A concept map of male partners in teenage pregnancy: implications for school counselors.

U.S. teen pregnancy rates remain high compared to those of other industrialized nations (Adolescent Health, 1995), despite the drop that has occurred since they peaked in 1991 (Ventura, Clarke, & Matthews, 1996). Because the well-documented social costs are substantial (Coley & Chase-Lansdale, 1998; Maynard, 1997), it is important that teenage pregnancy prevention remain a priority of public health professionals, policymakers, and counselors. Yet, a sometimes-unrecognized barrier to developing more effective prevention, intervention, and policy has been the scant research on the male partners of the young women who become pregnant. Kiselica, Gorcynski, and Capps (1998) confirmed that school counselors need information about teen fathers.

In the absence of research, perceptions of the male partners have remained ill defined and often based on the stereotype that they are sexual predators. This is the implicit suggestion of researchers who employ such terms as the adult liaison (e.g., Males, 1992) to discuss the consistent finding that these males are on average about 3 years older than their teenage female companions who become pregnant (Landry & Forrest, 1995; Lindberg, Sonenstein, Ku, & Martinez, 1997; Males, 1992; Males & Chew, 1996). This perception has been evident as well in the popular press (Roan, 1995) and among public officials who have responded by enforcing statutory rape laws more vigorously (Donovan, 1997).

But there are many pathways to teen pregnancy, and it is reasonable to assume that different types of young men might travel each of these diverse paths. Certainly some male partners do behave as sexual predators. But there is ample anecdotal evidence that many others behave from different motives. Curiously, though, no attempts have been made to develop more differentiated empirical classifications of those males. To have classifications that are valid and empirically grounded could aid school counselors to develop more effective interventions.

This study employed Trochim's (1989; n.d.) concept-mapping procedure to capitalize on the expertise of those people likely to have the best real-world understanding of pregnant teens and their partners: teachers, counselors, and case workers who work directly with them. Trochim's procedure uses multivariate statistics to depict visually a group's understanding of a particular phenomenon or issue. Counseling-related applications of Trochim's procedure have included descriptions of the experience of depression (Daughtry & Kunkel, 1993) and of giftedness (Kunkel, Chapa, Patterson, & Walling, 1995) as well as clients' perceptions of helpful experiences in counseling (Paulson, Truscott, & Stuart, 1999). The goal of this study was to develop a classification of young men who have been responsible for one or more teen pregnancies. To have this knowledge is a first step for school counselors to develop appropriate programs and interventions.

Method

Participants

The 41 (25 female; 16 male) participants reported a median of 3.0 years' direct experience working with pregnant teens and their male partners. Employed in high schools, continuation high schools, social service agencies, and teen outreach programs throughout Los Angeles County, they described their job titles as social work case workers (n = 17), counselors (n = 6), teachers (n = 5), program directors (n = 3), administrators (n = 3), child care technicians (n = 4), outreach workers (n = 2), and psychologist (n = 1). When asked to rate "the extent to which you feel knowledgeable about males who have been responsible for teen pregnancy" using a 7-point scale where 1 = not at all knowledgeable and 7 = extremely knowledgeable, their mean rating was 5.03 (SD = 1.11).

Participants' mean age was 36.00 years (SD = 10.92). Of the 39 who reported ethnicity, the breakdown was as follows: Hispanic (19; 48.7%), African American (9; 23.1%), White (8; 20.5%), Asian American (2; 5.1%), and mixed (1; 2.6%).

Materials

Stimulus materials were brief descriptions (about 120 words for each) of 20 males who had impregnated a teenager. These included information about the males' ages, their partners' ages, ethnicity, relationship history and status, and particulars of the pregnancy. These vignettes were developed by drawing from the approximately 1,900 case descriptions obtained in questionnaires and interviews conducted as part of a larger project. The relationships depicted in these vignettes were chosen to portray the broadest possible array of male and female characteristics and life circumstances.

The vignettes were sent to six counselors and social workers who were highly experienced with teen pregnancies. Each reviewed the vignettes to ensure that they represented the range of circumstances, participant ages, and participant ethnicities with which they were familiar. Based on their feedback, it was necessary to revise only two of the case descriptions to add breadth of description.

For example, one of the vignettes (Case 15) contained the following information:

* He is 23, she is 17; she is currently 7 months pregnant.

* Both are White, non-Hispanic.

* They have known each other for 2 years and 3 months and have been sexually involved for about a year.

* He has completed high school and works as an assistant manager for a fast food franchise.

* He has one child, a son, by a woman who is approximately his age. However, shortly after that baby was born, that woman took up with another man whom she then married. Since then, she has made it very difficult for him to maintain ongoing contact with his son.

* He is candid that he feels a strong loss of his son and that he believes that by having this baby he will "have another chance" at fathering.

* She will continue in high school and plans to marry him within the year.

The other vignettes are available for review online at http://www.usc.edu/dept/couns_psych/ conceptmap.htm.

Procedures

Initial participants were recruited because the researchers knew them to have professional experience with men who have impregnated teenagers. They, in turn, were asked to nominate other professionals they knew to have similar expertise. All participants were paid $35 for a task that required approximately 45 minutes.

Most materials were delivered to participants, then picked up one or two days later. Some participants received and returned materials by mail.

Participants were given two tasks to be completed sequentially. In the first task, participants were given the 20 descriptors, each on a different slip of paper (approximately 3" by 8.5" each) and asked to sort the slips of paper into piles, each of which consisted of descriptions of men who the respondent believed to be "similar in some important way." They were to use as many piles as they believed necessary and to place in any particular pile as many slips of paper as they liked. When they were finished, they were to staple each pile together and then to place all the piles in a manila envelope.

Participants then were to rate each of the 20 descriptors on four items using a 7-point, Likert scale where 1 = strongly disagree and 7 = strongly agree. These items were:

* He was in a committed, loving relationship with his partner at the time she became pregnant.

* He behaved in a predatory way with this woman.

* He is likely to behave as a responsible father to this child.

* This pregnancy likely was important to him as a symbol of masculinity or manhood.

The four items were chosen on the basis of their presence in the literature as descriptors of possible male behavior with respect to teen pregnancies.

Results

Data from the participants' sorting task were entered into a matrix in which a similarity score was recorded for each between-case comparison. Between any two cases, the similarity score consisted of the number of times that participants had paired them together. For example, if eight participants had sorted cases one and two into the same pile, that similarity score was eight; if two participants had put cases one and three into the same pile, that similarity score was two; and so on.

Tochim's (1989; n.d.) mapping procedure requires several steps. The first is to analyze the data via multidimensional scaling (MDS); the second is to analyze it via hierarchical cluster analysis; and the final step is to use the results of these two analyses to create the conceptual map.

Multidimensional Scaling

Trochim's (1989; n.d.) procedure requires that the MDS solution be constrained to two dimensions. The two-dimensional solution obtained in this case had stress and RSQ values of .14 and .94, respectively. Both values suggest that these two dimensions are a good representation of the structure of the data. Stress is a goodness-of-fit measure that can range from zero to one; small values (e.g., zero to .15) indicate a good fit. RSQ, the squared correlation, represents the amount of variance accounted for by the solution.

The solutions are summarized in Table 1 and depicted visually in Figure 1 (in the figure, Dimension One is depicted on the vertical axis; Dimension Two, on the horizontal). To illustrate, note in Table 1 that the man coded as Case 1 was located at .13 on Dimension One and at -.37 on Dimension Two. He is visually depicted in Figure 1 slightly above the horizontal axis and to the left of the vertical axis.

[FIGURE 1 OMITTED]

The dimensions obtained in MDS are similar to the factors obtained in factor analysis. As in factor analysis, there is a subjective step during which the obtained dimension is named based on an examination of the items that load on the dimension. The paragraphs that follow describe the meaning of the two dimensions.

Dimension One: Responsible commitment versus irresponsible noncommitment. Results obtained from both quantitative and qualitative procedures suggest the same description for this dimension. Using quantitative means, the correlations summarized in Table 2 suggest that young men located a tone end of the dimension (i.e., at the top of Figure 1) are described as being in a committed, loving relationship and as behaving in a responsible manner toward the young woman. Young men located at the other end of the dimension (i.e., the bottom of Figure 1) were rated as predatory and as seeking the pregnancy to establish their manhood.

Qualitative examinations of case descriptions selected from either end of the dimension suggested the same interpretation. For example, Cases 7 and 2 are both at what seems to be the responsible commitment end of the dimension whereas Cases 4 and 20 are at the other end of that dimension. Case 7 concerns a Mexican-American 18-year-old who has a 2-year-old with his girlfriend of the same age; they have been dating for 4 years, now live together, and plan to marry. Case 2 concerns a 16-year-old Vietnamese American who has a 3-month-old daughter with his Cambodian-American girlfriend; each lives with his or her respective parents, but they see each other daily and continue in high school. Case 4 concerns a 29-year-old gang member who has three other children with other women; he and the 16-year-old he impregnated had sexual relations only twice, drinking heavily both times. Case 20 concerns a 19-year-old who brags to his friends that he has fathered "at least" two children; however, recently when a girlfriend reported she was pregnant, he called her names and angrily asserted that he "knew" she was sleeping around with other guys and that he could not possibly be the father.

Dimension Two: Low family discord versus high family discord. The characteristics that defined this dimension were more difficult to identify than was true with the first. The dimension did not correlate with any of the four rating scales (see Table 2) and was more difficult to identify using qualitative means. However, after careful review of the cases at either end of the continuum, it appeared that this dimension must address level of discord with family of origin. For example, Case 16 has the most extreme position on the horizontal axis. It concerns a 20-year-old male partnered with a now-pregnant 16-year-old; both were runaways who had met at a homeless shelter about 8 months earlier. He supported them by selling street drugs and eventually they planned to live together after he completed his 6-month sentence for drug possession. Case 9, also on the right side of this axis, concerns a 19-year-old man in a relationship with a 16-year-old woman who moved into his family home after her parents kicked her out following a serious fight. None of the cases on the left side of the continuum exhibited family conflict.

Dimension One versus Dimension Two. The meaning of Dimension One is clearer than that for Dimension Two. As well, Dimension One seems much more useful in differentiating cases--that is, cases array well with the exception of the one extreme outlier (Case 16). There is very little spread of the cases along Dimension Two. Dimension One seems, therefore, to represent the dominant schema according to which participants in this study "mapped" male partners in teenage pregnancies.

Hierarchical Cluster Analysis

Hair and Black (2000) explained that cluster analysis is used to classify "objects" (in this study, the male partners) so that each object in a resulting cluster is similar to others in that cluster. They noted that cluster analysis has been especially helpful to mental health professionals in identifying categories of psychological disorders:

   For example, a researcher may want to know if it is important to identify
   different types of attention deficit hyperactivity disorder. A cluster
   analysis may reveal that a syndrome that specifies attention deficits
   should, perhaps, be separated from a syndrome that emphasizes
   hyperactivity. The cluster analysis can reveal what symptoms discriminate
   the two categories. (p. 147)

Cluster analytic techniques yield a dendogram, which is a tree-like structure that depicts a hierarchy of the "objects" the investigator is seeking to classify. For an example of how this works, examine the cluster analytic results for this study depicted in Figure 2. Each of the 20 men who were the focus of this study is listed at the first step in the analysis (far left) as his own "cluster." That is, at this first step, each of the men is listed alone, by identifying number.

[FIGURE 2 OMITTED]

At the next step, men who are most similar are clustered together; and with each successive step the resulting cluster is more inclusive. For example, men 4 and 18 are joined at the second step in the process; these two, in turn, join man 12 at the next step; then, at the subsequent step, men 4, 18, and 12 are joined by men 16, 3, 20, 6, and 17. Determining the best solution is a matter both of examining the patterns in the data and of considering issues of parsimony and interpretability.

At this stage in the analysis, a tentative "best solution" appears to be one with three clusters. One cluster (Cluster II in Figure 1) is comprised of cases 4, 18, 12, 16, 3, 20, 6, and 17; another cluster (Cluster I in Figure 1) is comprised of cases 11, 19, and 14. However, because it is impossible to make a meaningful distinction between these two in Figure 1, they are joined to make the single Cluster A. The third cluster, Cluster B, is comprised of men 7, 10, 2, 5, 9, 8, 15, 1, and 13.

Several statistical procedures are available for developing the hierarchies. Ward's (1963) method, which Hair and Black (2000) report is biased toward creating clusters with approximately the same number of observations, was used. Data used for analyses were the same used for the MDS analyses.

Combining the MDS and Cluster Analysis Results

The third step in Trochim's (1989; n.d.) mapping procedure is to combine the MDS results with those of the cluster analysis. This is represented in Figure 1, which depicts the concept map of the participants in this study. The areas depicted by the dotted lines define the two clusters that make up Cluster A; the solid lines define Clusters A and B.

Using these results, it is possible to characterize men in Cluster A as more predatory, less loving of and committed to their teenage partners, less likely to be responsible fathers, and more likely to regard the pregnancies as having been symbols of their masculinity. Men in Cluster B were less predatory, more loving of and committed to their teenage partners, more likely to be a responsible father to the child or children, and less likely to regard the pregnancy as having been a symbol of his masculinity.

It is an interesting as well that there were between-group age differences. Whereas the 12 men in Cluster B had a mean age of 18.67 (SD = 2.31), the 10 men in Cluster A had a mean age of 21.88 (SD = 3.93).

Discussion

This was the first study to use practitioners' expert knowledge to classify male partners of teenagers who have become pregnant. The results confirm that it is possible to cluster these men into different types according to their motivations and other attributes. This clustering is of potential use to those school counselors who engage in any of the three broad categories of teenage pregnancy interventions. These are interventions to: (a) prevent teenage pregnancy; (b) help make decisions about how to respond to a newly discovered pregnancy; and (c) help the young men and women with their educational, personal, and interpersonal functioning after a pregnancy.

Prevention

Prevention has the broadest scope of the several interventions. That is, each of the many pathways to teenage pregnancy (e.g., Coley & Chase-Lansdale, 1998) suggests a different prevention possibility. For example, pregnancies will be reduced to the extent that students of both genders are helped to learn about contraception, succeed academically, develop social support networks, develop assertion skills in dating situations, and clarify their life and career goals. Some prevention interventions target one gender. For example, with males, interventions to develop less rigid adherence to traditional conceptions of masculinity may indirectly reduce teenage pregnancies (e.g., Goodyear, Newcomb, & Allison, 2000; Sonenstein, Pleck, & Ku, 1991). Such rigid masculinity beliefs probably are true of men in this study's Cluster A.

Pistole (1999) argued that counselors who wish to prevent teenage pregnancies need first to acknowledge the emotional motivations for the young men and women in the relationships. It is a normal developmental process to begin looking beyond the family to develop new attachments and experience romantic love. Research such as Musick's (1993) suggested that teenagers may look to early sex to compensate for inadequate emotional support or stability. Pistole (1999) argued, therefore, that counselors should help the young people differentiate between needs for romantic love and attachment on the one hand and sexual expression on the other.

Pistole's (1999) model probably is most useful to those teenagers of both genders who are partnered in caring, committed relationships (Cluster B men in this study). Matters become more complicated with respect to relationships in which men of the type found in Cluster A are partners. In some of those cases, the young woman is dealing with attachment issues, regardless of what the young man's motives might be. In other cases, other of the young woman's attributes such as sensation seeking or substance abuse are factors and would be legitimate foci for interventions. For these young men, date rape programming could be a useful intervention.

One intervention possibility targeting women paired with men in Cluster B is suggested by findings in another study (Goodyear, Newcomb, & Locke, in press). It was found that pregnant teenagers' level of psychological distress (manifested as depression and suicidal ideation) strongly predicted their having partnered with a man who demonstrated misogynist attitudes and behaviors (operationalized as a willingness to use coercive sex and to psychologically maltreat women). This finding was consistent with Swann's (1983) self-verification hypothesis that depressed people will seek relationships with people who confirm their relatively negative self-views. One implication seems to be that to the extent that school counselors can help alleviate young women's depression levels, they may also reduce the young women's likelihood of partnering with men whose stance toward women is exploitive and puts them at particular risk of pregnancy.

School counselors interested in developing teenage pregnancy prevention programs that specifically target the male partner can find a superb resource free, online. Sonenstein, Stewart, Lindberg, Pernas, and Williams (1998) not only summarized research literature regarding sexual behavior of teenage males, but provided information including Web links to model teenage pregnancy prevention programs that target males.

Intervention at Time of Pregnancy

For the young woman, the consequences of an unplanned pregnancy are clear and direct. A primary focus of decision making regarding the pregnancy necessarily will be on the young woman and perhaps also her parents and guardians. But among the factors that affect young women's decision making in relationship to an unplanned pregnancy (e.g., Farber, 1991) are the nature and quality of her relationship with her male partner. Young men of the type in Cluster B are more likely to participate in the decision making and to be available to support the young women. In at least some of those cases, a counseling issue will concern the wisdom of this young man and woman partnering through either marriage or living together.

Young men of the type found in Cluster A are less likely to be available during decision making concerning unplanned pregnancies. In fact, in some cases they will themselves be a focus of counseling sessions, particularly if the young woman feels victimized.

Post-Pregnancy Interventions

The third broad category of intervention concerns post-pregnancy issues. These issues could include, for example, feelings of loss and guilt in cases of either abortion or having surrendered the baby for adoption. But at least as many post-pregnancy issues occur if the young woman gives birth and keeps the baby, for the psychological and other costs are substantial (Hotz, McElroy, & Sanders, 1997).

For all young men, an important intervention is to help them retain partnership with the young woman and their child or children. This is especially true of young men in Cluster A, but it applies as well to men in Cluster B, for good intentions alone do not ensure that the young men will remain as romantic partners with the young women. If they have had children together (and have not surrendered them for adoption), it is important that both be available as parents to those children. To help the young men and women retain an effective parenting relationship is an important counseling goal. One finding in a qualitative study (Goodyear, Arcinue, & Getzelman, 2000) was that the young woman's parents often have a great deal of power in affecting whether the young man stays in touch with the young woman and their child. This suggests that in many cases, to involve the young woman's family of origin in counseling may remove a barrier to the young man remaining connected to the young woman. Even if they do not remain romantically involved, it is important for them to be able to work together as parents.

Limitations of the Study and Future Research

It is somewhat surprising that there was not more differentiation of these young men; that is, that there were not more "types" reflected in the data. Perhaps this simply reflects the "cognitive maps" that this study's participants have of the universe of young men who are involved in teenage pregnancies. However, it also is possible that the results reflect some limitation of the stimulus items to which the raters responded. Given the steps taken to ensure that a broad range of cases were presented to participants, it is unlikely that representativeness was a factor in the results. It is possible, however, that the brevity of the descriptions affected the results. For example, if longer case examples had been used, it is possible that more differentiated classification of the men would have occurred. This would be reasonable to address in future research.

Participants included not only counselors, but also several other types of human service professionals as well. Logically, the most salient characteristic of these participants would be their familiarity with young men and women who have been involved in teenage pregnancy; there was no material on the stimulus items that required specialized professional knowledge. Moreover, the fact that these results embody the aggregated responses of participants with somewhat different professional affiliations reasonably makes the findings more generalizable. Nevertheless, because school counselors made up only a minority of the participants it would be useful to replicate this study using only school counselors as participants.

The methods of this study did not permit an estimate of the proportion of the men in each of these categories. It would be possible, though, for future researchers to estimate these proportions if they were to use statistical clustering techniques with either (a) male partners' self-descriptions, or (b) the teenage female partners' descriptions of them. Studies that employed these several data sources should together yield a richer and more complete description of these men.

In short, this study is the first of its type and therefore adds knowledge about the men involved in teenage pregnancies. It also adds to the relatively few articles on teenage pregnancy that have appeared in this journal or its predecessor (e.g., Barnes & Harrod, 1993; Eddy, McCray, Stilson, & DeNardo, 1983; Kiselica et al., 1998). But it is important that subsequent studies replicate and extend this initial work. To the extent that a fully differentiated classification model is developed, school counselors will have a more effective research base to guide their interventions.

Table 1. Dimension Loadings of the 20 Cases

                Dimension 1    Dimension 2
                 Loading         Loading

Case 1              .13          -.37
Case 2             1.6           -.63
Case 3            -1.02          -.25
Case 4            -1.81          -.65
Case 5             1.30           .16
Case 6            -1.50          -.40
Case 7             1.76           .60
Case 8              .11           .12
Case 9              .83           .48
Case 10            1.23          -.20
Case 11             .68          -.46
Case 12           -1.11          -.33
Case 13             .80           .32
Case 14            1.18          -.04
Case 15            1.22           .09
Case 16           -1.71          2.82
Case 17            -.99          -.24
Case 18           -1.22          -.40
Case 19             .05          -.33
Case 20           -1.58          -.32
Table 2. Between-Variable Correlations

                         1         2      3

1 MDS Dimension 1        --

2 MDS Dimension 2       .20        --

3 Committed, loving     .88 ***   .25    --
  relationship

4 Predatory behavior   -.84 ***  -.06   -.90 ***

5 Responsible           .89 ***   .39    .89 ***

6 Masculinity symbol   -.70 ***   .12   -.72 ***

                         4          5       6

1 MDS Dimension 1

2 MDS Dimension 2

3 Committed, loving
  relationship

4 Predatory behavior     --

5 Responsible          -.76 ***     --

6 Masculinity symbol    .85 ***   -.56 **   --

*** p < .001 ** p < .01.

References

Barnes, N. D., & Harrod, S. E. (1993). Teen pregnancy prevention: A rural model using school and community collaboration. The School Counselor, 41, 137-140.

Coley, R. L., & Chase-Lansdale, L. (1998). Adolescent pregnancy and parenthood: Recent evidence and future directions. American Psychologist, 53, 152-166.

Daughtry, D., & Kunkel, M. A. (1993). Experience of depression in college students: A concept map. Journal of Counseling Psychology, 40, 316-323.

Donovan, P. (1997). Can statutory rape laws be effective in preventing adolescent pregnancy? Family Planning Perspectives, 29, 30-34.

Eddy, J., McCray, E. H., Stilson, D., & DeNardo, N. (1983). Pregnancy counseling for teenagers. The School Counselor, 30, 398-402.

Farber, N. B. (1991). The process of pregnancy resolution among adolescent mothers. Adolescence, 26, 697-716.

Goodyear, R. K., Arcinue, F. A., Getzelman, M. A. (2000, April). Male partners in teen pregnancy: Lessons from their stories. Paper presented at the annual meeting of the American Educational Research Association in New Orleans.

Goodyear, R. K., Newcomb, M. D., & Allison, R. D. (2000). Predictors of Latino men's paternity in teen pregnancy: Test of a mediational model of childhood experiences, gender role attitudes, and behaviors. Journal of Counseling Psychology, 47, 116-128.

Goodyear, R. K., Newcomb, M. D., & Locke, T. F. (in press). Pregnant Latina teenagers: Psychosocial and developmental determinants of how they select and perceive the men who father their children. Journal of Counseling Psychology.

Hair, J. F., & Black, W. C. (2000). Cluster analysis. In L. G. Grimm & P. R. Yarnold (Eds.), Reading and understanding more multivariate statistics (pp. 147-206). Washington, DC: American Psychological Association.

Hotz, V. J., McElroy, S. W., & Sanders, S. G. (1997). The costs and consequences of teenage childbearing for mothers. In R. A. Maynard (Ed.), Kids having kids: Economic costs and social consequences of teen pregnancy (pp. 55-94). Washington, DC: Urban Institute.

Kiselica, M. S., Gorcynski, J., & Capps, S. (1998). Teen mothers and fathers: School counselor perceptions of service needs. Professional School Counseling, 1(2), 146-152.

Kunkel, M. A., Chapa, B., Patterson, G., & Walling, D. D. (1995). The experience of giftedness: A concept map. Gifted Child Quarterly, 39, 126-134.

Landry, D. J., & Forrest, J. D. (1995). How old are U.S. fathers? Family Planning Perspectives, 27, 159-161.

Lindberg, L. D., Sonenstein, F. L., Ku, L., & Martinez, G. (1997). Age differences between minors who give birth and their adult partners. Family Planning Perspectives, 29, 61-66.

Males, M. (1992). Adult liaison in the "epidemic" of "teenage" birth, pregnancy, and venereal disease. Journal of Sex Research, 29, 525-545.

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

Maynard, R. A. (Ed.). (1997). Kids having kids: Economic costs and social consequences of teen pregnancy. Washington, DC: The Urban Institute.

Musick, J. S. (1993). Young, poor, and pregnant: The psychology of teenage motherhood. New Haven, CT: Yale University.

Paulson, B. L., Truscott, D., & Stuart, J. (1999). Clients' perceptions of helpful experiences in counseling. Journal of Counseling Psychology, 46, 317-324.

Pistole, M. C. (1999) Preventing teenage pregnancy: Contributions from attachment theory. Journal of Mental Health Counseling, 21, 93-112

Roan, S. (1995, July 10). The invisible men (role of adult men in teen pregnancy). Los Angeles Times, p. E1.

Sonenstein, F. L., Pleck, J. H., & Ku, L. C. (1991). Levels of sexual activity among adolescent males in the United States. Family Planning Perspectives, 23, 162-167.

Sonenstein, F. L., Stewart, K., Lindberg, L. D., Pernas, M., & Williams, S. (January 1998). Involving males in preventing teen pregnancy: A guide for program planners. Washington, DC: Urban Institute. Retrieved May 6, 2001 from the World Wide Web: http://www.urban.org/family/invmales.html

Swann, W. B. (1983). Self-verification: Bringing social reality into harmony with the self. In J. Suis & A. G. Greenwald (Eds.), Social psychological perspectives on the self (Vol. 2; pp. 33-66). Hillsdale, NJ: Erlbaum.

Trochim, W. (1989). An introduction to concept mapping for planning and evaluation. Evaluation and Program Planning, 12(1), 1-16.

Trochim, W. (n.d.). An introduction to concept mapping for planning and evaluation. Retrieved May 2, 2001 from the World Wide Web: http://trochim.human.comell.edu/research/epp1/epp1.htm

Adolescent Health: State of the Nation--Pregnancy, sexually transmitted diseases and related risk behaviors among adolescents. (1995). Atlanta: U.S. Department of Health and Human Services, Public Health Service; CHHS publication no. (CDC) 099-4630.

Ventura, S. J., Clarke, S. C., & Matthews, T. J. (1996, December 19). Recent declines in teenage birth rates in the United States: Variations by state, 1990-94. Monthly Vital Statistics Report, 45(5): Supplement.

Ward, J. H. (1963). Hierarchical grouping to optimize objective function. Journal of the American Statistical Association, 58, 236-244.

Rodney K. Goodyear, Ph.D., is professor and chair, Division of Counseling Psychology, University of Southern California, Los Angeles. E-mail: [e-mail]

This research was funded by the California Wellness Foundation.