Guiding Good Choices

Study 1

Summary

Spoth et al. (1995) utilized a randomized controlled trial to examine 209 families assigned to an intervention group or a waitlist control group. Assessments at baseline and posttest measured parenting quality, parent-child relationship, youth self-esteem, youth affiliation with prosocial peers, and youth intentions and attitudes towards alcohol.

Spoth et al. (1995) and additional reports found that compared with the control group, the intervention group had:

• Improved quality of parent-child relationship, but only significant for fathers
• Better proactive communication from parent to child (protective factor)
• Improved intervention specific and general child management skills (protective factors)
• Reduced negative interaction with child, but only for mothers (risk factor)

Evaluation Methodology

Design: Families who were eligible for special financial assistance through a federally supported school lunch program, and who had a sixth or seventh grade student were recruited from six rural Midwestern school districts (nine public schools). Three hundred eighty-seven families were eligible for participation in the study. Of the families that agreed to participate, 209 (95%) completed the study's pretest (103 Intervention, 106 Control), and were then randomly assigned to the intervention condition or to a wait-list control condition. Of these, 175 (84% of the families providing pretest data) completed the posttest (85 Intervention and 90 Control group families). This study only included a pretest-posttest comparison. The posttest was completed 2-9 weeks after the intervention.

Sample: Among the families completing the posttest (Spoth et al., 1995), virtually all participants were Caucasian, and in slightly over half (54%) of the participating families, the target child was a boy. Among these families, 91% of the parents were married, 6% were divorced or separated, and 3% were unmarried but living in marital-like relationships. The average number of children was 3.2, and 65% of households had two or three children. Most mothers (64%) and fathers (54%) reported completing some post-high-school training; the median number of years of formal education was 13 for both mothers and fathers. The median annual family per capita income was $6,800; the median age of parents was 39 for mothers and 41 for fathers.

Measures: Pretest and posttest data collection procedures began with an in-home assessment consisting of a household composition interview, the administration of written questionnaires, and videotaped family interaction tasks. Following the in-home interview, families were given a mail-in questionnaire. The post-test in-home interview was conducted the same way as in the pretest, but in lieu of a mail-in questionnaire, a follow-up telephone interview was conducted. Each family was compensated at the rate of approximately $10 per hour for time devoted to the assessments. The videotaped portion of the assessment consisted of two structured interaction tasks. Interviewers asked the families to discuss a series of questions, first dealing with questions concerning family life, then concerning specific problems or issues cited by the families as sources of disagreement during an earlier portion of the interview. Families were given time allotments in order to complete their tasks, unsupervised by the interviewer.

Measures from Kosterman et al., 1997:

Proactive Communication - degree to which the parent is responsive in listening to the child and attempts to reason inductively with the child.

Negative Interaction - degree to which the parent escalates his or her own negative behaviors toward the child, expresses anger in a coercive manner, attempts to interrogate the child in an insistent, systematic manner, and behaves in an antisocial or antagonistic manner toward the child.

Relationship Quality - extent to which the parent and child display a high quality, emotionally satisfying relationship, and the parent behaves in a prosocial or considerate manner toward the child and has a high quality and significant amount of involvement in the child's life.

Measures from Spoth, Redmond, Hockaday, & Yoo, 1996

Behavioral Tendency Toward Alcohol Abstinence - index of two indicators of behavioral tendency to abstain from alcohol use (behavioral intentions and attitudinal factors influencing use), based on six questionnaire items.

Mastery Self-Esteem - Seven-item scale assessing the extent to which the adolescent sees himself or herself as having mastery over the forces that importantly affect his or her life; e.g., "I have little control over the things that happen to me."

Affectional Relationship with Parents - nine items regarding adolescent's mother and nine parallel items regarding the father; e.g., "How often did your mom/dad act lovingly and friendly toward you?"

Affiliation with Prosocial Peers - five items assessing the degree to which various statements describe their closest friends in community, school, and home contexts; e.g., "These friends sometimes break the law."

Analysis: Both self-report and observational measures were used in the analyses. Analyses were conducted in four steps. First, ANCOVAs controlling for pretest scores were conducted to test for intervention effects on intervention-specific and general child management outcomes for both mothers and fathers. Second, regression analyses were conducted to evaluate the relationships among the direct effects of the individual difference and intervention attendance variables on the intervention-specific parenting outcome measure, as well as the direct effect of this outcome measure on the general child management outcome. Third, analyses were conducted to rule out the possibility of important direct effects of the individual difference and intervention attendance variables on the general child management outcome. Finally, effects of the interactions between individual difference variables and measures of intervention attendance on the intervention-specific parenting outcome were examined.

Outcomes

(Spoth et al., 1995)

Baseline Equivalency and Differential Attrition: There were no significant differences in parents' age, fathers' education, or family income between the intervention and control groups. At pretest and posttest, control group families had more children than intervention families (3.4 vs. 2.9 at pretest and 3.3 vs. 3.0 at posttest). At posttest, intervention mothers also reported significantly higher educational attainment (14.0 vs. 13.4 years); the differences were not significant at pretest.

Outcomes:

Spoth et al., 1995. Results of the ANCOVA analysis indicated that there were significant intervention group versus control group differences on the intervention-specific and general child management skills outcome measures for both mothers and fathers, when controlling for pretest levels of these variables. Evidence from a regression path model also indicated effects of intervention attendance and readiness for parenting change in the intervention-specific parenting skills outcome among mothers and fathers. Two other outcomes not hypothesized in the original model were also found. In the case of fathers, a significant effect of the interaction between pretest level of intervention-specific parenting skills and intervention attendance was found on the intervention-specific parenting skills outcome. In the case of mothers, there was evidence of a direct effect of intervention attendance on the general child management skills outcome.

Protective factors and alcohol refusal skills: (Spoth, Yoo, Kahn, & Redmond, 1996)

This study describes a model which tests the additive effects of strong family attachments and peers' prosocial norms on early adolescent alcohol refusal skills. Both child attachment with parents and with peers' prosocial norms were hypothesized to be predictive of the alcohol refusal outcome. Additionally, child attachment with parents was expected to be significantly correlated with peers' prosocial norms. Finally, exposure to a parenting skills training intervention was hypothesized to indirectly influence peer refusal skills via its effect on child attachment with parents.

The results of one-tailed t -tests indicated that post-test child attachment with parents, pretest child alcohol refusal, and posttest level of peers' prosocial norms each exhibited highly significant direct positive effects on posttest child alcohol refusal skills in both the mother-child and father-child models. Additionally, parent skills training exposure, pretest child attachment with parent, and pretest level of peers' prosocial norms exhibited indirect positive effects on post-test child alcohol refusal skills in both models. Posttest child attachment with parents was directly and positively affected by parent skills training exposure in both cases.

Protective factors and alcohol abstinence: (Spoth, Redmond, Hockaday, & Yoo, 1996)

This study assesses protective process models that focus on specific adolescent problem behaviors and attitudes and also include intervention attendance and implementation effects. The hypothesized model suggests a positive relationship between Time 1 affectional relationship with parents and Time 2 mastery-esteem. Affectional relationship with parents at Time 1 was hypothesized to exert a positive influence on both affiliation with prosocial peers and mastery-esteem at Time 2. Affiliation with prosocial peers at Time 1 was expected to positively influence affectional relationship with parents and mastery-esteem at Time 2. It was further hypothesized that mastery-esteem, prosocial peer affiliation, and affectional relationship with parents would have positive contemporaneous effects on adolescents' tendency toward alcohol abstinence at Time 2.

There was no support for the cross-time effects of affectional relationship with parents on prosocial peer affiliation or affiliation with prosocial peers on affectional relationship with parents. There was support for the cross-time effects of affectional relationship with parents and affiliation with prosocial peers on young adolescent mastery-esteem.Thus, affectional relationship with parents and affiliation with prosocial peers exhibit both an immediate direct effect on behavioral tendency toward alcohol abstinence (at Time 2) and a longer-term indirect effect from Time 1 via Time 2 mastery-esteem on tendency toward alcohol abstinence. With regard to the intervention, GGC had a significant positive effect on affectional relationship with parents which then positively impacted the behavioral tendency toward alcohol abstinence.

Family interactions: (Kosterman, Hawkins, Spoth, Haggerty, & Zhu, 1997)

Results were extracted from actual family interactions collected in videotaped observations at the families' homes both prior to and following the intervention. Only parents participating in both the pretest and post test interactions were included in the analyses (n = 174 mothers, and n = 157 fathers). Analysis of covariance (ANCOVA) was employed in order to examine differences in post test scores while controlling for any differences in pretest scores. The GGC intervention was effective in promoting proactive communication from parent to child and in improving the quality of parent-child relationships (significant for fathers on 1 of 2 tasks and marginally significant for mothers on 1 of 2 tasks). GGC also reduced mothers', but not fathers', negative interactions with their children.

Targeted parenting behaviors: (Kosterman, Hawkins, Haggerty, Spoth, & Redmond, 2001)

This analysis focused on the session-specific effects of an implementation of the GGC curriculum. A controlled test of the effects of each session of the program on the specific parenting behaviors targeted in that session was conducted. MANCOVA and ANCOVA were employed in order to examine differences in post test scores while controlling for differences in pretest scores. MANCOVA results were significant for mothers but not fathers. Each of the outcome measures were examined independently through ANCOVAs. All results were in the expected direction, with the intervention group demonstrating more improvement than the control group. Seven of 14 outcomes were statistically significant for mothers, with another four approaching significance (p < .10). For fathers, 2 of 14 were significant, with one additional approaching significance. For mothers, the seven significant outcomes were: rewards for prosocial behavior, communicating rules around substance use, punishing for misbehavior, restricting child's alcohol use, expecting child to refuse beer from friend, expressing less conflict toward spouse, and working at being more involved with child. For fathers, they scored higher than controls on communicating rules around substance use and involvement with child.

Study 2

Summary

Spoth et al. (1999) and 12 additional reports utilized a cluster randomized controlled trial to examine 667 families with children in 33 schools. The schools were randomly assigned to one of three groups: the GGC intervention, another parenting intervention, or a minimal contact control group. Assessments occurred over seven waves: baseline, approximately 6, 18, 30, 48, and 72 months after baseline, and finally when the target children were age 21. Measures parenting behaviors and youth mental health, substance use, and risky sexual behavior.

Spoth et al. (1999) and additional reports found, compared to the control group, youth in the intervention group had:

• Greater likelihood of remaining in the no-use group through the 2-year follow-up for those youth who had not initiated substance use at the 1-year follow-up
• Greater likelihood of remaining in the same substance use status through the 2-year follow-up for those youth who had initiated substance use at the 1-year follow-up
• Fewer transitions to substance use at the two-year follow-up
• Lower alcohol initiation scores, lower frequency of past month drinking (among users), and reduced growth of alcohol use at the 3.5-year follow-up (ages 12-15 ½)
• Slower rates of increase in alcohol use, polysubstance use (alcohol, tobacco, and marijuana) and general delinquency (e.g., theft, vandalism, violence) over time across five waves of data
• Slower rates of increase in depressive symptoms and tobacco use from 6^th through 12 grade

Program Effects on Risk and Protective Factors:

• Increased protective parenting behaviors, parent-child affective quality, and general child management skills
• Strengthened parental norms against alcohol and other drug use by adolescents at the 3.5-year follow-up

Evaluation Methodology

Design:

Recruitment/Sample Size: Participants in the study were families of sixth graders enrolled in 33 rural schools in 19 contiguous counties in a Midwestern state. Schools were selected for participation if 15% or more families in the school district were eligible for free or reduced-cost school lunches and community size was 8,500 or fewer. All families of sixth graders in participating schools (1,309) were recruited for participation. Of these families, 667 (51%) enrolled in the project and completed pre-testing in the fall of 1993. These families were similar to the eligible families on sociodemographic characteristics.

Study type/Randomization/Intervention: The schools were randomly assigned to one of three experimental groups: a minimal contact control condition, or to one of the following universal family-focused preventive intervention programs: the Preparing for the Drug Free Years program (PDFY, now called Guiding Good Choices) or Iowa Strengthening Families Program (ISFP, now known as Strengthening Families 10-14). A randomized block design was used, in which schools were blocked on the proportion of students who resided in lower income households and on school size. Each group included 11 schools, and pretested families included 221 in the PDFY group, 238 in the ISFP group, and 208 in the control group. Among the intervention groups, 124 families assigned to the PDFY group participated in the program and 117 families assigned to the ISFP condition participated.

Assessment/Attrition: Assessments included self-reports and in-home interviews. After completion of the 5-week intervention, families were assessed approximately 6, 18, 30, 48, and 72 months following the pretest (when the students were in the sixth, seventh, eighth, tenth, and twelfth grades, respectively). An additional follow-up was conducted with the target children when they had entered young adulthood, at the approximate age of 21. These assessments constitute seven waves. From wave 2 (posttest) to wave 7 (young adult follow-up), overall attrition rates were 17.4%, 29.2%, 34.3%, 33%, 31.5%, and 27.4%.

Sample Characteristics:

The average number of children per family was 3.0, and in just over half of the families (51.2%), the target child was a girl. Representative of the study region, 307 (85%) families had two parents. Most parents had completed high school (97% of mothers, 95% of fathers), while 201 (57%) mothers and 168 (54%) fathers reported some post-high school education. Average age was 37.0 years for mothers and 39.6 years for fathers; nearly all families (98.6%) were Caucasian. Of the 176 post-tested intervention group families in this study, 199 (68%) had attended one or more PDFY sessions.

Measures:

Measures used in this study were drawn from both the self-report and observational portions of the in-home interviews. Pretest and posttest data collection procedures began with an in-home assessment, consisting of a household composition interview, the administration of written questionnaires, and videotaped family interaction tasks. Following the in-home interview, families were given a mail-in questionnaire. The posttest in-home interview was conducted the same way as in the pretest, but in lieu of a mail-in questionnaire, a follow-up telephone interview was incorporated. Each family was compensated at the rate of approximately $10 per hour for time devoted to the assessments. The videotaped portion of the assessment consisted of two structured interaction tasks. Interviewers asked the families to discuss a series of questions, first dealing with questions concerning family life, then concerning specific problems or issues cited by the families as sources of disagreement during an earlier portion of the interview.

Measures of substance use included child self-reports in response to questions pertaining to the following: lifetime alcohol use; lifetime alcohol use without parental permission; lifetime drunkenness; alcohol consumption in the past month; lifetime cigarette use; past month use of cigarettes; daily use of cigarettes; lifetime use of chewing tobacco; past month use of chewing tobacco; and lifetime marijuana use. The alcohol and tobacco items were combined in various ways to produce composite indices of alcohol and tobacco use. As reported by Spoth et al. (2004), the average alphas across the first six waves of data collection were .79 and .70 for alcohol and tobacco use indices, respectively.

Parent-reports of social emotional maladjustment were collected at wave 1 (pretest). Three subscales (for depression, anxiety, and hostility) of the Symptom Checklist 90-Revised were used to assess parental internalization and externalization. A total maladjustment score was calculated by summing the three subscales. In the case of dual-parent families, the two parent's mean scores were averaged. Internal consistency for the three subscales, as assessed with Cronbach's alpha coefficients, ranged from .75 to .89.

The measure of substance use included: lifetime cigarette use ("Have you ever smoked cigarettes?"), lifetime alcohol use ("Have you ever drunk beer, wine, wine coolers, whiskey, gin, or other liquor?"), and advanced substance use included several items because of low base rates (e.g., past month use of cigarettes and alcohol, lifetime drunkenness, lifetime use of illicit substances). The collapsed model had five statuses (no use, alcohol experimentation, tobacco experimentation, experimentation with both alcohol and tobacco, and more advanced use).

Depression was measured using eight items from the Child Behavior Checklist - Youth Self Report. Adolescents described feelings and behaviors on a three-point scale for the six-month period preceding the assessment, such as "I feel worthless or inferior," and "I think about killing myself." Polysubstance use was a self-report measure of the number of times the adolescent smoked cigarettes, chewed smokeless tobacco, and consumed alcohol within the past month. Three additional items asked about the number of times adolescents had used marijuana, inhalants, and other illicit drugs within the past year. Each item was coded into a dichotomous variable and then summed to compute an index that could range from 0 to 6.

A measure of adolescent methamphetamine use was added in the six-year follow-up and assessed with a single item: "For each substance listed below, please write down the number of times you used it during the past 12 months...took methamphetamines (meth)." The question was then dichotomized and recoded as 0 for no reported methamphetamine use and 1 for any reported methamphetamine use.

Wave 7 (young adulthood):

Measures of substance use frequency included

• drunkenness frequency (i.e., "How often do you usually get drunk?");
• alcohol-related problems (using a modified form of the Rutgers Alcohol Problems Index);
• past year cigarette frequency (i.e., "During the past 12 months how often did you smoke cigarettes?");
• past year illicit drug frequency (using nine open-ended items asking "How many times in the past 12 months did you use [specific substance]?");
• a polysubstance use index created using the three measures of substance use, dichotomizing each variable to indicate use (1) or no use (0) of substances, and summing the three dichotomous items to form an index with values ranging from 0 (indicating no use of any substance) to 3 (indicating at least some occurrence of all three substance use behaviors).

Measures of health-risking sexual behavior included

• self-reports of the number of sexual partners in the past year;
• condom use in the past year (with five responses ranging from "none of the time" to "always" and including a "don't know" option);
• substance use and sex to include two questions asking "When you have sex, how often have you been drinking alcohol" or "using drugs other than alcohol?" (including the same response options as for condom use);
• sexually transmitted diseases with participants asked "Have you ever been diagnosed with a sexually transmitted disease (STD or VD) other than HIV/AIDS, such as gonorrhea, genital warts, chlamydia, trich, herpes, or syphilis?"

Adolescent Alcohol, Tobacco, and Illicit Drug Initiation Index:

A substance use initiation index was calculated using data from waves 1 to 6. This index is the sum of the five individual substance initiation measures, each scored so that "Yes" = 1 and "No" = 0 (with measures corrected for consistency so that when "Yes" was reported for initiation of any substance at any wave, responses were scored "Yes" for each subsequent wave). Scores ranged from 0, indicating no initiation, to 5, indicating the initiation of alcohol use (without parental permission), drunkenness, tobacco, marijuana, and other illicit drugs. Internal consistency for this index, as assessed by Cronbach's alpha, averaged .60 across waves.

Analysis:

The types of analyses vary by report.

SEM, based on the sample covariance matrix, was conducted with LISREL 7 using maximum likelihood estimation. Overall model fit was assessed using four fit indexes: the likelihood ratio chi-square, goodness-of-fit index (GFI) and adjusted goodness-of-fit index (AGFI), and Critical N (CN). In addition to the hypothesized post-test direct and indirect intervention effects of primary interest to this investigation, a number of other parameters relevant to the measurement model or structural model fit were included in the analyses. These included (a) pretest effects paralleling those at post-test, (b) pretest to post-test effects of each parenting outcome, (c) correlated residuals of the global parenting constructs within each wave of data, (d) measurement method effects associated with observational and self-report indicators, (e) correlated pretest-post-test errors for each of the indicator variables, and (f) correlated pretest-post-test residuals of the latent method effect constructs. ANOVAS were conducted for each latent-variable parenting construct indicator scale to test for group inequivalence and differential attrition. T-tests and chi-square analyses were also conducted to examine group inequivalence with respect to a range of family characteristics. Finally, three sets of analyses were conducted to assess school effects.

Wave 4: Two-year Follow-up (Eighth Grade)

Spoth, Reyes, Redmond, & Shin (1999): The non-intent to treat analyses for this report included only those control and intervention families that actually attended the intervention (n = 101 PDFY adolescents, n = 91 ISFP adolescents, n = 137 control group adolescents). Loglinear analyses tested for intervention versus control group differences in (a) proportions of adolescents beginning and ending each of the two transition periods in the no-use status, and (b) proportions of adolescents having initiated substance use who did not progress in use.

Wave 5: Four-year Follow-up (Tenth Grade)

Spoth, Redmond, & Shin (2001): Multilevel (mixed model) analyses of covariance were used to test for intervention effects on alcohol and tobacco initiation indices and frequency of past month or past year use of alcohol, tobacco, and marijuana.

Mason, Kosterman, Hawkins, Haggerty, & Spoth (2003): Latent growth curve modeling was used to analyze the five waves of data to determine the growth in adolescent substance use and delinquency.

Guyll, Spoth, Chao, Wickrama, & Russell (2004): Analyzing alcohol and tobacco composite use indices, a three-level growth model, with individual observations clustered within adolescents and adolescents clustered within schools, was used to describe individual trajectories of substance use from waves 2 through 5. Defining the final level of substance use (that reported at wave 5) as the intercept, group effects at the wave 5 or four-year follow-up were tested by the coefficient for the condition dummy variables, and group-by-time interactions coefficients were used to test for growth effects.

A risk moderation analysis, analyzing the influence of parental social emotional maladjustment (reported at baseline) on program effects was analyzed by including in the models a risk-by-group variable (specifying moderating effects at the wave 5 or four-year follow-up) and a risk-by-group-by-time variable (specifying moderating effects on growth rates).

The results reported are based on an analysis of only the 373 families who provided complete data at all five data collection points; however, the researchers report that an intent-to-treat analysis using multiple imputation yielded the same results regarding statistical significance.

Wave 6: Six-year Follow-up (Twelfth Grade):

Mason, Kosterman, Hawkins, & Haggerty (2007): The effects of PDFY on trajectories of self-reported adolescent depressive symptoms from 6th through 12th grade were examined. The data were analyzed using intent to treat latent growth curve modeling, controlling for maternal depression, parent education gender of the target child, and the pretest measure of adolescent depressive symptoms. The individual-level growth curve analysis included an adjustment for clustering within schools.

Spoth, Clair, Shin, & Redmond (2006): In an intent-to-treat analysis, the Fisher exact test was used to assess group differences in methamphetamine use at wave 6.

Spoth, Redmond, Shin, & Azevedo (2004): Nonlinear growth curve models, incorporating school as a random effect, were conducted with school-level outcome variables aggregated over the available respondents in each school. Additional intervention effect tests were derived from model results and included analyses of intervention-control differences in the length of time (from baseline) to the growth curve inflection point (the point at which the growth rate is at its maximum). This analysis only included data from respondents who provided data at all six waves of the study and excluded schools in which data were available from five or fewer students from the analysis.

Wave 7 (Young Adulthood):

Spoth et al. (2009, 2014): The long-term, young adult follow-up controlled for baseline outcomes with growth models, used all subjects with available data, and adjusted for clustering within schools.

Following the use of hierarchical latent growth curve models to assess intervention effects on adolescent initiation of alcohol, tobacco, and/or drug use, the models examine adolescent initiation growth factor effects on 1) subsequent young adult substance use outcomes (Spoth et al., 2009), and 2) health-risking sexual behaviors and STDs (Spoth et al., 2014). Growth in initiation across time was modelled as linear, controlling for pretest effects on adolescent initiation and young adult substance use outcomes as well as gender, parent marital status, parent education level, and family income reported at pretest and accounting for school clustering by including school as a higher-level cluster variable. The influence of assignment to the intervention condition was incorporated via direct effects on both the intercept and slope factors of adolescent initiation, with resulting indirect effects on young adult outcomes through those growth factors to test for indirect intervention effects on young adult outcomes. The models use an intent-to-treat analysis with full-information maximum likelihood estimates for incomplete data.

Outcomes:

Implementation fidelity: Approximately 56% of pretested families attended at least one session with 94% of these families represented by a family member in three or more sessions, 93% attending four or five sessions, and 62% attending all five. Results from fidelity observations showed that all teams covered the key program concepts and the teams covered, on average, 69% of the component tasks in the group leader's manual (Spoth, Clair, & Trudeau, 2014).

Baseline Equivalence: Intervention and control group participants were equivalent on family sociodemographic characteristics and school and community characteristics at baseline. Regarding substance use, Guyll et al. (2004) reported that PDFY adolescents reported greater tobacco use, and Mason et al. (2007) reported that the polysubstance use index at pretest was higher among PDFY versus control participants.

Differential Attrition: At wave 5, Guyll et al. (2004) reported that families failing to complete all five assessments had younger parents, parents with less education, and children who reported more alcohol use at pretest when compared to those who completed measures at all assessment points. Mason et al. (2007) reported that assessment of dropouts and completers revealed a few differences at wave 6 with more highly educated parents remaining in the study, and adolescent alcohol drinkers being more likely to leave the study than nondrinkers. At wave 7, Spoth et al. (2009) stated that there were no significant differences in attrition across conditions.

Waves 2 and 3: Posttest and One Year Follow-up (Sixth and Seventh Grades)

Direct and Indirect Effects of Parenting Outcomes (Spoth, Redmond, & Shin, 1998)

PDFY: Three parenting outcomes were measured: the proximal measure of intervention-targeted parenting behaviors and two distal measures for general child management and parent-child affective quality. Results showed support for the hypothesized direct intervention effect on increase in protective parenting behaviors, with an indirect effect on parent-child affective quality and general child management skills. All hypothesized structural effects were significant at the .01 level.

One Year Follow-up (Redmond, Spoth, Shin, & Lepper, 1999)

The results of the above parenting model was tested with one-year follow-up data. Results indicated that statistically significant effects on parenting outcomes were sustained through a one-year period following post-test. The intervention effects were examined via structural equation modeling. These models included direct intervention effects on intervention-targeted parenting behaviors. Intervention effects on the two global parenting constructs were modeled as indirect, via the intervention-targeted parenting construct. Modeling results supported the hypothesized direct and indirect intervention effects for the PDFY program. When controlling for pretest levels and measurement method effects, all hypothesized effect paths at one-year follow-up were significant at the .01 level. The current findings provide further evidence that the tested interventions set in motion a sequence of changes in parenting practices that are sustained over time and are, thereby, more likely to yield long-term reductions in adolescent problem behaviors.

Waves 3 and 4: One- and Two-year Follow-ups (Seventh and Eighth Grades)

2.5 years following baseline (2 years post-intervention) (Spoth, Reyes, Redmond, & Shin, 1999)

Loglinear analysis showed that the two interventions were equally effective compared to the control group and did not differ significantly from one another. At the two-year follow-up, when youth were in the second semester of 8th grade, there was a significant experimental group x outcome interaction effect, and Z tests confirmed that three of four tests were statistically significant (both tests were significant for the PDFY intervention). This indicates that (a) youth in the PDFY group who had not initiated substance use at the time of the one-year follow-up were more likely to remain in the no-use group through the two-year follow-up than control group adolescents (88% probability of remaining a nonuser vs. 70%), and (b) youth in the PDFY group who had initiated substance use at the one-year follow-up were more likely to have remained in their one-year follow-up substance use status through the two-year follow-up than control group youth (77% probability of not advancing to more frequent/varied use vs. 50%). Thus, while substance use rates increased among all groups over the course of the study, transitions to substance use at the two-year follow-up were significantly lower among intervention group adolescents.

Wave 5: Four-year Follow-up (Tenth Grade)

Substance Use Outcomes (Spoth, Redmond, & Shin, 2001)

In this follow-up (four years after pretest), the PDFY group showed significantly lower alcohol initiation scores than the control group. New user proportions were lower among PDFY adolescents than among controls (approached significance p < .10) for two of five behaviors, lifetime drunkenness and lifetime use of marijuana. Among those adolescents who had used alcohol and tobacco during the past month and marijuana during the past year, analyses showed that PDFY adolescents had a lower frequency of past month drinking than the control group (relative reduction 40.6%). Examinations of mean substance frequencies to examine trends showed more positive outcomes in the 8th and 10th grades. The only significant effect in this study was for alcohol initiation. However, PDFY students showed a number of positive trends in lower rates of tobacco and marijuana use, compared to the control group. For example, PDFY adolescents were 37% less likely to have initiated marijuana use over the course of the study than control youth (p < .10). This program was compared not only to a non-treatment control group, but to another treatment program, Iowa Strengthening Families (ISF). ISF had more consistent and stronger outcomes than the PDFY program.

Park, Kosterman, Hawkins, Haggerty, Duncan, Duncan, & Spoth (2000)

These results, also at 3.5 years, showed a reduction in the growth of adolescent alcohol use from ages 12 - 15 1/2. Additionally, the PDFY adolescents significantly strengthened parental norms against alcohol and other drug use by adolescents over time. There was a marginally significant trend with the intervention helping to maintain proactive family management practices over time. There were no significant PDFY effects on growth in family conflict or refusal skills over time.

Mason, Kosterman, Hawkins, Haggerty, & Spoth (2003)

Latent growth curve modeling was used to analyze the five waves of data to determine the growth in adolescent substance use and delinquency. Compared to the control group, PDFY assignment was significantly associated with a slower rate of increase in polysubstance use (alcohol, tobacco, and marijuana) and general delinquency (e.g., theft, vandalism, violence) over time, controlling for pretest level of polysubstance use and delinquency, age, and gender.

Guyll, Spoth, Chao, Wickrama, & Russel (2004)

There was no significant difference between PDFY intervention group participants and controls in alcohol use at the four-year follow-up, but PDFY participants had significantly lower rates of increasing alcohol use from wave 2 to wave 5 when compared to controls. No significant intervention effects for tobacco use were found.

Parental social emotional maladjustment did not have a moderating effect on alcohol or tobacco use at five-year follow-up or on use trajectories for these substances from wave 2 to wave 5.

Wave 6: Six-year Follow-up (Twelfth Grade)

Depressive Symptoms (Mason, Kosterman, Hawkins, & Haggerty, 2007)

PDFY, relative to the control group, significantly reduced the rate of increase in self-reported depressive symptoms at the 12th grade follow-up (Cohen's d = .21). PDFY also slowed the rate of increase in polysubstance use. The model also examined reduced substance use as a mediator of the link between the intervention and adolescent depressive symptoms. Polysubstance use was linked with reductions in depressive symptoms, although the indirect effect of the intervention on depressive symptoms through polysubstance use was only marginally significant.

Methamphetamine Use (Spoth, Clair, Shin, & Redmond, 2006)

There was no significant difference between the PDFY intervention and control groups in self-reports of using methamphetamines in the past 12 months.

Substance Use Growth Rates (Spoth, Redmond, Shin, & Azevedo, 2004)

Significant differences in growth rates between the PDFY and control groups were observed for the tobacco-related outcomes only (the tobacco composite use index and lifetime cigarette use) with participants in the PDFY group showing slower overall growth in tobacco use relative to controls. There were no differences between the PDFY and control groups in inflection point.

Wave 7: Long-term Follow-up in Young Adulthood (approximately age 21)

Substance Use Frequency (Spoth, Trudeau, Guyll, Shin, & Redmond, 2009)

The PDFY condition had generally significant effects on the slope for the adolescent substance initiation index, indicating slower growth among intervention subjects. The direct effects models found significant PDFY program effects on cigarette frequency and alcohol-related problems.

Significant indirect effects of the PDFY intervention were observed for drunkenness frequency, illicit drug frequency, and on the polysubstance use index with lower reported values among the intervention group when compared to the minimal contact control group. Marginally significant indirect effects (p<.10) were observed for alcohol-related problems and cigarette frequency.

The models including both direct and indirect effects found no improvements in model fit (when compared to the indirect effects model) for any of the outcome variables, or a significant direct intervention effect. The analysis of relative reduction rates found significant or marginally significant indirect effects (p-values are not reported, but rather noted to be p<.05, one-tailed) for the PDFY intervention on all dichotomized young adult outcome measures: drunkenness frequency, alcohol-related problems, cigarette frequency, illicit drug frequency, and on the polysubstance use index.

Health-risking Sexual Behaviors and Sexually Transmitted Diseases (Spoth, Clair, & Trudeau, 2014)

For all three health-risking sexual behaviors and the measure of lifetime STDs, the PDFY condition either significantly or marginally significantly predicted the intercept and slope for the adolescent substance initiation index (p-values are not reported, but rather noted to be p<.05, one-tailed). The direct effects models found a marginally significant PDFY program effect on lifetime STDs (p=.046), but not on any of the health-risking sexual behaviors.

Significant or marginally significant (reported to be p<.05, one-tailed) indirect effects of the PDFY intervention were observed for lifetime sexually transmitted diseases, past year number of partners, and substance use during sex (but not condom use) with lower reported values among the intervention group when compared to the minimal contact control group. The models including both direct and indirect effects found no improvements in model fit (when compared to the indirect effects model) for any of the outcome variables.

Because substance use during sex and failure to use condoms is assumed by the researchers to be less likely in the context of a married or cohabitating relationship (when compared to single individuals), post hoc two-group analyses were run to test whether intervention effects for those two variables differed by relationship status. This analysis found no differences between single and married or cohabitating individuals in the PDFY condition.

An analysis of relative reduction rates found significant or marginally significant indirect effects (reported to be p<.05, one-tailed) for the PDFY intervention on lifetime STDs and substance use during sex.