Findings from a Controlled Field Trial

Transcription

1 A Health Belief Model-Social Learning Theory Approach to Adolescents Fertility Control: Findings from a Controlled Field Trial Marvin Eisen, PhD Gail L. Zellman, PhD Alfred L. McAlister, PhD We evaluated an 8- to 12-hour Health Belief Model-Social Learning Theory (HBM-SLT)- based sex education program against several community- and school-based interventions in a controlled field experiment. Data on sexual and contraceptive behavior were collected from 1,444 adolescents unselected for gender, race/ethnicity, or virginity status in a pretest-posttest design. Over 60% completed the one-year follow-up. Multivariate analyses were conducted separately for each preintervention virginity status by gender grouping. The results revealed differential program impacts. First, for preintervention virgins, there were no gender or intervention differences in abstinence maintenance over the follow-up year. Second, female preintervention Comparison program virgins used effective contraceptive methods more consistently than those who attended the HBM-SLT program (p < 0.01); among males, the intervention programs were equally effective. Third, both interventions significantly increased contraceptive efficiency for teenagers who were sexually active before attending the programs. For males, the HBM-SLT program led to significantly greater follow-up contraceptive efficiency than the Comparison program with preintervention contraceptive efficiency controlled (p < 0.05); for females, the programs produced equivalent improvement. Implications for program planning and evaluation are discussed. This demonstration project was supported by grants and contracts from the Texas Department of Human Services; The University of Texas at Austin Research Institute; The Lyndon B. Johnson School of Public Affairs, University of Texas at Austin; The Hogg Foundation for Mental Health; The William and Flora Hewlett Foundation; and NICHD grant HD The opinions and conclusions expressed herein are solely those of the authors. The authors are grateful to J.J. Card and Brent Miller for helpful comments on an earlier version of this paper; and to Robert Timothy Reagan and Max Nelson-Kilger for statistical and conceptual consultation. Marvin Eisen is Principal Research Scientist at the Sociometrics Corporation, Los Altos, California. Gail Zellman is a Research Psychologist at the Rand Corporation, Santa Monica, California. Alfred McAlister is Associate Director, Center for Health Promotion, UT HSC, Houston, Texas. Address reprint requests to Marvin Eisen, Sociometrics Corporation, 170 State Street, Suite 260, Los Altos, CA Health Education Quarterly, Vol. 19(2): (Summer 1992) 1992 by SOPHE. Published by John Wiley & Sons, Inc. CCC /92/ $04.00

2 250 INTRODUCTION In 1985 we reported the results of a pilot community-based sexuality education program that combined elements of the Health Belief Model (HBM) and Social Learning Theory (SLT). Those data revealed that most preintervention virgins remained abstinent through the follow-up period, consistent postprogram contraceptive usage increased significantly, and changes in HBM-based contraceptive perceptions and sexual knowledge at immediate posttesting predicted increases in contraceptive usage at the three- to six-month follow-up.2,3 Those encouraging findings led us to conduct a longer-term controlled field trial, the results of which are reported here. The HBM-SLT intervention was intended to increase teenagers awareness of: (1) the probability of personally becoming pregnant or causing a partner to become pregnant; (2) the serious negative personal consequences of teenage pregnancy; (3) the personal and interpersonal benefits of delayed and/or protected sexual activity; and (4) to decrease their perceptions of the barriers to abstinence and consistent contraceptive usage. The main objectives of the demonstration were: (1) to increase adolescents knowledge of, and motivation for, abstinence or consistent contraceptive use; (2) to measure the effect of changes in knowledge and motivation on changes in coital and contraceptive behavior over a one-year follow-up period; and (3) to compare the impact of the HBM-SLT-based intervention model against a range of community outreach and school-based interventions. It was hypothesized that the HBM-SLT (Experimental) program, relative to Comparison programs, would lead to fewer transitions from virginity to sexual activity, and a higher incidence of effective contraceptive use over the one-year followtup. EXPERIMENTAL DESIGN AND METHODS Overview of the Study Design The design involved a randomized field trial. Seven agencies that offered family planning services and one school district compared their own &dquo;usual care&dquo; outreach or regular classroom curriculum with our HBM-SLT intervention. Each agency recruited its study sample by its usual methods; the school district used its eighth and ninth grade population. Within the age range 13-19, participants were unselected with respect to gender, race/ethnicity, and preintervention virginity status. At each site adolescents were randomly assigned by classroom or individually (depending on agency operating constraints) to the HBM-SLT (Experimental) or the agency s usual (Comparison) program. Data were collected from individuals at three study points: prior to the intervention (Time 1); immediately afterwards (Time 2); and 12 months after the program completion date (Time 3). Sample Agencies Agencies were recruited from a pool of organizations that conducted sexuality education outreach programming with state reimbursement agreements in Texas

3 251 and with &dquo;innovative&dquo; information and education grants in California. The agencies included organizations ranging from an urban Planned Parenthood affiliate to rural community action program health clinics. A small independent school district in northern California also participated. The agencies and the school district recruited all teenagers who took part in the study, organized and coordinated the data collection using project materials, and delivered the actual intervention, and were paid to do so. Adolescent Participants Teenagers were recruited from agencies usual &dquo;client&dquo; groups. With the exception of the school system, all agency programs had family income ceilings or served low income, inner-city youth. A summary of study site and sample characteristics is presented elsewhere. 4,1 Between June 1986 and August 1987, 1,444 adolescents (mean age 15.5 years) completed the Time 1 interview and 1,328 (92%) received all or part of the Experimental or Comparison program and completed the Time 2 measures. Providers attempted to reinterview all adolescents who were exposed to any phase of the intervention 12 months after the completion date. Between July 1987 and September 1988, 888 participants (62% of the Time 1 sample and 67% of the Time 2 sample) were reinterviewed. Of the 1,444 Time 1 interviewees, 52% were females; 15% were white, 24% were black, 53% were Hispanic, 8% were Asian (including recent refugees from Cambodia, Laos, and South Vietnam); 62% reported some previous sexuality education; 37% reported they had had sexual intercourse; 49% of those said they used contraception at their last intercourse; and of the teens who used any method, 74% used a condom. Educational Curricula The HBM Curriculum The HBM-SLT curriculum, described elsewhere, included five lecture hours on reproductive biology and eight to ten hours of small group discussions. 2,4,5 The curriculum was transmitted through lectures, simulations, &dquo;trigger&dquo; films, leader-guided discussions and practice, and role-playing (including gender role reversals) based on participant-generated scripts. All activities were designed to increase participants fertility control behavioral skills and their sense of selfefficacy to perform appropriate behaviors through immediate feedback and verbal reinforcement by group discussion leaders. However, there were no specific self-efficacy enhancement exercises. 4,1 A training manual for group discussion leaders was prepared. The training manual and tapes, a group discussion curriculum guide, and the reproductive biology curriculum were employed in two-day training workshops for agency educators and school staff who would deliver the HBM-SLT intervention. * *The same people delivered their &dquo;usual&dquo; (i.e., the Comparison program) intervention in each site, thus serving as their own controls.

4 252 These were similar to the HBM-SLT curriculum and to each other in covering reproductive biology, contraception, sexually transmitted diseases (STDs), and sexual decision-making. They differed from the HBM-SLT intervention in not focusing on the four major perceptual components of the HBM, and in encouraging only limited role-playing or other active student involvement. Brief summaries of the duration, content, and format of Comparison programs, which differed one from another, are presented elsewhere,4.s Evaluation Instruments To assess program impact, an instrument was developed which assessed sexuality-related beliefs, attitudes, knowledge, and behaviors. Standard sociodemographic and educational items were included at Time 1. In addition, Time 1 virgins indicated whether they expected to become sexually active in the next three months. Items written for the pilot study which tapped HBM perceptions were subjected to scaling and psychometric studies,2.4 revised, and augmented with new items based on the pilot findings. Table 1 presents the final set of items (grouped by HBM perceptual construct). In addition, three items (nos. 1, 16, 18) were intended to assess aspects of self-efficacy. Sexual and contraceptive knowledge items discussed elsewhere2~s included reproductive physiology knowledge, pregnancy and sexuality myths, birth control and STD knowledge, and birth control method and STD prevention effectiveness. Finally, we collected behavioral measures: whether adolescents had sexual intercourse; whether they used a contraceptive method(s) at first and most recent sexual intercourse; what specific methods were used, if any; how consistently contraception was used before and after the intervention; and whether they or their partner(s) had become pregnant since the intervention program concluded. Abstinence continuation, transition from virginity to coital activity, and use of and changes in use of effective birth control methods at first and most recent postintervention intercourse were assessed- Procedure Participants usually were interviewed individually at Time 1 in each site. Upon completion of the intervention they were immediately retested on their knowlteffective contraceptive methods included pill, condom, diaphragm, foam/jelly, and sponge; ineffective methods were withdrawal, rhythm, and douched To assess teenagers consistent use of effective methods, we computed a weighted composite contraception index. The Contraceptive Efficiency index for participants with coital experience at Time 1 was computed by the formula: (effective method at first preintervention sex [no 0; yes 1] + effective method at last preintervention sex (no 1; yes 3)) x Time 1 consistency (range: 1-5) Time 1 score (range: 1-20). The index for all Time 3 nonvirgins was: (effective method at first postintervention sex [no 0; yes 1] + effective method at last postintervention sex (no 1; yes 3)) x Time 3 consistency (range 1-5) Time 3 score (range: 1-20).

5 253 Table 1. Health Belief Model Items Grouped by Construct Used to Assess Pregnancy and Birth Control Use Perceptions (N 35)d (table continued on next page)

6 254 Table 1. (Continued) a Response categories are five points: (1) strongly agree to (4) strongly disagree; (8) not sure, except Items 3, 5, and 6. Response categories for Item 3 are: (1) very likely to (4) very unlikely; (8) not sure. Response categories for Items 5 and 6 are: (1) very worried to (4) not at all worried; (8) not sure. edge and HBM perceptions in a group setting. One year after the date of intervention program completion, participants were reinterviewed individually following up to five attempts to make contact by phone, personal visit, or letter. Plan of Analysis Our overall analysis plan involved a four-group approach: we examined Time 1 virgins abstinence, transition to coital activity, and initiation of contraceptive use separately for each gender; we examined changes in Time 1 nonvirgins contraceptive behavior separately for each gender as well. This approach allowed for the likelihood that different issues would surround the initiation of sexual intercourse, the initiation of contraceptive use, and the maintenance of contraceptive use. Moreover, it enabled us to take into account the contraceptive histories of the already sexually active. Thus, for Time 1 nonvirgins we determined whether exposure to either program significantly increased contraceptive use from Time 1 to Time 3, and whether the Experimental program was significantly more effective than the (combined sites) Comparison programs with Time 1 contraceptive use, Time 2 sex knowledge and HBM perceptions, and demographic variables controlled. For preintervention virgins who remained virgins during the follow-up year, the analysis focused on detecting treatment group differences in abstinence maintenance ; for those who became sexually active, we examined contraceptive use and the role of treatment, Time 2 HBM scales and sex knowledge, and demographics in predicting adoption and continuation of fertility control behaviors.

7 255 Comparability RESULTS Across Interventions and Sites There were no significant randomization differences in Time 1 sexual behavior or contraceptive use between HBM-SLT and Comparison conditions., Moreover, data could be combined across sites because there were no significant differences between any two sites on salient Time 3 outcome variables for either gender.4.s HBM Scales Construction The 35 HBM items were first grouped by the four HBM constructs; then Seriousness and Benefits were each divided into two dimensions, as shown in Table 1. These items were subjected to a confirmatory factor analysis using a principal components approach and oblique rotation. Based on standard criteria,~ the original six groupings were judged to form an adequate basis for constructing scales. Internal-consistency reliability extimates for each (Table 2) revealed that five of six scales were sufficiently reliable for making group comparisons (alpha > 0.50), but the three-item self-efficacy subscale was not. Time 1 to Time 2 Changes in Sexual Knowledge and Health Beliefs In bivariate analyses, all sexual knowledge and HBM scales showed statistically significant increases for both the Experimental and Comparison intervention groups ( ps < ). In multivariate analyses, Experimental participants had greater Time 2 sex knowledge than Comparisons (p < 0.05) with Time 1 knowledge and demographics controlled; however, Experimental and Comparison participants did not differ on any Time 2 Health Beliefs.4.s Time 3 Changes in Coital Activity Status by Time 1 Virgins Among the 567 Time 1 virgins who provided data at the three collection points, 167 (29.5%) reported becoming coitally active since Time 1. Thus, 400 (70.5%) teens maintained their abstinence. In addition, 39 (15%) of 262 Time 1 nonvirgins reported no sexual intercourse during the follow-up period. Bivariate analyses revealed no treatment nor age group differences in abstinence maintenance. Females were more likely to remain abstinent than males (77% vs. 61%, p < 0.001) and race/ethnicity differences were associated with continuing abstinence (whites 66%, blacks 62%, Hispanics 73%, Asians 79%, p < 0.06). Contrary to our hypothesis, HBM-SLT program males exhibited no greater continued abstinence than the Comparison males when relevant variables were controlled (Table 2, right half). Four variables-being nonblack, attending

8 256 Table 2. Final Step in Regressions of Time 3 First Intercourse, for Time 1 Virgins on Selected Demographic, Cognitive, Health Belief, and Treatment Variables (Females and Males) a Dummy variable: Yes 1; No 0 (relative to all others, i.e., Comparison group). b Yes 0; No 0. C HBM-SLT 1; Comparisons -1. p < 0.05; **p < 0.01; z < church more frequently, being at or above grade level for one s age, and having less expectation of becoming sexually active in the next three months-were significant predictors of abstinence stability. About 13% of the variance was accounted for by this model. For females, another four-variable model, also without the hypothesized HBM-SLT effect, emerged as the best fit (Table 2, left half). Having higher educational aspirations, being below grade level for one s age, having less expectation of becoming sexually active in the next three months, and perceiving fewer Benefits of Effective Birth Control Use at Time 1 predicted abstinence. This model accounted for 8% of the variance in females Time 3 virginity status. Time 3 Initiation of Contraceptive Use for Time 1 Virgins Bivariate analyses for males revealed no differences between HBM-SLT and Comparison participants in use of an effective method at first intercourse (64% vs. 71%), use of an effective method at most recent intercourse (55% vs. 61%), or in Contraceptive Efficiency at follow-up (Time 3 means 9.87 vs ). There was no HBM-SLT effect on Contraceptive Efficiency (Table 3, right half). Blacks, those who had lower educational aspirations, those who had a prior sexuality education class, and those who saw fewer Barriers to Birth Control Use at Time 1, and fewer Time 1 Interpersonal Benefits of Birth Control Use were better contraceptors than other participants. Those five variables accounted for almost 24% of the variance in males Contraceptive Efficiency scores.

9 Table 3. Final Step in Regressions of Time 3 Contraceptive Efficiencya for Time 1 Virgins on Selected Demographic, Health Belief, Knowledge, and Treatment Variables (Females and Males) 257 a Contraceptive Efficiency is a weighted composite of consistency x effective method (first + last sex). b Dummy variable: Yes 1; No 0 (relative to all others, i.e., Comparison group). Yes 1;No 0. d HBM-SLT 1; Comparisons -1. * p < 0.05; **p < Fewer females in the HBM-SLT group reported using an effective method at their last sexual encounter than in the Comparison groups (38% vs. 62%, p < 0.07). The groups also differed on Contraceptive Efficiency scores over the follow-up (HBM-SLT mean 7.41; Comparison mean 11.21, p < 0.05). Multivariate analyses confirmed the bivariate treatment differences on Contraceptive Efficiency for females (Table 3, left side). There also were significant positive effects of being older at first intercourse and having more favorable Time 2 perceptions of Benefits of Effective Birth Control Use (with Time 1 perceptions controlled). The three variables accounted for about 19% of the variance in Contraceptive Efficiency. Time 3 Changes in Contraceptive Use for Time 1 Nonvirgins Among males, bivariate analyses showed that Contraceptive Efficiency scores improved significantly for both the HBM-SLT group (Time 1 mean 8.78; Time 3 mean 12.08, p < 0.005) and the Comparison groups (Time 1 mean 7.69; Time 3 mean 9.58, p < 0.10). However, neither condition showed significant improvement in effective contraceptive use at most recent intercourse.

10 258 U :E m 1-0 b0 o jn O &dquo; 0 h u (I) T UJ co y- m <~ (I)~ &dquo;u ~ 11 Wz ~ ~ ; % $ C1:S -- m j ~ ~ 0- U M cc ~ U E«U S S 6iJ (I). ~ c&dquo; O o ~g Z C1:S - ~ E &dquo; &dquo; - > o -c o u.- s 1-0 (I) g 01)0 Q F/- <U (u... a. ~ E~ E u r/5 co a 01) -~- <u 3.D 0 P ~ X <u co + s C 0 J3 ~ S S o0 a~ > o U ~.r.i cd in. u g v u 0 X «~ 4j.&horbar; % j ~, C M 0 ~3 0 ~ ~ O ~L S s 045. v-4 3 o II ~: 0 co z 0 jj &dquo; # ~ ~ i. ~s &dquo;$> Qj O u.-v 0 co Il... C t 8 zq_ fl9y ct E VUAx n. m m w *

11 As expected, HBM-SLT program males reported greater improvement in Contraceptive Efficiency than those in the Comparison groups with other variables controlled (Table 4). Teens who were more efficient contraceptors and had greater sex knowledge at Time 1, those who were non-hispanic, and those who attended the HBM-SLT program were better contraceptors. Approximately 6% of the variance in Time 3 Contraceptive Efficiency was accounted for by Time 1 Efficiency; exposure to the HBM-SLT program contributed only 3% to the total variance explained (Step 3, p < 0.05). Together, the four variables accounted for 15% of the variance in contraceptive use. HBM-SLT program females showed no improvement in Contraceptive Efficiency (Time 1 mean 8.51; Time 3 mean 9.94) nor in use of an effective method at most recent sex. Contraceptive Efficiency in Comparison females increased from a mean of 9.06 to (p < 0.05). However, effective method use at most recent sexual encounter did not improve for those in Comparison groups either. Contrary to our hypothesis, the HBM-SLT and Comparison programs were equally effective in improving females Time 3 Contraceptive Efficiency (Table 5). Those who were more efficient contraceptors at Time 1, those who had sex education before Time 1, and those who reported fewer Time 1 Barriers to Birth Control Use were more efficient Time 3 contraceptors. 259 Time 3 Incidence of Pregnancy Responsibility or Pregnancy Five percent of males (11/212) and 10% of females (18/177) reported some personal involvement in a pregnancy. The small number of conceptions pre- Table 5. Stepwise Regression of Time 1 Nonvirgins Time 3 Contraceptive Efficiencya on Selected Personal History, Time 1 Contraceptive Usage, Health Belief, and Treatment Variables (Females, N 92) a Contraceptive Efficiency is a weighted composite of consistency (first + last sex). b Yes 1 ;No 0. I HBM-SLT 1; Comparisons -1. * p < 0.05 ; **p < 0.01 ; ***p < x effective method

12 260 cluded multivariate analyses; however, it was clear that there were no differences between treatment programs in pregnancy responsibility for males (HBM- SLT 7%; Comparison 4%) or pregnancy for females (HBM-SLT 9%; Comparison 12%). DISCUSSION Summary of Program Outcomes Preintervention virginity or nonvirginity and gender were important mediators of treatment group effects on the key fertility control measures. For preintervention virgins the HBM-SLT and Comparison programs led to equal likelihood of maintaining their abstinence over the follow-up year. Female preintervention virgins who attended Comparison programs were more likely to use an effective contraceptive method at most recent intercourse and to be more efficient contraceptors than those attending the HBM-SLT program ; however, further examination of method use showed that they relied heavily on condoms-primarily a male method (condom only 60%; pill and condom 10%; pill only 30%). Thus, it is not clear to what extent females contraceptive decision-making and choices were determined or affected by exposure to the Comparison programs and to what extent by their partners. For males who were virgins at Time 1, the intervention programs were equally effective. Only one Time 2 Health Belief scale, Benefits of Effective Birth Control Use, predicted Contraceptive Efficiency-and this for females only. Both interventions significantly increased Contraceptive Efficiency from Time 1 to Time 3 for preintervention nonvirgins. For males, the HBM-SLT program led to significantly greater follow-up Contraceptive Efficiency than the Comparison programs. The Experimental and Comparison interventions produced equivalent improvement for females. Finally, no Time 2 Health Beliefs appeared to mediate improvement in Contraceptive Efficiency for Time 1 nonvirgins of either gender. The proportion of variance accounted for by the intervention programs and Time 2 Health Beliefs was lower than expected for both males and females. Three factors contributed. First, being sexually active was a relatively low incidence behavior in the study sample. More than half of follow-up participants were still virgins. Second, the study compared the HBM-SLT intervention against other programs rather than assessing it against a &dquo;no treatment&dquo; control group. Third, the HBM-SLT program was cut to a maximum of 12 hours in every site, while most agencies stretched or adjusted their regular programs to match the time allocated for the Experimental program. Thus, the HBM-SLT and Comparison programs became very similar in terms of number and length of sessions-and probably their potential for impact. Study Limitations Two important limitations to the study must be noted. First, the design did not include a true &dquo;no treatment&dquo; control group by intention of the original funders. However, even if one had been desired, there is ample evidence from

13 261 recent national surveys that it would have been impossible to achieve. By 1988 more than 90% of adolescent (15-19) males and 80% of females reported exposure to some formal sex educationll (also, F. Sonenstein, personal communication) ; in the present low income sample almost two-thirds had had some previous sex education. Moreover, because there are no published reports of sex education interventions that incorporated an untreated comparison group and followed teenagers for a year to assess changes in contraceptive behavior, 5,11 it is difficult to estimate the absolute impact of the intervention programs studied here. Second, one-third of intervention participants were not available for reinterview at one-year follow-up. However, less than 5% of those who were located actually refused to participate; almost all teenagers who were not reinterviewed had moved away. Extensive attrition analyses revealed that among males, more Time 1 nonvirgins than virgins left the sample at Time 3, but they were not poorer contraceptors than those nonvirgins who remained in the study. Among females, no sexual behavior variables significantly differentiated Time 1 virgins or Time 1 nonvirgins who completed the follow-up interview from those who did not complete it.1,5 Study Implications Differences in program impact as a function of previous sexual experience, gender, race/ethnicity, and prior sexuality education suggest that intervention programs need to be client-group specific-one size does not fit all. In the present demonstration, the HBM-SLT curriculum appeared to work best with males who were sexually experienced at Time 1, especially in increasing consistent use of effective methods (including condoms). This finding may have important implications for developing effective acquired immunodeficiency syndroms (AIDS) education efforts for high risk males. One hypothesis for the relatively greater success with males in the HBM-SLT program is that its active involvement and role-playing format forced them to examine and think about their dating and sexual interactions with females in new ways. When they did, they gained new awareness of the pregnancy risks they and their partners face (Experimental males scored higher on the Time 2 Susceptibility to Pregnancy and STDs scale than did Comparison males), as well as some means with which to improve their pregnancy avoidance behaviors. The Comparison programs seemed to be most effective with females who were not yet sexually active at Time 1. In contrast to males, females-even the youngest ones in our sample-may have been saturated with the threat of pregnancy (the perceived susceptibility and seriousness elements in the HBM). Consequently, they may have felt that they had less to &dquo;learn&dquo; from the HBM-SLT intervention. Significantly higher scores reported by Comparison group females than by Experimental females or by Comparison males on the Time 2 Susceptibility to Pregnancy and STDs scale seem to support this notion. Future demonstrations that apply an HBM-SLT approach should aim to be client-specific. Refined and strengthened curriculum components and restoration of the HBM program to its originally intended hours may also improve the success of the intervention.

14 262 References 1. Rosenstock I, Strecher V, Becker M: Social learning theory and the health belief model. Health Educ Q 15: , Eisen M, Zellman G, McAlister A: A health belief model approach to adolescents fertility control: Some pilot program findings. Health Educ Q 12: , Eisen M, Zellman G: Changes in the incidence of sexual intercourse of unmarried teenagers following a community-based sex education program. J Sex Res 23: , Eisen M: Testing an Intervention Model for Teen Fertility Control, Final Grant Report to NICHD. Los Altos, Sociometrics Corporation, Eisen M, Zellman G, McAlister A: Evaluating the impact of a theory-based sexuality and contraceptive education program. Fam Plan Perspect 22: , Sonenstein F, Pleck J, Ku L: Sexual activity, condom use and AIDS awareness among adolescent males. Fam Plan Perspect 21: , Howard M, McCabe J: Helping teenagers postpone sexual involvement. Fam Plan Perspect 22:22-26, Eisen M, Zellman G: The role of health belief attitudes, sex education, and demographics in predicting adolescents sexuality knowledge. Health Educ Q 13:9-22, Bernstein I, Keith J: Reexamination of Eisen, Zellman, and McAlister s health belief model questionnaire. Health Educ Q 18: , Helmstadter G: Principles of Psychological Measurement. New York, Appleton- Century-Crofts, Inc, Eisen M: Unpublished tabulations from the National Survey of Family Growth, Cycle 4, Los Altos, Sociometrics Corp., Kirby D: Sexuality Education: An Evaluation of Programs and Their Effects. Santa Cruz, Network Publications, 1984.