Parenting in 2 Worlds: Testing Improved Parent-Adolescent Communication about Sexuality in Urban American Indian Families

The rapidly growing majority of American Indians (AIs) (70%) who now live in cities contend with severe health disparities while being underserved by culturally appropriate prevention programs (Anastario et al., 2013; Gone & Trimble, 2012). AI adolescents and adults experience relatively high rates of inter-partner violence, sexual assault, and co-occurring risky sexual behavior and substance use (Malcoe et al., 2004; Ravello et al., 2014; Tuitt et al., 2023). Thus a substantial need exists for interventions that promote the sexual health of AI youth (Suellentrop & Hunter, 2009; Yuan et al., 2015). Culturally tailored interventions that address sexual health disparities among AI adolescents have been effective in reducing risk behaviors (Craig Rushing et al., 2018; Kenyon et al., 2019; MacIntosh, 2012). However, research on parent-adolescent interventions that specifically address communication with adolescents about risky sexual behaviors is not as prevalent despite demonstrated efficacy in preventing adolescent pregnancies and sexually transmitted infections in the general population (Lowe, 2008; Rouner et al., 2015; Widman et al., 2016).

Sexual health risk factors among American Indian adolescents

The history of colonization and the continued prevalence of structural racism and inequality contribute greatly to the many negative statistics that characterize the health of Indigenous peoples world-wide (Monchalin et al., 2019). American Indian/Alaska Native youth, relative to non-Hispanic White youth, experience higher rates of sexually transmitted infections (STIs), being forced to have unwanted sex, becoming pregnant or impregnating someone, and having their first sexual intercourse before age thirteen (Centers for Disease Control [CDC], 2018; Urban Indian Health Institute, 2009; Vernon, 2001). Consequences of early and risky sexual behavior include poorer mental health outcomes, such as depression (Spriggs & Halpern, 2008; Vasilenko et al., 2012), which is more likely to occur in girls following early sexual onset (Firestone et al., 2006). Sexual health risks for AI youth often co-occur with substance use and exposure to violence (Hellerstedt et al., 2006; Mitchell et al., 2007).

On a more positive note, avoidance of risky situations and intentions to not engage in sexual behavior are associated with delayed sexual initiation in AI adolescents (Markham et al., 2015). Thus, helping parents or guardians learn how to discuss sex with their adolescent is potentially a powerful tool that may delay the debut of consensual sexual activity and decrease the associated risks of adverse sexual health outcomes.

Parent-child communication as a protective factor for adolescent sexual health

Parent-child relationships are powerful agents of socialization through which children acquire knowledge and emotion skills that can improve the quality of their relationships as they mature (Bush & Peterson, 2013; Grusec & Davidov, 2010). While research has often traced the influence of parental socialization in dyadic parent-child relationships (Bush & Peterson, 2013), research is increasingly focused on understanding how the combination of maternal and paternal communication influences the child (Axpe et al., 2019). Parent-child communication about sexual health is part of the parental socialization process, and a potentially protective factor that can encourage safer sex practices and prevent adolescent pregnancies and STIs. In the general population, communication with a parent regarding sexual health is consistently associated with safer sex behavior among adolescents (Guilamo-Ramos et al., 2016; Harris, 2016; Widman et al., 2016; Stewart et al., 2019). These conversations typically contain four components: information about condoms and contraceptives; consequences of sexual behavior and the advantages of safe sex; effectiveness of condoms and contraceptives for preventing pregnancy and STIs; and where to obtain more information about condoms and contraceptives (McCallister et al., 2019).

There is evidence in the general population that the gender composition of dyadic parent-adolescent communication regarding sex and sexual health, and its timing, can affect the frequency, focus and influence of these conversations. These parent-adolescent conversations are generally more strongly predictive of safer adolescent sexual behaviors when involving mothers rather than fathers (Dilorio et al., 1999), daughters rather than sons (Widman et al., 2016), and younger rather than older children (Beckett et al., 2010). Dyads of fathers and daughters communicate the least on these topics, and focus less on sex-positive topics than on sexual risk and exploitation topics (Evans et al., 2020). Fathers generally are less likely than mothers to have characteristics associated with successful communication with their child regarding sex and sexual health, such as a sense of self-efficacy and expectations of positive outcomes (Wilson & Koo, 2010). Other factors that can affect dyadic father-son communication are the son’s pubertal development, the father’s sexuality values, education, and communication with his own father (Lehr et al., 2005). Overall, the observed variations in parent-adolescent communication about sexual health have been attributed to gender differences in socialization about dating and sexual relationships, such as greater concern about sexual exploitation of daughters than of sons (Perilloux et al., 2008), and the more complex and intense social pressures to engage in sex faced by older, more sexually developed adolescents (Beckett et al., 2010).

Specific to AI families, parent-adolescent communication is a demonstrated protective factor in sexual health, delaying the onset of an adolescent’s sexual debut, reducing the odds of recent sexual intercourse, and increasing condom use during sex (Chewning et al., 2001; Greene et al. 2018; Hellerstedt et al., 2006 ) and providing greater understanding of sexual health topics, such as HIV/AIDS (Ramirez et al., 2002). Parents are key positive adult role models for their adolescents, acting as a protective factor against risky sexual behavior and early debut of sexual initiation. Intervention programs focused on parent-adolescent communication regarding sexual health in AI communities have the potential to create significant change.

A Culturally Tailored Intervention: Parenting in 2 Worlds (P2W)

Although urban AI parents have been included among the participants in assessments of culturally informed parenting programs for AI populations, parenting interventions are rarely customized to urban AI families and their community contexts. In a multi-phase cultural adaptation process, P2W preserved the core components of an existing efficacious parenting program while modifying the content to address cultural and social challenges facing urban AI families and altering formats to accommodate distinctive AI learning styles (Kulis et al., 2016). Guided by ecodevelopmental theory, the overall goal of the cultural adaptation process was to integrally embed cultural elements that serve to protect and strengthen urban AI families and communities into the curriculum. Specific goals of the manualized parenting intervention were to: (a) empower parents to help their youth resist substance use and other risk behaviors; (b) build and strengthen family functioning to promote pro-social youth behaviors; and (c) increase the family’s problem solving and communication skills in ways that resonated culturally with urban AI families. The adaptation process, detailed elsewhere (Kulis et al., 2015), modified an existing effective parenting program, created a pilot adapted version for urban AI parents, tested the pilot version, and made further refinements. The adaptation utilized feedback from participating AI parents, AI workshop facilitators and observers, and tribally enrolled curriculum experts, with diverse tribal backgrounds and different family histories of migration to urban areas. Qualitative and quantitative data from curriculum lesson feedback forms and focus groups were analyzed by an AI research team to identify cultural, linguistic and learning style changes needed to make the P2W curriculum appropriate for urban AI families. The adaptation addressed the challenge of tribal diversity in cultural practices, traditional languages, indigenous beliefs and value systems by identifying and incorporating cultural elements shared by diverse urban AI communities. These values included spirituality, sacred history, sacred lands, notions of kinship expressed through clans or bands, the importance of ritual, respect, and traditional language and beliefs (Jumper-Reeves et al., 2014). P2W also addressed distinctive challenges for urban AI families, such as cultural disruptions accompanying tribal to urban migration, acculturation, enculturation and identity issues, and a variety of living arrangements. One-third of AI children are estimated to live in households without either parent (Shah et al., 2022). Recognizing the diversity of family arrangements in urban AI communities, P2W focuses on the role of those caring for children in the family, without assumptions about the nature of the family’s composition, their biological relationship to the child, or the gender identity, sexual identity or tribal background of the adult and the child.

The 10 P2W curriculum lessons are designed to last two hours and be delivered once per week. Topics include: an introduction to the program and to other participants; how to build supportive parenting communities; identifying family cultural traditions and values; adolescent development and how to know your teen’s world; communicating with your child; creating supportive relationships with teens; guiding their behavior effectively; practicing monitoring strategies; talking with teens about risky behaviors; and putting it all together. The lessons have strengths-based starting points that focus on AI cultural heritage as a foundation for raising healthy children and increasing family resilience and wellbeing. Parents discuss the values of their tribal backgrounds, how they were transmitted intergenerationally, and how cultural values can protect their families in the face of the cultural disruptions and the social challenges of life in the city. The curriculum emphasizes guiding children’s behavior through modeling, rather than intrusive managing or monitoring. One lesson is devoted solely to the role of parents in sex education with the goals of increasing the adolescent’s awareness of sexual risk behaviors and their consequences, and teaching parents to be prepared for sensitive conversations about sexual risk behaviors with their adolescents. The parents discuss how to provide advice to their teens, and where the parents and youth could find accurate information and resources within their social networks. The lesson provides helpful hints about effective strategies to consider. These include more general matters of sexuality and sexual development. For example, the lesson advises parents to start teaching about sex and body parts from an early age (e.g. body ownership, permission); answer sex-related questions truthfully; call body parts by their correct names; teach following your values, but leave the door open for conversation and questions. The parents also discuss: ways to prepare adolescents to make decisions about sexual behaviors; providing guidance about waiting to have sex until the adolescent is older; how the teen can set limits (e.g., avoid being alone in cars; go out in groups); effective communication strategies with the adolescent for resisting pressure to engage in risky behaviors; and to be aware of and prepared for safe sex practices.

P2W has been shown to be efficacious in improving parental agency and positive parenting practices, reducing discipline problems with the child, and enhancing the cultural engagement of the participating parents, including their sense of AI identity, AI spirituality, and positive bicultural orientation (Kulis et al., 2016, 2020). In this article we tested the effects of P2W in increasing parent-adolescent communication about sex on two dimensions: discussion of sexual development and safe sex, and discussion of how to make decisions about having sexual intercourse. In addition to a test of the overall effects of P2W on these outcomes, we explored whether the effects varied depending on the gender and ages of the participating parent and their child. We expected that P2W participants would report significantly larger increases in communication about sex with their children on both dimensions, compared to participants in Healthy Families in 2 Worlds (HF2W), a non-culturally grounded, information-only intervention that provided parents with family health information in 10 weekly lessons.

HF2W lessons were suggested by groups of urban AI parents and professionals as important to the community, including vision and eye care, oral health, emergency response systems, CPR, first aid, media influences on children’s health, and safe dating practices. In the safe dating lesson, HF2W parents were taught what a healthy relationship looks like, the warning signs of an unhealthy relationship, and teen dating violence. They discussed what advice to give to a girl about dating, what advice to give to a boy, and how much supervision is needed for girls and for boys. As with other family health topics in HF2W, participants brainstorm family and community barriers to health and resources and practices to promote better health.

Based on findings in the general population showing variations in parent-adolescent communication about sex by the gender of the parent and of the child, as well as by the age of the child, it might be expected that in this urban AI sample there would be more communication about sex between teens and their mothers than with their fathers, more communication with daughters than with sons, and more communication with older than with younger children. However, intervention effects might reflect a different pattern because effective prevention programs may have larger effects on the segments of the population least likely to engage in this communication. Thus, without specific directional hypotheses, we explored whether the effects of the P2W program on parent-adolescent communication varied depending on the gender composition of the dyad, the age of the child, and also the age of the parent.

Method

Participants

Data for the study were collected in a randomized controlled trial (RCT) in the three Arizona cities with the largest AI populations (see Figure 1, CONSORT diagram). The participants were 575 parents/guardians of AI children living in urban areas. To be eligible, participants had to be: (1) primary caregivers of an AI child 10–17 years old attending an urban school; (2) responsible for making health, educational and social decisions for the child; and (3) residing in the city with the youth. Recruitment focused on this caretaker role, without exclusions based on family or household composition: participants could be parents or guardians of an urban AI child, either legal or informal; in a single-parent, dual-parent, extended or foster family household; and with opposite- or same-sex partners. Both of the child’s parents/guardians (hereafter, parents) were encouraged and allowed to attend the intervention workshops, but they chose only one parent to serve as a responding participant in the trial. Participants with multiple children between the ages of 10 and 17 were asked to report on parent-adolescent relationships with a single “focal” child, the one closest in age to 14.

Figure 1. CONSORT Diagram for Parenting in 2 Worlds Efficacy Trial.

*Based on number allocated; **Based on number participated in pre-surveys

The three cities from which the participants were drawn—Phoenix, Tucson and Flagstaff—have urban AI communities that differ markedly in tribal composition and migration histories. Those connected to northern Arizona tribes (Navajo, Hopi) are predominant in Flagstaff, those from southern Arizona tribes (Tohono O’odham, Pasqua Yacqui) constitute a majority in Tucson, and Phoenix has a much more diverse AI population connected to tribes across Arizona and other states.

Table 1 presents the demographic characteristics of the participants. For the sample as a whole, the gender composition of parents was unbalanced, 77% female and 23% male, while the “focal” children they reported on were split evenly between females and males. Adult participants were diverse in marital status and age. They were almost evenly divided between those married currently or previously, those unmarried but living with a partner, and those never married and not cohabitating. They ranged in age from 18 to 71, with a majority (62%) between 30 and 45. Those at the extremes of this age range may have been older siblings, aunts/uncles, or grandparents who had assumed a caretaker role for the child (adults were not asked to report their legal relationship to the child). To be eligible for the study participants did not have to be the child’s legal guardian. Informal guardianships of this nature are not uncommon in AI families (Garwick et al., 2008; Rink et al., 2016; Shah et al., 2022). Thus, the study could represent a wide array of family compositions. Although most participants had completed high school or some post-secondary education, incomes were very low: half reported annual family incomes below $10,000. On average, their households had between four and five members, including about two dependent children. The “focal” child was just under 14 years old, on average. Most parents had extensive past and current connections to reservations. Over half spent most of their childhood on a reservation, three-fourths had lived some time on a reservation, and all except a small minority had relatives living currently on a reservation. The participants came from 37 different tribal backgrounds, most (90%) tracing their heritage to one of the 22 tribal nations in Arizona.

Table 1.

Demographic Profile of Participants at Pretest, by Intervention Condition

	Total		Parenting in 2 Worlds		Healthy Families in 2 Worlds
Parent and Child Gender:
Mother/Daughter	38.3%		40.7%		35.4%
Mother/Son	38.7%		39.5%		37.7%
Father/Daughter	11.6%		12.0%		11.2%
Father/Son	11.4%		7.8%		15.7%
Parent Marital Status
Married Currently	20.5%		19.6%		21.2%
Divorced, Separated, Widowed	15.4%		16.0%		14.7%
Unmarried, Living with Partner	28.7%		27.8%		29.8%
Unmarried, Not Cohabitating	35.6%		36.6%		34.4%
Parent Education
Less than High School Degree	25.6%		25.2%		25.9%
High School Degree	29.8%		31.5%		27.8%
Education Beyond High School	44.6%		43.3%		46.3%
Annual Family Income
< $10,000	50.5%		49.0%		52.2%
$10,000 - $29,999	30.7%		30.5%		30.9%
$30,000 or More	18.8%		20.5%		16.9%
Most of Childhood on Reservation	49.8%		57.4%		49.8%
Relatives Living on Reservation	92.2%		91.5%		93.1%
	M	SD	M	SD	M	SD
Parent Age	36.72	(9.13)	37.32	(8.90)	36.01	(9.37)
Child Age	13.77	(2.37)	13.76	(2.30)	13.79	(2.45)
Household Size	4.65	(2.12)	4.57	(2.27)	4.74	(1.96)
Years Residing in Urban Area	16.70	(13.01)	16.67	(13.07)	16.73	(12.97)
Years Residing on Reservation	13.45	(11.76)	14.14	(11.71)	12.62	(11.78)
N	575		312		263

Measures

In addition to parenting skills, family functioning, and youth and parent risk behaviors, the questionnaires asked parents about their communication with their child about sex. This study analyzes two multi-item measures of the frequency of this type of communication.

Communication about Sex

Sexuality and Safe Sex

(Cox et al., 2008). As a general measure of discussion of sexual development and sexual risks, parents reported how often they: encouraged their child to ask questions about sex and sexuality; talked to their child about body changes in adolescence, HIV, sexually transmitted diseases, preventing pregnancy, and using condoms; and set rules for their child about sex. Responses to these seven items, scored from 1 (Never) to 4 (Often), formed a highly reliable mean scale (α = 0.94).

Sexual Decision-Making

(Kelley et al., 2003). Parents also reported discussing with their child decisions about whether and when to have sex. They indicated how often they talked about four issues in having sexual intercourse: negative consequences of becoming or getting someone pregnant; dangers of sexually transmitted diseases; moral issues in having or not having sexual intercourse; and when the child may have sexual intercourse. Responses to these items, scored from 1 (Never) to 4 (Often), formed another highly reliable mean scale (α = 0.93).

Demographics.

Gender.

Parents reported the gender (female or male) of themselves and of their focal child; the questionnaire did not supply non-binary gender options. We created dummy variables to contrast the four gender dyads: mother/daughter, mother/son, father/daughter, and father/son.

Age.

Parents reported own their age and the age of their focal child in whole years, at the pretest.

Procedure

Collaborating non-profit urban Indian centers recruited the participating parents, who were randomized individually to receive the P2W intervention or to receive the information-only HF2W curriculum that was not culturally tailored to urban AI families. Staff at the urban Indian centers recruited, screened, and enrolled eligible participants. Recruitment took place at schools, churches, youth centers, health fairs, pow-wows, other community events, and through word of mouth. Both of the child’s parents were encouraged and allowed to attend the intervention workshops, but they chose only one parent to serve as a responding participant in the trial. Participants with multiple children between the ages of 10 and 17 were asked to report on parent-adolescent relationships with a single “focal” child, the one closest in age to 14.

The P2W and HF2W interventions each contained 10 workshops delivered weekly by AI facilitators, drawn from the local community and trained by the research team. Both interventions included one workshop that addressed risk factors associated with sexual health with P2W delivering a culturally adapted curriculum and HF2W simply providing factual information about teen dating. The two interventions were implemented over two years in 26 workshop cycles (11 in Phoenix, 9 in Tucson, and 6 in Flagstaff). In each cycle, separate facilitators delivered the P2W and HF2W workshops, on the same timetables but in separate areas. Workshops had an average of 11 participants. To facilitate attendance and retention of participants, free childcare and food were provided at the workshops, parents could make up lessons at flexibly arranged times, and participants formed voluntary WhatsApp groups that contacted absentees and encouraged them to keep coming back. Attendance declined to 85% at the third workshop and to 66% at the tenth (mean = 7.6 workshops).

The study followed university IRB approved protocols for data collection, including a second level cultural review for research with AI populations. Parents in P2W and HF2W completed half-hour, paper and pencil pretest questionnaires during Workshop 1, posttests at Workshop 10, and a follow-up questionnaire one-year later. Parents received incentives to attend the workshops and defray transportation costs ($15) whether they consented to complete the questionnaires (98%) or declined. The immediate posttest was completed by 63% of the participants in both interventions, and 55% completed the one-year follow-up. Workshop attendance and attrition to posttest and follow-up did not differ significantly by the participants’ gender, age, number of children, education, or length of urban residence.

Analysis Strategy

We tested the effects of P2W, relative to HF2W, separately on the two measures of parent/child communication about sex. We first examined pairwise t-tests (assessing changes from pretest to immediate posttest, and pretest to follow-up) and then estimated baseline-adjusted regression models in Mplus 7.0 (Muthén & Muthén, 2012). These models incorporated full information maximum likelihood (FIML) adjustments for attrition and any other missing data, and also corrected the models to take account of any random effects at the site level (Phoenix, Tucson, Flagstaff) and at the facilitator level (10 P2W implementers; 11 HF2W implementers), using numerical identifiers for each facilitator and site. We created interaction terms to test whether P2W effects were moderated by the gender combinations in the parent/child dyad, and ages (mean centered) of the child and of the parent.

Results

Equivalence Tests of P2W and HF2W Demographics

Table 1 contrasts the demographic profiles of the P2W and HF2W samples. With one exception, the two samples did not differ significantly on demographic characteristics, as indicated by Chi-square and independent sample t-tests (not shown in tables). The single difference was that the P2W sample had a somewhat smaller proportion of father/son dyads than the HF2W sample, which produced a significant association between the parent-adolescent gender combinations (FF, FM, MF, MM) and intervention condition (χ² (3) = 9.2, 3 df, p=.03).

In additional tests not reported in tables, we examined whether there were pretest differences in parent-adolescent communication about sex by the gender of the parent or by the adolescent’s gender. Although fathers did report less communication than mothers on both measures (sexuality and safe sex; sexual decision making) [t(552) = 5.02, p < .001; t(553) = 5.91, p < .001], there were no significant differences in communication with daughters than with sons on the same measures [t(548) = 1.056, p = .291; t(547) = 1.89, p = .060].

Changes in Immediate and Long-term Outcomes

Table 2 reports the means on the two measures of communication about sex at each survey wave, separately for the P2W and HF2W participants. The first is the more general assessment of discussing pubertal changes and safe sex, while the second is specific to decision-making about having sex. At the pretest (T1), both outcome measures had means near the midpoint of the 1–4 range, which corresponds to communicating about sex “a few times” or “occasionally.” There were no significant differences between the P2W and HF2W participants on these baseline measures (t = .13, p = .90; t = .70, p = .48). The means for both measures increased from the pretest (T1) to the immediate post-test (T2) and from pretest to the one-year follow-up (T3) in both P2W and HF2W. In P2W, increases from T1 to T2 and from T1 to T3 were significant for both measures of communication about sex (t(311) = 5.90, p < .001; t(311) = 5.89, p < .001; t(311) = 5.06, p < .001; t(311) = 4.99, p < .001). In HF2W, there were no significant changes from T1 to T2, but the T1 to T3 increases were significant for both measures (t(261) = 3.85, p < .001; t(261) = 4.28, p < .001). Cohen’s d estimates show that increases in communication about sex on both measures at the immediate follow-up were larger in P2W (d > 0.30) than in HF2W (d < 0.10), while at the one-year follow-up the effects were closer but still somewhat larger in P2W (d > 0.33) than in HF2W (d = 0.28; d = 0.29). Comparing the changes in P2W to those in HF2W, t-tests show that the increases from T1 to T2 were significantly greater in P2W than in HF2W on both measures (t(573) = 2.83, p = .002; t(573) = 2.62, p = .005). Cohen’s d estimates of the P2W intervention effect relative to the HF2W effect were above 0.20 for these T1 to T2 changes. Relative changes from T1 to T3 in the two interventions, however, were not significantly different on either outcome measure.

Table 2.

Changes in Immediate and One-Year Outcomes, by Intervention Condition

			Sexuality and Safe Sex	Sexual Decision Making
Parenting in 2 Worlds N=312	T1	M (SD)	2.497 (1.009)	2.432 (1.058)
	T2	M (SD)	2.821 (0.918)	2.750 (1.044)
	T3	M (SD)	2.859 (1.029)	2.770 (0.948)
	Difference test (t)	T1-T2	5.903***	5.064***
		T1-T3	5.892***	4.991***
	Cohen’s d	T1-T2	0.336	0.302
		T1-T3	0.355	0.336
Healthy Families in 2 Worlds N=263	T1	M (SD)	2.500 (1.019)	2.357 (1.054)
	T2	M (SD)	2.602 (1.042)	2.453 (1.044)
	T3	M (SD)	2.764 (0.873)	2.666 (1.072)
	Difference test (t)	T1-T2	1.824	1.172
		T1-T3	3.853***	4.283***
	Cohen’s d	T1-T2	0.099	0.092
		T1-T3	0.278	0.291
P2W versus HF2W	Difference of Differences (t)	T1-T2	2.829**	2.618**
		T1-T3	0.091	0.024
	Cohen’s d	T1-T2	0.236	0.203
		T1-T3	0.081	0.020

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

Tests of Relative Intervention Effects

Table 3 presents the tests of intervention effects using linear models adjusted for baseline (i.e., controlling for the measure at pretest), for missing data or attrition, and for any random effects due to the site/city and facilitator. The table contains separate panels for short-term (T2) and long-term (T3) effects, for both measures of communication about sex. All models control for the gender composition of the parent-adolescent dyads, with mother/daughter pairs serving as the omitted reference group, and for the parent’s age and the child’s age. There are two models for each outcome—one with intervention effects (P2W v. HF2W as a dummy variable) controlling for gender and age, and a second adding interactions of the intervention effect with the gender pair combination, and with the ages of the parent and of the child.

Table 3.

Immediate and One-Year Intervention Effects on Parent-adolescent Communication about Sexual Health (N=572)

	Sexuality and Safe Sex (T2)				Sexual Decision Making (T2)
	b	SE	b	SE	b	SE	b	SE
Intercept	0.397	(0.241)	0.495	(0.420)	0.254	(0.324)	0.782*	(0.369)
Dependent Variable @ T1	0.605***	(0.046)	0.607***	(0.048)	0.552***	(0.037)	0.549***	(0.038)
Mother/Son	−0.024	(0.072)	−0.183**	(0.066)	−0.048	(0.070)	−0.181	(0.163)
Father/Son	−0.304**	(0.089)	−0.348***	(0.099)	−0.247*	(0.114)	−0.256†	(0.140)
Father/Daughter	−0.305**	(0.089)	−0.517**	(0.157)	−0.363**	(0.119)	−0.398†	(0.206)
Child Age	0.048***	(0.011)	0.039†	(0.021)	0.061**	(0.021)	0.024	(0.018)
Parent Age	0.004	(0.004)	0.007	(0.005)	0.005	(0.004)	0.006	(0.005)
P2W versus HF2W	0.218**	(0.081)	0.067	(0.077)	0.240***	(0.068)	0.147	(0.148)
P. v. H. X Mother/Son			0.295**	(0.091)			0.239	(0.224)
P. v. H. X Father/Son			0.085	(0.158)			−0.010	(0.173)
P. v. H. X Father/Daughter			0.356†	(0.201)			0.044	(0.261)
P. v. H. X Child Age			0.019	(0.025)			0.074***	(0.022)
P. v. H. X Parent Age			−0.007	(0.008)			−0.003	(0.008)
R2	0.525		0.533		0.467		0.474
	Sexuality & Safe Sex (T3)				Sexual Decision Making (T3)
	b	SE	b	SE	b	SE	b	SE
Intercept	1.129**	(0.341)	0.675	(0.495)	0.961***	(0.330)	0.715	(0.502)
Dependent Variable @ T1	0.555***	(0.049)	0.563***	(0.049)	0.513***	(0.055)	0.527***	(0.059)
Mother/Son	0.033	(0.102)	−0.239	(0.190)	0.033	(0.120)	−0.248	(0.207)
Father/Son	−0.116	(0.154)	−0.308*	(0.146)	−0.043	(0.182)	−0.397	(0.261)
Father/Daughter	−0.192	(0.160)	0.035	(0.160)	−0.344*	(0.159)	−0.413†	(0.219)
Child Age	0.012	(0.020)	0.040	(0.031)	0.039*	(0.019)	0.071*	(0.030)
Parent Age	0.003	(0.004)	0.008	(0.007)	0.001	(0.005)	0.000	(0.009)
P2W versus HF2W	0.080	(0.093)	−0.119	(0.135)	0.074	(0.077)	−0.219	(0.136)
P. v. H. X Mother/Son			0.483*	(0.202)			0.498*	(0.243)
P. v. H. X Father/Son			0.494	(0.326)			0.869**	(0.284)
P. v. H. X Father/Daughter			−0.362†	(0.216)			0.169	(0.234)
P. v. H. X Child Age			−0.045	(0.041)			−0.062†	(0.034)
P. v. H. X Parent Age			−0.011	(0.010)			0.000	(0.011)
R2	0.394		0.424		0.365		0.394

Note: Estimates are unstandardized regression coefficients.

^†p < 0.10.

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

The initial models, without interactions, mirror the t-test results, showing desired overall effects of P2W, compared to HF2W, from T1 to T2, but not from T1 to T3. For both measures, increases in communication about sex to the first posttest were significantly greater in P2W than in HF2W (b = 0.218, p < .01; b = 0.240, p < .001). The initial models also show that, compared to mother/daughter pairs, the father/son and father/daughter pairs reported significantly smaller increases in communication about sex on both measures from T1 to T2, and father/daughter pairs had smaller increases on sexual decision making from T1 to T3. Parents communicated about sex more with older children from T1 to T2 on both measures, but from T1 to T3 only on the sexual decision-making measure. Parental age was not a significant predictor of changes in sexual communication in any model.

The second models with interactions identify sub-groups where increases in parent-adolescent communication about sex were greatest for P2W participants. P2W had the strongest immediate effects (T1-T2) in increasing communication about sexuality and safe sex among cross-gender pairs: larger effects for mothers/sons and fathers/daughters than for mothers/daughters. However, long-term (T1-T3) this effect persisted only for mother/son pairs while reversing direction for father/daughter pairs (i.e., weaker effects compared to mother/daughter pairs). For the more specific sexual decision-making measure, the interactions indicate that P2W had stronger immediate effects for older children, but no variation in the effect by the gender composition of the pair. Long-term effects for this outcome show that P2W improved communication especially with sons, both in mother/son and father/son pairs, but the interaction effect with child’s age reversed direction (i.e., weaker P2W effects for older adolescents). None of the interaction models showed that effects varied by the age of the parent.

Discussion

This study tested whether the P2W program increased parent-adolescent communication about sexual health, relative to an alternate program; whether the effects were both short-term and long-term; and whether effects varied by the gender and ages of parent and adolescent. Results showed that P2W parents significantly increased communication on both measures examined at both the immediate and one-year posttest. Although there were no significant short-term changes among those in the alternate program, this group did report an increase in communication in the long-term. Evaluating the effects of the two programs relative to each other, the increases in communication were significantly greater in P2W than the alternate program in the short term, but not long-term. This pattern of effects could be interpreted as the fading of initially strong P2W effects over time, and/or that parents in the alternate program delayed communication about sexual health until their teens became more mature sexually or alternatively old enough to date. Noting that means for P2W parents leveled off but did not decline from the immediate to the long-term posttest, P2W may have triggered conversations about sexual health at earlier adolescent ages than they may otherwise have occurred, and then helped maintain their frequency.

The short-term effect sizes for P2W were between 0.2 and 0.3 for both measures, which would be considered “small” by conventional thresholds once recommended by Cohen (1988), benchmarks that are increasingly questioned (Owens et al., 2021). However, these effect sizes are within the typical range of many parenting programs (Cedar & Levant, 1990; Leijten et al., 2019). Further, because P2W effects are relative to an alternate treatment, the effect sizes are likely to underestimate what the effects would be compared to no-treatment controls.

The culturally tailored P2W intervention focused on the history, strengths and resilience of AI tribes, encouraging parents to strengthen, connect, or reconnect to their individual tribal histories and embrace shared values with other parents of different tribal backgrounds in their groups. This process of cultural revitalization led to significant increases in AI ethnic identity and spirituality in P2W parents (Kulis et al., 2020). Other studies have found that cultural revitalization and connectedness (e.g., knowledge of history and culture, engagement in traditional practices, and connections to kin, the earth, and ancestors) are powerful protective factors that promote health and well-being (Allen et al., 2018; O’Keefe et al., 2022; Shea et al., 2019; Ullrich, 2019). Thus, the cultural nature of this intervention may have led to the demonstrated increase in communication about sex with their adolescents in comparison to the information only HF2W curriculum. For historically matrilineal tribes in particular, researching their cultural beliefs likely led them to remember or discover the importance of women as the conveyors of tribal and cultural knowledge, with mothers and children regarded as sacred (Dalla & Gamble, 2001; Ullrich, 2019). Sharing this knowledge with sons and daughters might provide a way for parents to instruct their adolescents on how they should treat and be treated in intimate relationships and the importance of protecting themselves. Moreover, the demonstrated increase in parental AI ethnicity and spirituality likely helped strengthen the parent/adolescent relationship as the fostering of cultural values, historical cultural knowledge, and teaching of cultural lifeways all contribute to enhanced connections, and increased health and well-being among Indigenous youth (O’Keefe et al., 2022).

Findings on whether there were differential effects by the gender of the parent and adolescent varied short-term and long-term, and for different types of communication about sexual health. For the measure of general communication about sexuality and safe sex, the desired effects of P2W were stronger for cross-gender dyads at the immediate posttest, persisted for mother/son dyads one-year later, but reversed long-term for father/daughter dyads. These effects are best interpreted relative to the reference group of mother/daughter dyads, the gender pair that reported the most frequent communication of this type at the pretest and both post-tests. Thus, the effects of P2W on this measure point to its ability to improve communication in the gender dyads least likely to otherwise engage in it.

Why this intervention effect eroded and reversed direction over time for father/son dyads is less clear. Fathers might have delayed these conversations until their sons were seen as more sexually mature, prompting more of this communication overall among fathers and sons at the long-term posttest in both the P2W and alternate intervention groups, thereby narrowing the impact of the P2W intervention. It is important to note however, that targeting both mothers’ and fathers’ parenting behaviors has led to increased interaction by both parents over the long term (Jeong et al., 2019). It is possible that fathers were slow to start, but after observing the increased interactions with the mother became more comfortable over time. If so, the increased interaction of fathers with sons at the long-term posttest might actually lead to more sustained effects over time. Communication barriers might also have led to less communication over time as research among Black and Latino fathers found communication barriers existed due to fathers not knowing how to correctly instruct their teens about sex education (Guilamo-Ramos et al., 2019). Future intervention efforts with AI youth might want to specifically recruit fathers as they are an important source of information for both sons and daughters (Guilamo-Ramos et al., 2023; Stewart et al., 2019).

For the more specific measure of communication about sexual decision-making, short-term effects of P2W did not vary by gender but were stronger long-term for mothers and fathers of adolescent sons. This may reflect different types or levels of concern about the sexual and reproductive health of daughters and sons. In the general population, these parent-adolescent conversations are more frequent and more strongly linked to positive sexual health outcomes when they involve daughters, reflecting heightened concern about the immediate and long-term consequences of pregnancy for females (Widman et al., 2016). However, AI families may view these concerns relative to cultural values that welcome the arrival of a daughter’s child into the family regardless of the daughter’s age, and with the expectation that the extended family will help support the mother and child (Hafner & Rushing, 2019). For many traditionally matrilineal tribes, a daughter’s child will stay with the family whereas a son’s child stays with the mother’s family (Dalla & Gamble, 2000). If a teenage son’s child is expected to live with the mother’s family, the parents of a son may have more conversations with him about his sexual decision making and choice of a partner as he gets older. That said, it is important to note that cultural values have shifted and not all parents will accept a teenage daughter’s child into their home (Dalla et al., 2009). The results may also reflect that P2W parents of sons became more aware of possible consequences for their sons and families, and acquired the skills to discuss them. Other research has highlighted the importance of parental conversations with male AI adolescents about contraception (Chambers et al., 2018)

Differences in P2W effects for younger and older adolescents can be interpreted relative to findings on the overall age effects, that parents reported more communication with their older adolescents about safe sex and sexuality (short-term) and about sexual decision-making (both short- and long-term). P2W effects by adolescent age were confined to the measure of sexual decision-making, with stronger P2W effects short-term, and weaker effects long-term. While P2W may at first have heightened parental concern about having these conversations with older adolescents, having had them, they were not as likely to be repeated. P2W parents of younger adolescents who did not have the conversations short-term may have been motivated long-term to have them when their adolescent had matured by a year. Notably, parental age was unrelated to either measure of parent-adolescent communication about sex, either as a main effect or as a moderator of P2W effects. P2W appears to have robust effects for parents of all ages.

Limitations and Future Directions

Certain limitations are important to note in interpretations of the findings. First, the study assessed the amount of parent-adolescent communication about sex, not its content, and did not include measures of proposed mechanisms by which this communication may lead to better adolescent sexual health outcomes (Rogers, 2017). The proposed models rest on assumptions that parent-adolescent communication about sex influences the adolescent’s cognitions, which then affect intentions and behaviors. How well these models apply broadly to AI populations is unclear, and a proper subject for future research. Moreover, our study had no information from adolescents to verify the degree of communication or assess its direct impact on the adolescent.

Second, the gender imbalance in the adult participants placed constraints on the ability to assess how P2W influenced communication about sex in parent-child dyads with different gender compositions. Only 23% of the parents/guardians were male, and attrition reduced their numbers to 57 in the final posttest, with about half reporting on sons and half on their daughters. Although appropriate statistical adjustments for missing data were applied in tests of intervention effects, it is possible that larger numbers of male participants would result in additional significant differences among the gender dyads.

Differences between P2W and HF2W parent-adolescent communication about sex may be due to differences in the content and focus of the lessons that addressed adolescent sexuality. P2W addressed the issues more comprehensively with information parents need to know about the child’s sexual development, how to locate accurate sources of information, when and how to have conversations about sexuality, and how to prepare youth to navigate sexual risks. HF2W was more narrowly focused on recognizing, avoiding and dealing with unhealthy sexual relationships and dating violence. The study was unable to delve into the particular components or combinations of components in P2W that prompted more communication about sex. Future research can include qualitative work to gain more insight into how parents actually applied these lessons to learn what they talked about with their adolescents when they had the often difficult conversations about sex and sexuality.

Third, we could not assess whether parent-adolescent communication about sex may have differed for parents and/or adolescents with non-binary gender identities or non-heterosexual sexual orientations, as that information was not collected in the parent surveys. Given the study sample size, it is unlikely that these subgroups could be meaningfully examined in separate or comparative analyses. Although P2W was designed to help all caretakers of all urban AI adolescents, regardless of their relationship to each other or their other identities, additional research would be required to show the generalizability of its effectiveness with these and other diverse groups, such as differences by tribal background, or past experience living on tribal land.

For parents, practitioners, and interventionists, this study highlights the importance of coming from a strengths-based approach that accounts for the family’s traditions and values as an effective way to increase parent-adolescent communication about sex. Moreover, this study supports the importance of strengthening, reconnecting, and connecting urban AI parents to cultural knowledge and practices as a powerful way to promote the well-being of AI children (Ullrich, 2019). Specifically, we found evidence that this culturally grounded parenting intervention significantly increased parent-adolescent communication about sexual health. With the many threats to the safety and well-being of AI adolescents today, and the disproportionate number of both boys and girls who go missing or are murdered by intimate partners as well as strangers (Ficklin et al., 2022; Yuan et al., 2015), the importance of increasing these conversations cannot be overstated. AI parents today are well aware of the threats to their children’s safety and have adopted child-centric parenting approaches to protect their children by monitoring them closely and shielding them from exposure to negative influences (McKinley et al., 2021). Culturally tailored parenting interventions can provide AI parents with the connections to traditional knowledge and practices needed to keep adolescents safe.