere & Runtz, 1989). The measure was repeated later in the questionnaire, this time including specific ages of the sexual partner. The ages of the adolescent sexual partner specified was "younger than 12 years" (LSA < 12), "between 13 and 14 years" (LSA 13-14) and "15 years" (LSA 15). "Sex" was not operationally defined to represent sexual intercourse only, but was left open for the respondent to interpret and define. The measure was modeled after Malamuth's (1981) "likelihood of raping" (LR) measure.

By defining the hypothetical sexual partner to be "younger than 12 years" we intended to construct a category that has an unequivocal meaning of "child" for most people, both in the cultural, psychological, and biological sense. Among Norwegian youths between 13 and 14 years, 55% of the females and 88% of the males are still pre-/beginning/mid-pubertal by Tanner's stages (Young in Norway Study, 1992). In defining the sexual partner as "between 13 and 14 years," it was our intention to create a category for an ambiguous and transitional stage of life, but still within the limits of what is commonly regarded as a child in a cultural, psychological, and legal sense and as such nonlegitimate as a sexual partner. By defining the sexual partner as "15 years," we constructed a category closer to the informants themselves, more relevant for sexual encounters, but still below the age of consent in Norway.

The mean scores on a scale of positive attitudes toward sexual abuse of children (ATSA), developed by Briere and Runtz (1989) (described below), were compared between the different scores on the LSA questions using ANOVA. The mean score on ATSA was significantly higher for respondents with values 4-1 than for respondents with the value 5 (no likelihood) for all three age-specific LSA questions. The answers were thus dichotomized as 0 (no likelihood) and 1 (some likelihood) as per Briere and Runtz (1989).

Attitudes toward sexual abuse (ATSA). ATSA (Briere & Runtz, 1989) is an additive index of the score on 15 items regarding positive attitudes toward sex between children and adults, and sexual abuse of children (scale 15-70, mean 29.6, SD = 8.0; Cronbach's [alpha] = 0.77). Sample items are "An adult and a child should be allowed to have sex together if they both want to," "Sex with children is relatively harmless," "Sex between a child and an adult is likely to hurt the child in major ways."

Problem behaviors. Frequency of drinking was measured using a self-reported assessment of average frequency of alcohol consumption per year from 0 (never) to 4 (twice a week or more). To highlight the two extremes of the scale, the mid-category was chosen as the reference group in the logistic regression. We also wanted to know if the respondent had an alcohol dependency according to DSM-IV, with negative social or psychological consequences (e.g., fighting, truancy, periods of "sadness." Alcohol-related problems were thus measured using 5 items from Rutgers Alcohol Problem Index (RAPI, Cronbach's Alpha = 0.69) (White & Labouvie, 1989). The scale was categorized into three groups labeled "no problems," "moderate problems," and "severe problems" (0-2).

The respondents were also asked to report their frequency of cannabis use during the previous 12 months. The question was open-ended, allowing categorization of respondents as nonusers, experimental users (1-4 times during the previous 12 months) and regular users (5 + times during the previous 12 months). Report of use of other illegal drugs (e.g., amphetamines, ecstasy, cocaine, heroin during the previous 12 months were also included in the analysis.

Conduct problems were measured using 6 items selected from Olweus's scale of antisocial behavior (Olweus, 1989) and the National Youth Longitudinal study (Windle, 1990). Respondents reported the frequency of their involvement in these behaviors during the previous 12 months, as well as their age at the time of the first incident. The items cover types of behavior that most likely will be nonproblematic when low in frequency, such as playing truant, and more serious behaviors, such as burglary or car theft. Consequently, the cut-off point representing an indicator of a conduct problem was set differently for different items (Wichstrom, Skogen, & Oia, 1996) according to the distribution of respondents on each variable, in order to isolate groups of reasonable problem severity for each item. The respondents were assigned one point for each of the following: sneaking onto a bus or into a cinema more than 30 times/first incident <13 years (10.7%), playing truant more than 30 times/first incident <13 years (2.6%), staying out all night without parental permission more than 10 times/ first incident <13 years (0.7%), theft of more than 100 USD more than once (5.9%), stealing a car/motorcycle (2.7%), and breaking in to steal something (2.3%). Subjects meeting one or more of these criteria were classified as having conduct problems, and given the value 1 on a dichotomized variable.

Early sexual initiation is included in the problem behavior syndrome described by Jessor and Jessor (1977). The respondents were asked to recall the age of their first sexual intercourse. Sexual intercourse before turning 15 years was given the value 1 (6.7).

Nonconforming Sexual Activity

To assess the level of sexual experience we asked for both age of sexual initiation and number of partners. Given the respondents young age (18-20 years), the respondents were asked to report lifetime prevalence of coital sexual partners (0 = none to 4 = more than five). The scale was dichotomized into a measure where more than 5-sexual partners during lifetime was given the value 1. Furthermore, we asked if the respondents had ever used pressure or coercion to obtain sexual favors from someone. A record of prostitution experiences was measured using lifetime prevalence of buying or selling sexual favors.

Use of pornographic/sexually explicit material was measured using a question on frequency of use of pornographic films/books and magazines (Briere & Runtz, 1989), and a question on frequency of looking at pornography on the Internet, ranging from weekly to never on a six-point scale. Respondents were also asked to rate the extent of their viewing of pornography in magazines and/or movies over the last year, on a six-point scale. Due to almost universal access to the Internet, a sum score for frequency of pornography viewing in both written, visual, and "cyber" media was calculated. The scale was given 3 categories ranging from 0 (never or seldom) to 2 (weekly), in order to identify the small group of frequent viewers of pornography in all three media. We also wanted to know if the respondents had ever seen pornographic material depicting children (film/magazines/Internet). Friends viewing of and interest in pornograpic material was measured using four statements describing what "some of my friends" like to do--watching pornographic videos often, watching violent pornographic videos, watching child pornography on the Internet or thinking that hard-core pornography is objectionable (reverse)--rating the correctness of the statements from 1 (totally wrong) to 6 (totally correct) (Cronbach's alpha = 0.69).

The respondents' own experiences of unwanted sex were covered fairly extensively in the questionnaire. The respondents were asked if they had ever had sexual contact which they themselves had experienced as "unwanted and involuntary," followed by questions about their own age and the perpetrator's age at the time of the first incident. The age of consent in Norway is 16 years. Therefore, for the purpose of this study, child sexual abuse victimization was operationally defined as having had unwanted involuntary sexual contact at the age of 15 or younger.

Psychosocial Adjustment.

We included several measures of psychosocial adjustment: a measure of depression (Cronbach's Alpha = 0.81) based on 6 items taken from Hopkins Symptom Check List (HSCL) (Derogatis, 1977), a measure of relational skills satisfaction (Cronbach's Alpha = 0.60) measured using three items from the Sex Role Satisfaction Scale (Burt, 1980), and a measure of perceived loneliness (Cronbach's Alpha = 0.73) measured using four items from the UCLA Loneliness Scale (Russel, Peplau, & Cutrona, 1980).

Analysis

Prevalence of LSC and LSA are presented as cross tabulations by gender. Mean value of the scale ATSA is presented by gender for those with no and some likelihood of having sex with a child or younger adolescent of different ages. To compare characteristics of the older adolescent males with some likelihood for having sex with a younger adolescent partner (13-14 years) to other males of the same age group, distributions on the independent variables are shown for each group of respondents. The domain-specific variables were explored in three separate models where variables were excluded using the backwards step-wise procedure, aimed at assessing an efficient model in terms of explained variance.

RESULTS

The sample consisted of 307 males (43%) and 403 females (57%), of whom 71% were 18 years of age (M = 18.4 years, SD = 0.54). Two-thirds lived with both of their parents. Most Oslo youths in this age group drink alcohol regularly, and are allowed to buy alcohol from age 18. The average frequency of drinking in this sample was 2-4 times per month. A minority of the males (15%) and females (9%) experienced severe problems related to their drinking habits; the gender difference was not statistically significant. Thirty-six percent of the males and 33% of the females reported having used cannabis during the previous 12 months. This is in accordance with findings from other studies of adolescents in Oslo (Bakken, 1998). A small, nonsignificant gender difference was also noted in the incidence of use of other illegal drugs such as ecstasy (3-4 methylenedioxymethamphetamine, MDMA), amphetamines, cocaine, and heroin, where 12% of the males and 10% of the females reported illegal drug use.

A slightly larger proportion of males (38%) compared to females (31%) in this age group were still virgins ([chi square] = 3.29, p = .07). However, one in ten males and females alike reported a life prevalence of more than 5 sexual partners. Only 5.4% of the males reported never having seen pornographic material in magazines, in films, or on the Internet, while 40.7% of the females reported that they had never seen such material ([chi square] = 365.2, p < .001). One in four had viewed pornography in magazines or films at least once a week, while 13% had the same frequency of pornography use via the Internet. Child pornography was much less common, but it was nevertheless reported by a considerable minority among males (males = 17.3%, females = 2.5%, X2 = 47.09, p < .001), perhaps reflecting easier access to the Internet over the last couple of years. At some point, 7% of the males had used coercion to get sexual favors, compared to 1 percent of the females ([chi square] = 17.77, p < .001). Conversely, 14% of the females had been the victim of unwanted, involuntary sexual contact before the age of 16. The corresponding figure for males was 1.7%.

Table 1 shows the prevalence of males and females who reported a hypothetical likelihood of having sex with a child, as well as the distribution on the LSA measure where the potential sexual partner was defined as less than 12 years old, 13-14 years old, and 15 years old. It also includes a comparison of mean scores of ATSA in groups with different attitudes toward younger adolescents and children as sexual partners.

We found that the proportion of respondents indicating some likelihood of having sex with a younger adolescent differed greatly according to the age stated in the questionnaire (Table 1). While two-thirds of the males did not exclude at least some likelihood of having sex with a 15-year-old, some likelihood of having sex with a child less than 12 years was indicated by only 6%. Among females, the corresponding figures were 16.8% and 0.8%, respectively. Thus, we found a marked difference according to gender for all the age-specific questions. The percentage of some likelihood for males was more than three times greater than for females. In comparing the mean score of the ATSA scale (Briere et al., 1992), the notion that some likelihood of having sex with partners of different ages reflected a positive attitude toward child sexual abuse was confirmed for males but not for females (Table 1). The difference between ATSA mean scores for males with some and no LSA was more distinct as the age of the young adolescent in question decreased, as can be seen from the increase in F-value.

For comparison, we also included the original measure where the word child was used with no further definition of age, adopted from Briere & Runtz (1989) (Table 1). It is worth noting that the percentage indicating a hypothetical likelihood of having sex with a child was greater than the percentage with the corresponding likelihood when the age was defined as "under 12" or "between 13 and 14."

For further analysis, we identified a subgroup of 58 males who reported some likelihood of having sex with a younger adolescent 13-14 years (LSA 13-14) by indicating 1-4 on the LSA-question where the child's age was defined as age 13-14 in Table 1. In this table this subgroup included some males who also reported a likelihood of having sex with a child less than 12 years of age. In the comparison group, only males with no such likelihood or some likelihood of having sex with a 15-year-old was included (n = 57).

Only 12 females reported some likelihood of having sex with a younger adolescent of 13-14 years. A bigger sample is needed for a more thorough analysis of this group, Thus, only male respondents were included in the further analysis. Table 2 shows cross tabulation analyses of independent variables representing the three approaches, comparing the 58 males with some LSA 13-14 to the rest of the males. Interestingly, even if the age difference between the respondents was minimal (18-20 years), age was negatively related to LSA 13-14. Males with no likelihood of having sex with a child age 13-14 was older (M = 18.37) compared to the comparison group (M = 18.16, F = 7.52, p < .01).

Problem Behavior

While the relationship between experiences of alcohol-related problems (RAPI) and LSA 13-14 was linear, the relationship to frequency of alcohol consumption is more complex. Both male abstainers and respondents with a pattern of weekly drinking were overrepresented among males with some LSA. This curvilinearity was tested by including drinking frequency squared to a bivariate logistic regression of the LSA 13-14-drinking frequency relationship. The squared component gave a significant improvement to the model (L[R.sub.[chi square] of drinking frequency = 3.02, df = 1, ns; L[R.sub.[chi square]] improvement when drinking frequency squared was entered = 3.88, df = 1, p < .05). Descriptive analysis of the relationship between LSA 13-14, alcohol consumption and RAPI revealed a tendency toward an interaction effect, resulting in an even higher probability of LSA 13-14 in males who drank frequently and had severe alcohol-related problems. Nineteen percent of heavy drinkers with no alcohol-related problems scored on the LSA 13-14. The corresponding figure for heavy drinkers with severe RAPI problems was 36%, but the interaction effect was not significant.

The association with use of cannabis and other illegal drugs was not significant. Almost one-third of the respondents with some LSA 13-14 scored more than 0 on the scale of conduct problems, compared to 14% in the comparison group. Likewise, early sexual initiation was more common among the LSA 13-14 males.

Based on 5 indicators from this model (drinking twice a week or more, severe alcohol-related problems, illegal drug use, conduct problems, and early sexual initiation), we constructed a composite measure of problem behaviors. Among the LSA 13-14 males, 45% reported at least one of these problem behavior indicators, compared to 29% of the males in the comparison group ([chi square] (df = 4) = 15.39, p < .01).

Multivariate analysis of this block of independent variables following a backward stepwise procedure showed that the best fit of the model included drinking frequency, conduct problems, and early sexual debut (Table 3). This model's [chi square] (df = 6) = 18.90 (p < .01). There were distinct positive effects of alcohol abstinence, drinking once a week, and conduct problems, with alcohol drinking frequency showing the strongest effect (Alcohol: Wald (df = 4) = 11.47, p < .05. Conduct problems: Wald (df = 1) = 5.04, p < .05).

Nonconforming Sexual Experiences

A high number of sexual partners did not seem to be associated with the likelihood of a sexual interest in younger adolescents. However, experience of using coercion to get sexual favors or prostitution experience (selling/buying) was more common in the some likelihood-group. Correspondingly, there was a relationship to frequent use of pornography, and a strong relationship to friends' use of and interest in pornography. Experience of child pornography seemed to be more frequent among males with a sexual interest in 13-14 year olds, but the difference between the two groups was not significant.

Only 5 males in the total sample reported experiences of sexual abuse before the age of 16. Three of these five males reported some likelihood of having sex with a younger adolescent 13-14 years old. Thus, we found a clear relationship between experiences of child sexual abuse and LSA 13-14, but this result must be interpreted with caution due to the small number of respondents.

Multivariate analysis of this block of independent variables following a backward stepwise procedure showed that the best fit of the model included sexual favors by coercion, respondents' own experiences of unwanted, involuntary sexual contact before the age of 16, use of pornography and friends' use of and interest in pornography (Table 3). (This model was [chi square] (df = 5) = 19.31, p < 0.01)). None of the model variables had significant effects in the multivariate analysis, but frequency of pornography use showed the strongest effect of the four (Wald = 4.12 (df = 2)).

Psychosocial Adjustment

Males in the LSA 13-14 subgroup showed signs of poor psychosocial adjustment, given the higher mean score for depression, lower relational skills satisfaction, and a higher frequency of emotional loneliness in this group. Twice as many scored on LSA 13-14 in the lonely group (34.4%) than in the comparison group (17.7%); ([chi square] = 5.07, p < .05). The correlation between loneliness and depression was high (r = 0.41, p < .001). Thus, the multivariate analysis of this block of independent variables following a backward stepwise procedure resulted in the best fit of the model including relational skills satisfaction and depression, yielding a fit for this model: [chi square] (df = 2) = 10.43 (p < .01) (Table 3). This block showed the lowest explained variance, due to a low number of included variables. Both of the variables were significant, with relational skills satisfaction showing the strongest effect (RelSkills: Wald = 5.65 (df = 1), p < .05). Depression: Wald = 3.94 (df = 1), p < .05).

DISCUSSION

We found that specifying the age of the younger adolescent partner had an important influence on the attitudes toward having sex with younger adolescents. Attitudes toward child sexual abuse were most positive among those indicating a hypothetical likelihood of having sex with a partner younger than 12 years of age. Compared to other male respondents, the older adolescent males with some likelihood of having sex with a partner 13-14 years of age, were more often alcohol abstainers or frequent drinkers, and had more alcohol- and conduct-related problems. They had also engaged in coercive sex as doers, buying and selling sexual favors, as well as having been subjected to sexual abuse. They were more often frequent pornography users as well as having friends with positive attitudes toward pornography. Finally, these males were also more often very lonely, showing signs of depressive mood and low relational skills satisfaction. Multivariately, all three approaches seem to be important to varying degrees, with psychosocial adjustment the least influential.

The sample included students in nonvocational high schools only. Students in vocational training experience their first sexual intercourse at an earlier age than students in theoretical training (Edgardh, Lewin, & Nilsson, 1999), which may reflect different attitudes regarding sexual age norms. Thus, we have reason to believe that the prevalence is underestimated compared to the prevalence in the general population of adolescents in this age group. However, the lower probability for including adolescents outside the school system, for instance with conduct problems/substance problems, is not likely to have affected the associations between the dependent and independent variables in major ways.

The relatively small age difference between the respondent and the sexual partner in the follow-up questions could be raised as an objection to this study. With an age difference of a mere 4 years, an 18-year-old and a 14-year-old could be perceived as peers. However, junior and senior high schools in Norway do not share the same buildings or the same schoolyards. Most students choose a senior high school outside their neighborhood, while the younger students still attend a local junior high school. The respondents included in this study would therefore not normally perceive themselves as a "peer group" in relation to younger adolescents in the neighboring junior high schools. It is also possible, since the questionnaire did not unambiguously state that any sexual contact with a person 13-14 years of age is illegal by Norwegian law, that not all respondents knew that they could be breaking the law if they behaved as the question hypothesized. This could result in an overestimate of youths with a likelihood of having sex with a 13-to 14-year-old.

The fact that a majority of males in the age group 18-20 years regarded young people of 15 years as probable sexual partners implies that the age of legal consent is not perceived as an absolute limit for consensual sex, but rather a norm that can be transgressed and negotiated in individual cases. The results also indicate that the age of the sexual partner is important, and that the males do not view pubescent (13-14 years) or pre-pubescent sexual partners (12 years or younger) as "sexual" to the same degree as they do 15-year-olds. In comparison, only a fraction of the females of 18-19 years regarded younger adolescents of 13-14 years as possible sexual partners.

A large minority of the respondents, males in particular, reported a hypothetical likelihood of having sex with a younger adolescent of 13-14 years. In this age group, sexuality and dating is becoming more relevant and the boundaries between consensual sex and abuse may be more blurred than is the case where the partner is distinctly preadolescent or of the same age. The questionnaire did not indicate the level of physical or mental maturity of the younger partner, nor were any conditions included regarding use of force or the nature of the "sex." However, the power differential related to age differences and the lesser ability of the younger partner to "say no" is an issue of concern in its own right. Given the undesirability of an early sexual debut, this group of males also represents possible sexual partners for much younger adolescents, thus being of interest as such. These positive attitudes toward sex with younger adolescent partners could also be perceived as a social force in the legitimization of sexual contacts with underage adolescents in general among older adolescent males. The considerable difference found between males and females is worth noting. The difference in pubertal timing between males and females could be one underlying factor in this gender difference, if we presuppose a heterosexual sexual relationship. While pubescent males could find younger females to be on nearly the same developmental level as themselves, teenage females are not as inclined to regard younger, less developed males as sexual partners. In addition to this, there is a clear tendency for females of this age to date older rather than younger males (Edgardh et al., 1999). Across many different studies and measures, men have been shown to have more frequent and more intense sexual desires than do women, as reflected in spontaneous thoughts about sex, frequency and variety of sexual fantasies, desired frequency of intercourse, desired number of partners, and other measures (Baumeister, Catanese, & Vohs, 2001). This gender difference in sex drive and sexual motivation could also be related to the gender difference in the present research.

We found that several of the variables included in the problem behavior approach were related to a positive attitude toward younger adolescents as sexual partners. For some of these males, positive attitudes towards younger adolescents as sexual partners seem to fall into a general domain of problem behavior. The trimmed model revealed that high levels of alcohol use and conduct problems in particular increased the odds of inclusion in the LSA 13-14 group.

However, the present study shows a curvilinear (U-shaped) relationship between drinking and LSA 13-14. As expected, males with a high level of alcohol consumption or experiences of negative consequences from their heavy drinking, showed a higher probability of scoring on LSA 13-14. More surprising is the fact that males with no or only limited experience of drinking alcohol scored almost the same as the heavy drinkers on LSA 13-14.

The general use of alcohol in the adolescent population indicates that drinking may serve an important role in the normal process of growing up. Norwegian male abstainers and males with a late onset of drinking have been found to have more mental health problems and indications of delayed adjustment to the adult role compared to those who follow the mainstream (Pape & Hammer, 1996). Late starters have also been found to have fewer close friends and a lower frequency of contact with friends (Leifman, Kuhlhorn, Andreasson, & Romelsjo, 1995; Pedersen, 1993a, 1993b). In this context, including the small but significant effect of loneliness, low relational skills satisfaction, and depression in the present study, the U-shaped relationship between frequency of drinking and LSA 13-14 indicates that the males of the LSA 13-14 group could perhaps also be described by a mechanism related to poor social and psychosocial adjustment. The respondents in the LSA 13-14 group are less satisfied with their relational skills and report a higher score on depression symptoms.

Relationships between the self-reported hypothetical likelihood of having sex with a younger adolescent of 13-14 years and the various measures related to other nonconforming sexual experiences, e.g., to coercive sex, experiences of prostitution, use of pornography and to friends' use of and interest in pornographic material are also noteworthy. The absence of a relationship to what we could term "experimental" use of pornography suggests that the use of pornography in this sense is not to be viewed as nonconformist sexual behavior in this group of young men. A frequency of weekly use of pornography may indicate that pornography is a component of habit or lifestyle and that as such it corresponds with positive attitudes toward having sex with an underage adolescent. Friends' interest in child pornography and violent pornography was also related to the LSA 13-14 measure. Our findings therefore also indicate that we may need to view these experiences in a peer group context, where a positive attitude toward younger adolescents as sexual partners, could be learned, sustained or reinforced.

We also found a strong relationship to a history of having been sexually abused as a child, which increased the probability of LSA 13-14 by 6.5. Because of the small number of male child abuse victims (n = 5), this finding should be interpreted with caution.

There has traditionally been a strong focus on offenders' prior sexual histories, both in explaining sexual offending and in the treatment of sexual offenders (Daleiden, Kaufman, Hilliker, & O'Neil, 1998). Non-conformist sexual behavior could be the result of learning and subsequent deviant conditioning and reinforcement of sexually nonappropriate behavior (Marshall & Eccles, 1993; McGuire, Carlisle, & Young, 1965). Studies of adolescent sexual offenders report both more frequent use of pornographic magazines (Zgourides, Monto, & Harris, 1997), and indications of poor psychosocial adjustment such as loneliness and feelings of social isolation (Awad & Saunders, 1989; Carpenter, Peed, & Eastman, 1995; Marshall, 1989) and depression (Becker, Kaplan, Tenke, & Tartaglini, 1991) have also been found. Intoxication during or before a sexual offense is believed to contribute to lowering internal inhibitors against sexual abuse (Ageton, 1983; Finkelhor, 1984), and offenders reporting a high level of alcohol consumption or reporting that alcohol increased their arousal have been found to abuse a higher number of victims (Becker & Stein, 1991). As Jessor and Jessor (1977) argue, sexually abusive behavior can be seen as one of several elements in a syndrome of conduct disorders. Young males with a history of norm-violating or antisocial behavior could be more prone to crossing the line in the sexual domain as well. Do the findings in this study imply that the group we have identified through a question of a hypothetical likelihood for having sex with a person 13-14 years of age may also be at risk to act out this likelihood?

We have no means of predicting future action on the grounds of positive attitudes toward having sex with younger adolescents. In this case, the subgroup of males indicating positive attitudes toward younger adolescents as sexual partners is in all likelihood comprised of males with a differentiated risk for sexually abusive behavior. Some of these young males may have checked "some likelihood" on the grounds that they cannot rule out even the "impossible" when it comes to their future actions, depending on the situation. Others may have sexual fantasies about young girls or boys, but will never actually act on these fantasies. These and other kinds of scenarios will be true of the majority of the males in this subgroup. However, our findings indicate that the subgroup of males in this study do share several characteristics with groups of adolescent offenders. This could imply an increased likelihood to actually have sex with an underage adolescent among a minority of these males.

As a phenomenon found among normal population males, positive attitudes toward younger adolescents as sexual partners should primarily be understood as attitudes reflecting acceptance of early sexual involvement, rather than sexual abuse. Based on the scores on the ATSA scale, we know that these males believe that it is not harmful for children to have sex with adults, and that the child should decide for herself/himself whether she/he is ready for sex. In the case of a 13-14 year old sexual partner rather than a child, these beliefs may seem to be even more true to the older adolescent male. What these older adolescent males do not seem to consider, is that it might not be the responsibility of the 13-14 year old partner to "say no," but rather the older partner's responsibility not to ask.

CONCLUSION

This study revealed that about one in five of the older adolescent males reported positive attitudes toward having sex with a younger adolescent partner of 13-14 years. Thus, these positive attitudes seem to be rather widespread among Norwegian urban high school males. Generally, there is no controversy over sexual education in Norwegian schools today, and education on reproductive health and contraceptives is provided from the 9th grade. However, the results of this study provide a strong indication that sexual education should pass on not only technical information but also relational issues in order to heighten boys' consciousness about peer pressure, the power differential in relationships with younger partners, and their responsibility in a sexual relationship. This is especially important since boys generally seek sexual and contraceptive counseling from health personnel to a lesser degree than do girls and could therefore be more effectively reached through general programs.

Positive attitudes toward having sex with younger adolescent partners were associated with other problem behaviors, nonconforming sexual behavior, and psychosocial maladjustment, implying that preventive efforts should also be targeted at particular adolescent male populations. Professionals working with youth at risk must include sexual and relational issues in their preventive intervention and communication with the older boys that typically represent the romantic or sexual partners for young girls most at risk for unwanted pregnancies.

REFERENCES

Ageton, S. (1983). Sexual assault among adolescents. Lexington, MA: Lexington Books.

Andersson-Ellstrom, A., Forssman, L., & Milsom, I. (1996). The relationship between knowledge about sexually transmitted diseases and actual sexual behavior in a group of teenage girls. Genitourinary Medicine, 72, 32-36.

Awad, G. A., & Saunders, E. B. (1989). Adolescent child molesters: Clinical observations. Child Psychiatry and Human Development, 19, 195-206.

Bakken, A. (1998): Ungdomstid i storbyen [Urban adolescence]. NOVA-Rapport 7:98. Oslo: Norwegian Social Research, NOVA.

Barbaree, H. E., Marshall, W. L., & McCormick, J. (1998). The development of deviant sexual behavior among adolescents and its implications for prevention and treatment. The Irish Journal of Psychology, 19, 1-31.

Baumeister, R. F., Catanese, K. R., & Vohs, K. D. (2001). Is there a gender difference in strength of sex drive? Theoretical views, conceptual distinctions, and a review of relevant evidence. Personality and Social Psychology Review, 5, 242-273.

Becker, J. V., Kaplan, M. S., Tenke, C. E., & Tartaglini, A. (1991). The incidence of depressive symptomatology in juvenile sex offenders with a history of abuse. Child Abuse and Neglect, 15, 531-536.

Becker, J. V., & Stein, R. M. (1991). Is sexual erotica associated with sexual deviance in adolescent males? International Journal of Law and Psychiatry, 14, 85-95.

Briere, J., Henschel, D., & Smiljanich, K. (1992). Attitudes toward sexual abuse: Sex differences and construct validity. Journal of Research in Personality, 26, 398-406.

Briere, J., & Runtz, M. (1989). University males' sexual interest in children: Predicting potential indices of "pedophilia" in a nonforensic sample. Child Abuse and Neglect, 13, 65-75.

Burt, M. R. (1980). Cultural myths and support for rape. Journal of Personality and Social Psychology, 38, 217-230.

Carpenter, D. R., Peed, S. F., & Eastman, B. (1995). Personality characteristics of adolescent sexual offenders: A pilot study. Sexual Abuse: A Journal of Research and Treatment, 7, 195-203.

Daleiden, E. L., Kaufman, K. L., Hilliker, D. R., & O'Neill, J. O. (1998). The sexual histories and fantasies of youthful males: A comparison of sexual offending, nonsexual offending and nonoffending groups. Sexual Abuse: A Journal of Research and Treatment, 10, 195-209.

Derogatis, L. R. (1977). SCL-90R: Administrative, scoring and procedure manual I for the R version. Baltimore: John Hopkins University School of Medicine, Psychometrics Research Unit.

Edgardh, K. (2000). Sexual behavior and early coitarche among 17-year-old Swedish girls. Sexually Transmitted Infections, 76, 98-102.

Edgardh, K., Lewin, B., & Nilsson, B. R. (1999). Sexual experience and behavior as reported by 17-year-old girls and boys in Sweden. Scandinavian Journal of Sexology, 2, 41-60.

Fergusson, D. M., Horwood, L. J., & Lynskey, M. T. (1994). The comorbidities of adolescent problem behaviors: A latent class model. Journal of Abnormal Child Psychology, 22, 339-354.

Finkelhor, D. (1984). Child sexual abuse: New theory and research. New York: Free Press.

Hayes, C. D. (1987). Risking the future: Adolescent sexuality, pregnancy and childbearing. Vol. I. Washington, DC: National Academy Press.

Hingson, R., Heeren, T., Winter, M. R., & Wechsler, H. (2003). Early age at first drunkenness as a factor in college students' unplanned and unprotected sex attributable to drinking. Pediatrics, 111, 34-41.

Jakobsen, R., Rise, J., Aas, H., & Anderssen, N. (1997). Noncoital sexual interactions and problem behavior among young adolescents: The Norwegian longitudinal health behavior study. Journal of Adolescence, 20, 71-83.

Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. New York: Academic Press.

Kaltiala-Heino, R., Rimpela, M., Rissanen, A., & Rantanen, P. (2001). Early puberty and early sexual activity are associated with bulimic-type eating pathology in middle adolescence. Journal of Adolescent Health, 28, 346-352.

Kinsman, S. B., Romer, D., & Schwarz, D. F. (1998). Early sexual initiation: The role of peer norms. Pediatrics, 102, 1185-1192.

Lanctot, N., & Smith, C. A. (2001). Sexual activity, pregnancy, and deviance in a representative urban sample of African American girls. Journal of Youth and Adolescence, 30, 349-372.

Leifman, H., Kuhlhorn, E., Andreasson, P., & Romelsjo, A. (1995). Abstinence in late adolescence: Antecedents and covariates to a sober lifestyle and its consequences. Social Science and Medicine, 41, 113-121.

Leitenberg, H., & Saltzman, H. (2000). A statewide survey of age at first intercourse for adolescent females and age of their male partners: Relation to other risk behaviors and statutory rape implications. Archives of Sexual Behavior, 29, 203-215.

Malamuth, N. (1981). Rape proclivity among males. Journal of Social Issues, 37, 138-157.

Males, M., & Chew, K. S. Y. (1996). The ages of fathers in California adolescent births, 1993. American Journal of Public Health, 86, 565-568.

Marin, B. V., Coyle, K. K., Gomez, C. A., Carvajal, S. C., & Kirby, D. B. (2000). Older boyfriends and girlfriends increased risk of sexual initiation in young adolescents. Journal of Adolescent Health, 27, 409-418.

Marshall, W. L. (1989). Intimacy, loneliness and sexual offenders. Behavioral Research and Therapy, 27, 491-503.

Marshall, W. L., & Eccles, A. (1993). Pavlovian conditioning processes in adolescent sex offenders. In H. E. Barbaree, W. L. Marshall, & S. M. Hudson (Eds.), The juvenile sex offender (pp. 118-142). New York: Guilford Press.

McGuire, R. J., Carlisle, J. M., & Young, B. G. (1965). Sexual deviations as conditioned behavior: A hypothesis. Behavior Research and Therapy, 2, 185-190.

O'Donnell, L., O'Donnell, C. R., & Stueve, A. (2001). Early sexual initiation and subsequent sex-related risks among urban minority youth: The reach for health study. Family Planning Perspectives, 33, 268-275.

Olweus, D. (1989). Prevalence and incidence in the study of antisocial behavior: Definitions and measurement. In M. W. Klein (Ed.), Cross-national research in self-reported crime and delinquency (pp. 187-201). Dordrecht, NL: Kluwer Academic Publishers.

Orr, D. P., Beiter, M., & Ingersoll, G. (1991). Premature sexual activity as an indicator of psychosocial risk. Pediatrics, 87, 141-147.

Pape, H., & Hammer, T. (1996). Sober adolescence: Predictor of psychosocial maladjustment in young adulthood? Scandinavian Journal of Psychology, 37, 362-377.

Pedersen, W. (1993a). The majority fallacy reconsidered. Acta Sociologica, 36, 343-355.

Pedersen, W. (1993b). Adolescents, friendship and use of drugs. Tidsskrift for Den Norske Laegeforening, 113, 1747-1749.

Pedersen, W., & Samuelsen, S. O. (2003). Nye monstre av seksualatferd blant ungdom [New patterns of sexual behavior among young people]. Tidsskrift for Den Norske Laegeforening, 123, 3006-3009.

Rosenthal, D. A., & Smith, A. M. A. (1997). Adolescent sexual timetables. Journal of Youth and Adolescence, 26, 619-636.

Russell, D., Peplau, L. A., & Cutrona, C. E. (1980). The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 472-480.

Statistics Norway (1999). Statistical Yearbook 1999. Oslo: Statistics Norway.

Tyden, T. (1996). It will not happen to me: Sexual behavior among high school and university students and evaluation of STD-prevention programs. Doctorial thesis. Uppsala, Sweden: Uppsala University.

White, H. R., & Labouvie, E. W. (1989). Toward the assessment of adolescent problem drinking. Journal of Studies of Alcohol, 50, 198-207.

Wichstrom, L., Skogen, K., & Oia, T. (1996). Increased rate of conduct problems in urban areas: What is the mechanism? Journal of the American Academy for Child and Adolescent Psychiatry, 35, 471-479.

Windle, M. (1990). A longitudinal study of antisocial behavior in early adolescence as predictors of late adolescence substance abuse: Gender and ethnic group differences. Journal of Abnormal Psychology, 99, 86-91.

Zabin, L. S., Kantner, J. F., & Zelnik, M. (1979). The risk of adolescent pregnancy in the first months of intercourse. Family Planning Perspectives, 11, 215-222.

Zgourides, G., Monto, M., & Harris, R. (1997). Correlates of adolescent male sexual offense: Prior adult sexual contact, sexual attitudes, and use of sexually explicit materials, International Journal of Offender Therapy and Comparative Criminology, 41, 272-283.

The authors wish to thank Willy Pedersen, Ph.D. and Ingeborg Rossow, Ph.D. for their valuable comments on earlier drafts of this paper.

Svein Mossige, Norwegian Social Research, Oslo, Norway.

Lars Wichstrom, Norwegian University of Science and Technology, Trondheim, Norway.

Request for reprints should be sent to Kristinn Hegna, NOVA, Postbox 3223 Elisenberg, N-0208 Oslo, Norway. E-mail: kristinn.hegna@nova.no

COPYRIGHT 2004 Libra Publishers, Inc.
COPYRIGHT 2005 Gale Group