Same-sex and opposite-sex best friend interactions among high school juniors and seniors.

Lundy, Brenda ; Field, Tiffany ; McBride, Cami 等

Intimate friendship seems to be one of the most salient characteristics of adolescence, more so that in previous developmental stages (Buhrmester, 1990; Richey & Richey, 1980). A study that compared preadolescents with adolescents found that intimate friendship is more important to adolescents (Buhrmester, 1990). The study also noted that competence in peer relationship skills is a greater concern for adolescents than for preadolescents. A review of the literature on friendship revealed that children with close friends show better academic performance, are less likely to drop out of school, and have lower rates of juvenile delinquency and adult psychopathology (Parker & Asher, 1987). Reisman (1985) noted that the DSM-III-R uses lack of close peer relationships as a criterion for a number of childhood psychiatric disorders.

Preadolescent best friend pairs versus acquaintance pairs show more matching of positive behavioral states and experience lower stress levels during their interactions (Field, Greenwald, Morrow, Healy, Foster, Guthertz, & Frost, 1992). Few studies, however, have compared same-sex friendships with opposite-sex friendships for adolescents. Sharabany, Gershoni, and Hofman (1981) investigated changes in peer intimacy of same-sex and opposite-sex friendships during preadolescence and adolescence. A 32-item intimacy questionnaire, which assessed eight dimensions of intimacy (Sharabany Intimacy Scale), was administered to 480 Israeli schoolchildren in the 5th, 6th, 9th, and 11th grades. Half rated their same-sex best friend and the other half rated their opposite-sex best friend on intimacy level. Females reported a greater level of intimacy with their same-sex friend than did males. The intimacy scores for opposite-sex friends were low for both males and females in the 5th grade, but increased after that for females more rapidly than for males. By the 11th grade, the level of opposite-sex peer intimacy was similar to that of same-sex intimacy.

Although intimacy ratings provide information regarding the "comfort level" experienced in friendships, the inclusion of behavioral and physiological measures can provide additional information related to the level of stress during actual peer interactions. In a study using behavioral and physiological measures (McBride & Field, 1997), high school juniors were videotaped in same-sex and opposite-sex best friend pairs during a face-to-face conversation. The videotapes were rated for concordance of behavior states, saliva samples were collected to determine the subjects' cortisol levels before and after the interaction, and the subjects completed questionnaires in which they rated their interaction and the likability and characteristics of their partner. They were also asked to complete self-esteem, peer intimacy, depression, and anxiety scales. The most comfortable, playful interactions were found to be those between females. Females rated their same-sex interactions as more comfortable and their female partners more likable, and they engaged in more playful behavior together than did male-male or opposite-sex dyads. Although it had been anticipated that by their junior year these adolescents would be spending more time in heterosexual relationships and therefore rating them more optimally and showing more playful behavior, that did not appear to be true for this eleventh-grade sample. In terms of behavioral concordance, or synchrony, the adolescents spent similar amounts of time together in an interested state for both same-sex and opposite-sex interactions, and in an animated state for male same-sex pairs and a playful state (the highest level of intimate behavior on the scale) for female same-sex pairs.

The purpose of the present research was to determine (1) whether a "comfort level" similar to that of female-female interactions in a face-to-face situation had developed for male-male interactions one year later, and (2) whether females had come to experience the same level of comfort and matching of playful behavior in their interactions with males by their senior year of high school as they had experienced in same-sex interactions in their junior year.

METHOD

Subjects

Eighteen high school adolescents (11 females, 7 males) who had participated in the McBride and Field (1997) research during their junior year were again studied during their senior year. They were asked during both years to name their best same-sex friend and best opposite-sex friend who attended the same high school. Their best friends were then invited to participate with them. The average age of the adolescents was 16.3 years (range = 15-17) in eleventh grade and 17.3 years (range = 16-19) in twelfth grade. Their ethnic distribution was 33% Hispanic, 28% white, 28% African American, and 11% other, consistent with the high school distribution. Forty-four percent of their parents were married, 94% had at least some college, and 44% had a graduate school degree. Seventy-two percent of their parents earned more than $30,000, and the sample averaged 2.3 (middle to upper middle SES) on the Hollingshead Index.

Procedure

Selection of best friend pairs. The students were asked to provide demographic information (age, sex, race) and the names of their best same-sex friend and best opposite-sex friend within the same high school grade. The questions used to ascertain best friends were: "I spend the most time with _____"; "I know _____ the best"; and "I have lunch with _____ the most." The students named as best friends were then asked to participate and to answer the same questions. The students who designated one another on two out of three of the above questions were matched in pairs.

Same-sex and opposite-sex best friend interactions. The best friend pairs were seated face-to-face across a small table and asked to have a conversation about anything they desired (ideas were suggested, such as talking about their day at school or their summer plans). The 10-minute conversations were videotaped by a tripod-mounted camera placed approximately six feet away from the students so as to be unobtrusive. A mirror was propped next to one student so that the image on the video screen showed the dyad side by side. Microphones were also clipped to the students' shirts to ensure an audible recording. The videotaped interactions were subsequently coded by an observer who rated each adolescent on his or her behavior.

Following the 10-minute interaction, students were administered questionnaires that required 15 minutes to complete. Saliva samples were collected prior to the interaction and following the questionnaire period to assay cortisol levels as an index of arousal level or interaction stress. Because saliva cortisol levels have a 20-minute lag time, the saliva sampled after the questionnaire period reflected cortisol levels midway through the 10-minute interaction. It was speculated that by the middle of the interaction, the adolescents would have adapted to the situation and probably be at their greatest ease. For the sampling of saliva, each student simply placed a dental swab (dipped in lemonade crystals) along his or her gum line for 10 seconds.

Questionnaires

The students were administered two sets of questionnaires. The first consisted of a series of interaction rating scales: how they felt during the interaction, how they rated their interaction partner, and an assessment of level of intimacy with best friend. The second set contained self-rating scales that dealt with self-esteem, anxiety, and depressed mood. The interaction rating scales were administered after both the same-sex and opposite-sex interactions. The self-rating scales were administered only after the first session. The same-sex/opposite-sex interaction order was counterbalanced to minimize any effects on questionnaire responses.

Feelings. Conversation comfort, or how the adolescent felt during the interaction, was assessed using the Feelings Scale (Warner, Malloy, Schneider, Knoth, & Wilder, 1987). Twelve positive items were rated on a 5-point Likert scale: relaxed, friendly, interested, calm, unself-conscious, enthusiastic, confident, involved, happy, in charge of the situation, natural, and pleasant. These items had been generated on a face-validity basis (Warner et al., 1987). The scale was the same as that used in the Field et al. (1992) study.

Partner rating. The adolescents were asked to rate how much they liked their conversation partner on a 5-point Likert scale. The 12 items were the same as those used in the study by Field et al. (1992): (a) the way he/she looks, (b) his/her personality, (c) the things he/she said, (d) the way he/she talks, (e) the way he/she smiles/laughs, (f) his/her ideas, (g) the way he/she listens, (h) the way he/she laughs at things I said, (i) the way we share ideas, (j) the way he/she looks at me, (k) the way we take turns talking, and (1) the way we have fun together talking.

Peer intimacy. The Peer Intimacy Scale (Blyth & Foster-Clark, 1987) was used to assess level of intimacy with best friend (Cronbach's alpha = .85; test-retest reliability = .81). An example of one of the 8 questions is: "How much do you share your inner feelings or secrets with your best friend?" Responses range from not at all to very much, and higher scores signify greater intimacy.

Self-Esteem. The Self-Esteem Scale (Field & Yando, 1991) asks students to compare themselves with peers on 20 descriptors: confident, anxious, happy, fearful, competitive, ambitious, hard-working, good-looking, good in sports, creative, independent, angry, honest, generous, caring, expressive, outgoing, sentimental, good at schoolwork, and moody (Cronbach's alpha = .66; test-retest reliability = .83). Responses include less, same, and more.

Anxiety. The State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970)assesses level of current situational anxiety. It consists of 40 items rated on a 4-point Likert scale (not at all, somewhat, moderately, and very much so).

Depression. The 20-item Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1991) was included to assess depressive symptoms. The scale has been standardized for high school populations (Radloff, 1991), and has adequate test-retest reliability (.80-.90), internal consistency, and concurrent validity (Wells, Klerman, & Deykin, 1987). The adolescents reported their feelings during the preceding week on items representing the primary symptoms of depression.

Coding by Observers

Each of the videotaped interactions was coded for behavior, which ranged from negative to positive states of attentiveness and affect on a linear scale: interested was defined as occasionally looking intently, leaning toward, interested affect, and talking intently; animated was defined as looking intently, leaning toward, animated gestures, exaggerated facial expressions, and fast/animated talk; playful was defined as looking intently, leaning toward with some physical contact, playful/exaggerated affect (e.g., funny faces), laughing while talking, and various forms of playfulness (e.g., mimicry, joking, playing a game). Although each of these mutually exclusive behavior states was operationally defined to include several behaviors (attention, body position, affect, and talking), they were basically intended to capture different levels of attention and affect, ranging from being disengaged to highly engaged.

The coders were blind to dyad classification and were trained to .90 reliability. Intercoder reliabilities for behavior states, based on one-fourth of the observations, were calculated using Cohen's kappa and ranged from .73 to .85 (M = .78). The videotapes were coded in real time for the behavior states of one member of the dyad and subsequently replayed for the coding of the other member. A laptop computer was used to record behavior states by simply pressing a key. A particular state was recorded until a new state was observed and a different key pressed. The coding program provided the second-by-second behavior state of each adolescent in parallel time series (Guthertz & Field, 1989). The program also provided a data matrix of the proportion of time spent in each behavior state by dyad members, both individually and jointly.

RESULTS

Repeated-measures analysis of variance (ANOVA) was conducted with gender as the grouping variable and interaction condition (same-sex/opposite sex) and grade level (eleventh/twelfth) as repeated measures. Interaction effects were subsequently tested by post hoc ANOVAs and Bonferroni t tests.

Self-Rating Scales

A marginal gender by grade interaction effect was found for depression, with the twelfth-grade males having the lowest CES-D scores, F(1, 14) = 3.54, p = .08. Differences were also found between the eleventh and twelfth graders, with the twelfth graders reporting fewer depressive symptoms, F(1, 14) = 4.70, p [less than] .05. No effects were noted for anxiety (STAI) or self-esteem scores. (See Table 1.)

Peer Interaction Rating Scales

A significant condition (same-sex/opposite-sex) effect was found for feelings, F(1, 12) = 9.07, p [less than] .01. The "How I felt" scores were higher for same-sex interactions than for opposite-sex interactions, t(13) = 3.34, p [less than] .005. (See Table 2.)

A group (gender) by condition (same-sex/opposite-sex) interaction effect was found for partner rating, F(1, 13) = 4.58, p [less than] .05. Post hoc tests indicated that the highest partner ratings were assigned by females for same-sex interactions, t(8) = 2.77, p [less than] .05.

A marginal gender by grade interaction effect was found for peer intimacy, F(1, 12) = 4.15, p [less than] .06. The analysis also revealed a significant effect for gender, F(1, 12) = 4.75, p [less than] .05, and grade level, F(1, 12) = 6.35, p [less than] .05. Post hoc tests indicated that twelfth-grade males received the lowest peer intimacy ratings, F(1, 13) = 8.92, p [less than] .01. Males' peer intimacy ratings also decreased between the eleventh and twelfth grades, t(5) = 2.84, p [less than] .05.

Table 1
Mean Scores (and Standard Deviations) for the Self-Rating Scales

 11th Grade 12th Grade
 Male Female Male Female
 Mean (SD) Mean (SD) Mean (SD) Mean (SD)

Depression
(CES-D) 21.86 (7.7) 24.33 (8.3) 15.57 (4.4) 23.89 (8.9)

Anxiety
(STAI) 44.71 (6.8) 42.09 (8.5) 40.29 (7.3) 40.27 (6.7)

Self-Esteem 47.14 (1.8) 44.30 (3.6) 39.29 (13.8) 44.10 (4.7)

Behavior States and State Matching

A gender effect was found for interested state, F(1, 11) = 5.64, p [less than] .05. Males were more often in an interested state than were females. No significant differences were found for synchrony of interested state.

A gender by grade interaction effect was found for animated state, F(1, 11) = 8.40, p [less than] .01. The analysis also revealed a significant effect for grade level, F(1, 11) = 46.38, p [less than] .001, and a marginal effect for gender, F(1, 11) = 4.49, p [less than] .06. Post hoc analyses indicated that twelfth-grade females showed the most animated behaviors, F(1, 12) = 6.62, p [less than] .05. Paired t tests revealed an increase in females' animated behaviors between the eleventh and twelfth grades, t(5) = -2.99, p [less than] .01.

A condition (same-sex/opposite-sex) by grade interaction effect was found for synchrony of animated state, F(1, 11) = 7.38, p [less than] .05. The analysis also revealed a significant effect for condition, F(1, 11) = 11.90, p [less than] .005, and grade level, F(1, 11) = 17.92, p [less than] .001. Post hoc analyses indicated that the greatest synchrony of animated state occurred for same-sex pairs in the twelfth grade, t(14) = -4.31, p [less than] .001.

Another gender by grade interaction effect was found for playful state, F(1, 11) = 5.99, p [less than] .05. The analysis also revealed a significant effect for grade level, F(1, 11) = 74.39, p [less than] .001. Post hoc analyses indicated that eleventh-grade females showed the most playful behaviors, although the effect was only marginally significant, F(1, 12) = 81.80, p [less than] .07. Paired t tests indicated a significant decrease in playful behaviors between the eleventh and twelfth grades for both males, t(5) [TABULAR DATA FOR TABLE 2 OMITTED] = 5.63,p [less than] .001, and females, t(6) = 6.97, p [less than] .001. A significant effect for grade level was revealed for synchrony of playful state, F(1, 11) = 46.14, p [less than] .001. Synchrony occurred less often in the twelfth grade. (See Table 2.)

Table 3
Changes in Cortisol Levels

 Same Sex Opposite Sex
 Male Female Male Female
Grade Mean (SD) Mean (SD) Mean (SD) Mean (SD)

11th -.18 (0.7) -.76 (1.0) .75 (1.0) -.41 (0.8)
12th .08 (0.7) .25 (0.4) .47 (1.1) .08 (0.3)

Note: negative scores indicate & decrease in cortisol levels and
positive scores indicate an increase in cortisol levels.

Cortisol Levels

A marginal gender by grade interaction effect was found for changes in cortisol levels, F(1, 12) = 4.11, p [less than] .07. The analysis also revealed a significant effect for gender, F(1, 12) = 5.59, p [less than] .05, and a marginal effect for grade, F(1, 12) = 3.93, p = .07. Post hoc analyses revealed that the greatest decrease in cortisol levels was for eleventh-grade females during same-sex interactions and the greatest increase was for eleventh-grade males during opposite-sex interactions, F(1, 13) = 5.75, p [less than] .05. Thus, eleventh-grade males demonstrated more arousal or stress during their interactions with females. (See Table 3.)

DISCUSSION

The findings indicate that males in the twelfth grade had fewer depressive symptoms than did eleventh- and twelfth-grade females, as well as fewer symptoms than they had in the eleventh grade. Whether the older males actually experienced fewer symptoms of depression or were less willing to report them cannot be determined. Twelfth-grade males may have felt more inhibited in expressing their feelings as a result of a socialization factor, although many investigators have reported diverging trends for depression in males and females at this age.

As found in the McBride and Field (1997) study, both males and females appeared to feel more comfortable in same-sex interactions. Females also continued to rate their same-sex partner more favorably than they did their opposite-sex partner.

Twelfth-grade males had the lowest peer intimacy scores. This is consistent with the study by Sharabany et al. (1981), in which males were found to develop intimacy mere slowly than did females. According to Sharabany et al., males "de-emphasize the affective components (e.g., emotional support and understanding, trust and loyalty) and stress the instrumental aspects (e.g., they support one another in trouble and meet specific concrete needs)" (p. 801). Further, "in contrast to the girls' socialized need for intimacy, boys are socialized against intimacy" (p. 801).

In the present study, males showed lower interactional involvement (greater interested behavior). Females, in contrast, showed greater involvement (animated behavior), which increased between the eleventh and twelfth grades. Playful behaviors, however, decreased between the eleventh and twelfth grades for both males and females. The increase in animated behaviors and decrease in playful behaviors may reflect increasing inhibition and maturity.

According to the cortisol results, the most stress was experienced by eleventh-grade males during opposite-sex interactions. The eleventh-grade males may have had less experience with opposite-sex interactions than did females and twelfth-grade males. The unfamiliarity of the situation, a face-to-face conversation with a female, may have resulted in greater stress.

The changes appeared to be subtle between high school juniors and seniors. Although females may be socialized to be more intimate, twelfth-grade males and females appeared to display more of the same behaviors and more synchrony of those behaviors. Further, although males and females reported feeling more comfortable in same-sex interactions, they continued to develop comfort in opposite-sex interactions (e.g., the increase in cortisol level was lower for males in the twelfth grade than in the eleventh grade).

The affective, behavioral, and physiological data converged in this study to suggest that female-female interactions were the most involved/intimate, but that both females and males were becoming more involved with the opposite sex as they grew older. To better understand adolescent intimacy, future research might focus on specific behaviors and the dynamics of the more involved female-female interactions and the less involved male-male interactions. In addition, more frequent observations across adolescence would enhance our understanding of the later-developing intimacy in opposite-sex interactions.

The authors gratefully acknowledge the adolescents who participated in this study and the research associates who assisted in the data collection and coding. This research was supported by an NIMH Research Scientist Award (#MH00331) and an NIMH Research Grant (#MH46586) to Tiffany Field.

REFERENCES

Blyth, D. A., & Foster-Clark, F. S. (1987). Gender differences in perceived intimacy with different members of adolescents' social networks. Sex Roles, 17, 689-719.

Buhrmester, D. (1990). Intimacy of friendship, interpersonal competence, and adjustment during preadolescence and adolescence. Child Development, 61(4), 1101-1111.

Field, T., Greenwald, P., Morrow, C., Healy, B., Foster, T., Guthertz, M., & Frost, P. (1992). Behavior state matching during interactions of preadolescent friends versus acquaintances. Developmental Psychology, 28, 242-250.

Field, T. M., & Yando, R. (1991). Adolescents' Self-Perceptions Scales. Unpublished scales.

Guthertz, M., & Field, T. (1989). Lap computer or on-line coding and data analysis for laboratory and field observations. Infant Behavior and Development, 12, 305-320.

McBride, C., & Field, T. (1997). Adolescent same-sex and opposite-sex best friend interactions. Adolescence, 32(127), 515-522.

Parker, J. G., & Asher, S. R. (1987). Peer relations and later adjustment: Are low-accepted children at risk? Psychological Bulletin, 102, 357-389.

Radloff, L. S. (1991). The use of the Center for Epidemiological Studies Depression Scale in adolescents and young adults. Journal of Youth and Adolescence, 20(2), 149-166.

Reisman, J. M. (1985). Friendship and its implications for mental health or social competence. Journal of Early Adolescence, 5, 383-391.

Richey, M. H., & Richey, H. W. (1980). The significance of best-friend relationships in adolescence. Psychology in the Schools, 17, 536-540.

Sharabany, R., Gershoni, R., & Hofman, J. E. (1981). Girlfriend, boyfriend: Age and sex differences in intimate friendship. Developmental Psychology, 17, 800-808.

Spielberger, C. D., Gorsuch, R. C., & Lushene, R. E. (1970). The State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.

Warner, R., Malloy, D., Schneider, K., Knoth, R., & Wilder, B. (1987). Rhythmic organization of social interaction and observer ratings of positive affect and involvement. Journal of Nonverbal Behavior, 11(2), 57-74.

Wells, V. E., Klerman, G. L., & Deykin, E. Y. (1987). The prevalence of depressive symptoms in college students. School Psychiatry, 22, 20-28.