Connectedness as a Predictor of Sexual and Reproductive Health Outcomes for Youth

Family connectedness 

Ninety studies (34 longitudinal and 56 cross-sectional) examined the direct association between family connectedness and ASRH outcomes; 15 studies also examined indirect effects. Specific measures used to assess family connectedness varied across studies. Although the majority (n = 49) used measures of parental closeness (e.g., warmth, caring, support, attachment to family), other studies (n = 32) used measures such as satisfaction with the parental relationship, quality of relationship, and family cohesion. Fifteen studies examined parental involvement (e.g., how many activities the parent and child do together). The majority of studies (n = 76) assessed family connectedness from the adolescent perspective; however, six studies used parent report, six used both adolescent and parent report, and two used observer ratings. Although the majority assessed family connectedness through multi-item scales, nine studies used single-item measures. Cronbach's alphas for multi-item scales ranged from .53 to .95, with over 90% being greater than or equal to .70.

Table 1 shows the findings of direct association between family connectedness and ASRH outcomes. Overall, there was sufficient evidence that family connectedness can be protective, with at least two longitudinal studies demonstrating a protective association for four ASRH outcomes (ever had sex, early sexual debut, frequency of sex, and pregnancy/birth). No longitudinal studies reported a risk association.

Table 1. Distribution of reviewed studies' findings related to the association between family sub-constructs and adolescents' sexual and reproductive health outcomes

	Sexual behaviors by sub-construct	Nature of association
		Protective association	Risk association	No association
	Family connectedness (34 longitudinal and 56 cross-sectional studies)
	Ever had sex	14a, b[18], [19], [20], [21], [22], [23], [24]		31a, b[18], [19], [21], [24], [25], [26], [27], [28], [29], [30], [31], [32]
		12b, c[13], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42]		24b, c[16], [33], [34], [36], [37], [43], [44], [45], [46], [47], [48], [49], [50]
	Recent sex/current sexual activity	6c[51], [52], [53], [40]		3a[54], [55]
				7c[36], [46], [51]
	Early sexual debut	2a, b[56], [60]	2c[57], [58]	4a[56], [57], [59]
		7c[36], [52], [58], [61]		10c[36], [57], [58], [62], [63]
	Use of contraception	1a[23]		2a[64]
		4c[40], [46], [53], [43]		6c[42], [43], [46]
	Use of condom	2a[67]		2a[67]
		3c[68], [69], [70]		15c[36], [44], [46], [68], [70], [71]
	No. sexual partners	8c[51], [52], [45], [58], [72], [73]		1a[27]
				11c[51], [58], [73], [74], [75]
	Frequency of sex	3a, b[76], [77]		3a[76]
		8c[52], [77], [78], [79]		13c[42], [74], [78], [80]
	Sexual risk index	2a[81]		2a[81]
		4c[69], [82], [84]		32c[83], [68], [84], [85], [86]
	Contracted an STI			1a[87]
	Pregnancy/birth	4a, b[23], [88], [89]		12a, b[64], [77], [88], [89], [90], [91], [92], [93], [99]
		5b, c[94], [36], [95]		15b, c[35], [36], [63], [77], [95], [96], [97], [98]
	Intentions	1c[100]		2c[38], [100]
	Subtotal	28a	0a	61a
		59c	2c	135c
	General parent–adolescent communication (four longitudinal and 12 cross–sectional studies)
	Ever had sex	2a, b[101], [102]		4b, c[103], [83]
		5b, c[35], [101], [102], [103], [104]
	Recent sex/current sexual activity	3a[105]		4a[105]
	Early sexual debut	2a[105]		5c[50], [106], [107]
	Use of condom	1a[102]		5c[103], [106], [107]
		3c[102], [103], [108]
	No. sexual partners	3c[106]		6c[106], [107]
	Frequency of sex	4c[106], [107]		5c[106], [107]
	Sexual risk index			1c[109]
	Contracted an STI			1c[103]
	Pregnancy/birth	2c[96], [110]		4a[110]
				5c[35], [110]
	Subtotal	8a	0a	8a
		17c	0c	32c
	Parent-adolescent communication about sex (12 longitudinal and 46 cross-sectional studies)
	Ever had sex	3c[111], [112], [113]	3a[114]	11a[18], [31], [114], [115], [116], [117]
			5c[119], [40], [45], [46], [118]	9c[46], [112], [113], [119], [120], [121], [122], [123]
	Recent sex/current sexual activity		2c[46]	1c[120]
	Early sexual debut	2a, b[19], [115]	1a, b[115]	10a, b[19], [115]
		8b, c[103], [124], [125], [126], [127], [128]	2b, c[124]	15b, c[106], [107], [112], [125], [127]
	Use of contraception	8c[120], [123], [124], [125], [129], [130]	1a[115]	3a[115]
				9c[31], [40], [113], [120], [124], [125]
	Use of condom	1a[131]	1a[115]	3a[115]
		15c[103], [122], [126], [130], [132], [133], [134]		26c[70], [103], [106], [107], [112], [124], [132], [133], [134], [135]
	No. sexual partners	5c[124], [132]	4c[106], [45], [124]	5a[115], [131]
				6c[106], [107], [132]
	Frequency of sex	1a[131]	2c[106]	6c[40], [78], [106], [107], [136]
		1c[107]
	Sexual risk index	1a[81]		3a[81], [137]
		4c[138], [139], [140]		8c[86], [109], [138], [140]
	Contracted a STI	1c[141]
	Pregnancy/birth	4a[142]	3c[124]	14a[142]
		3c[176], [143]		2c[113], [124]
	Intentions	1a[137]		1a[137]
		1c[112]		5c[112], [136], [144]
	Subtotal	9a	6a	50a
		49c	18c	87c
	Parental monitoring (27 longitudinal and 34 cross-sectional studies)
	Ever had sex	8a, b[27], [116], [145], [146], [147]	2a, b[29], [148]	13a, b[20], [29], [54], [60], [145], [146], [149], [150], [151], [152]
		13b, c[33], [39], [49], [50], [122], [153], [154], [155], [156]	2b, c[157]	7b, c[38], [50], [155], [156], [157]
	Recent sex/current sexual activity	3c[51], [106], [158]	2c[157]	4c[46], [157]
	Early sexual debut	6a, b[19], [28], [61], [159], [160], [161]		6a, b[19], [61], [159], [161]
		7b, c[122], [155], [162], [163]		4b, c[111], [155], [163]
	Use of contraception	2a, b[152], [164]		1a, b[149]
				4b, c[43], [155], [156]
	Use of condom	3a, b[102], [164]	1b, c[68]	4b[102]
		6b, c[157], [69]		13b, c[68], [122], [154], [156], [157], [158]
	No. sexual partners	2a[164]		2a[27], [149]
		7c[51], [75], [106], [158], [165]		4c[51], [74], [158]
	Frequency of sex	5c[51], [78], [79], [158]		1a[149]
				6c[74], [78], [158]
	Sexual risk index	1a[166]		1a[167]
		9c[68], [69], [82], [86], [109], [140], [168], [169]		2c[68], [140]
	Contracted a STI	4a, b[164], [170]		3a, b[167], [170]
		2b, c[165], [171]		5b, c[157]
	Pregnancy/birth	1a[160]		6a[91], [172]
		2c[153]		2c[158]
	Intentions	2a[150]	2c[157]	2a[150]
		1c[38]		2c[157]
	Subtotal	29a	2a	39a
		55c	7c	53c

a Indicates the studies were longitudinal in design. b Indicates that it met the standard of evidence (i.e., findings from at least two longitudinal studies provided evidence for a protective or risk association). c Indicates the studies were cross-sectional in design.

Family connectedness was found to be protective of ever having sex in 14 findings from two longitudinal studies [18], [19], [20], [21], [22], [23], [24]; however, 31 findings from 11 longitudinal studies [18], [19], [21], [24], [25], [26], [27], [28], [29], [30], [31], [32] demonstrated no association. For example, two studies using data from the National Longitudinal Study of Adolescent Health (Add Health) indicated that greater levels of mother–child connectedness, parental closeness, and parent–child shared activities were all protective of sexual initiation [18], [21]. Although there was a protective effect for boys between mother–child connectedness and initiation of sex for 10th and 11th graders, no association was found for girls of the same age [21]. There was also no association between family connectedness indicators between fathers and sons and sexual initiation [21]. Notably, more studies reporting “no association” used measures referring to “parents” in general, rather than to a specific parent [21], [25], [28], [29], [30], [31], [32]. Measures referencing specific parents (e.g., mothers vs. fathers), especially those assessing parental involvement [21], [22], were more likely to show a protective effect for ever had sex than parental involvement measures which referenced “parents” in general [25], [28], [32]. One study [31] reported a protective association at the bivariate level only, indicating possible mediation or moderation by other factors.

There was sufficient evidence to suggest that family connectedness can be a protective factor for early sexual debut with two findings from two longitudinal studies indicating a protective association [56], [60]. However, four findings from three longitudinal studies [56], [57], [59] reported no association. Bingham and Crockett (1996) [56] assessed quality of parent relationship and reported a protective association; though, this association became non-significant after adjusting for self-esteem and positive affect prior to first intercourse. Browning and colleagues [60] assessed “family attachment and support” and reported a protective association when controlling for individual, family, and neighborhood predictors. Only one longitudinal study reporting findings of no association reported bivariate analysis [59], indicating a protective association which was not sustained in multivariate analyses.

There was sufficient evidence to suggest that family connectedness can be a protective factor for frequency of sex, with three findings from two longitudinal studies indicating a protective association [76], [77]. However, one of these studies [76] also reported three findings of no association. Sabo et al. [77] assessed “family cohesion” and reported a protective association for males and females among African American and white youth. In contrast, Lauritsen [76] assessed “attachment to family” and reported a protective association for white males only (no association was reported for white females or for African American youth). Given limited descriptions of measures, it was not possible to examine exact differences in item wording or operationalization. Bivariate analyses were not reported.

There was sufficient evidence to suggest that family connectedness can be a protective factor for pregnancy/birth, with four findings from three longitudinal studies [23], [88], [89] reporting a protective association. However, 11 findings from eight longitudinal studies [64], [77], [88], [89], [90], [91], [92], [93] reported no association. One longitudinal study [88] measuring satisfaction with the father–child relationship reported a protective association for pregnancy among athletes but not among nonathletes, whereas another longitudinal study [89] measuring parental involvement reported a protective association among white and African American youth but not among Latinos. Four of the eight longitudinal studies reporting findings of no association reported bivariate analyses [64], [65], [77], [88], [90]; of these, three [64], [88], [90] reported a protective association which was not sustained in multivariate analyses.

There were too few longitudinal studies to draw conclusions about the relationship between family connectedness and other ASRH outcomes.

Fifteen studies examined the indirect effects of family connectedness on ASRH outcomes. Path analysis from longitudinal and cross-sectional studies revealed that family connectedness had indirect protective effects through individual factors (e.g., academic performance [90], depression [90], substance use [90], self-control [27]) and external factors (e.g., peer deviance [90], number of partners [25], and eating meals with a parent [25]). One longitudinal study from England reported an indirect risk effect for early sexual initiation [26]. This risk effect was mediated through decreased conflict and through lack of explicit parental standards or expectations for their child's behavior. Another longitudinal study [173] reported a more complex pathway involving both family connectedness (operationalized as family involvement) and parental monitoring. Family involvement mediated the effects between family conflict and parental monitoring; parental monitoring then affected “problem behavior” (e.g., sexual risk-taking), both directly and through youth association with deviant peers.

Regarding generalizability of findings for specific sub-groups, there was sufficient evidence to support mother–child connectedness, specifically, as a protective factor for both males (three findings from two longitudinal studies [18], [21]) and females (two findings from two longitudinal studies [19], [24]) for ever had sex. There were insufficient longitudinal studies to make any definitive conclusions about racial/ethnic or age-level sub-groups.

Parent–adolescent general communication 

Sixteen studies (four longitudinal and 12 cross-sectional) examined the direct association between parent–adolescent general communication and ASRH outcomes. Three studies [101], [102], [105] included longitudinal and cross-sectional analyses. One study examined the indirect effect of parent–adolescent general communication [179]. Measures used to assess parent–adolescent general communication varied widely. Three studies used a single item to assess quality of communication [101], [103], [110]. Six studies used the Parent–Adolescent Communication scale for open communication and problem communication [83], [96], [102], [104], [174], [179]. Two studies adapted the Parent–Adolescent Communication scale, and three developed original scales, one of which included an item about sexual and reproductive health-related communication. One study used a parent–adolescent conflict-focused communication indicator, whereas another used a combined index score for quality of maternal–adolescent general communication and mother–daughter relationship. The majority of studies (n = 16) used adolescents' self-report; two studies used mother and adolescent reports [106], [107]. Of the 11 studies that presented a Cronbach's alpha, scores ranged from .76 to .91.

Table 1 indicates that there was sufficient evidence that parent–adolescent general communication can be a protective factor for ASRH outcomes, with two longitudinal studies indicating a protective effect for one outcome (ever had sex). No longitudinal studies reported a risk association.

Regarding ever had sex, two findings from two longitudinal studies [101], [102] demonstrated a protective association, whereas no longitudinal studies supported a nonassociation. The two longitudinal studies used measures of general mother–adolescent and father–adolescent communication among males and females in a white clinic-based sample [101] and among African American inner-city youth [102].

Too few longitudinal studies examined the relationship between parent–adolescent general communication and other ASRH outcomes or reported stratified results to assess the generalizability of findings by sub-groups. Regarding indirect effects, one study reported that the influence of parent–adolescent general communication was mediated by peer factors and sex-role attitudes [179].

Parent–adolescent sexual communication 

A total of 58 studies (12 longitudinal and 46 cross-sectional) examined the direct association between parent–adolescent sexual communication and ASRH outcomes. One study included both longitudinal and cross-sectional analyses [31]; 13 included examination of indirect effects. Three additional studies examined indirect effects only.

The measures used to assess parent–adolescent sexual communication were varied. Content ranged from the type and amount of sexual topics discussed between parent and child to the quality or ease of communication. Twelve studies used a single item to assess the construct, 31 used scales ranging from 2 to 12 items (in several articles not enough information was provided to assess the number of items), and four used a combination of both scales and individual items. Eight studies analyzed multiple measures of communication; 10 studies provided insufficient information to assess the type of measure used. Of the 22 studies that presented a Cronbach's alpha, alphas ranged from .65 to .91; 21 studies reported an alpha greater than or equal to .70.

Overall, there was only sufficient evidence to support a protective association between parent–adolescent sexual communication and early sexual debut. Two findings from two longitudinal studies [19], [115] demonstrated a protective association; however, one study also reported a risk association [115], and 10 findings from these same studies indicated no association. Wight et al. [115] examined comfort of communication with fathers and mothers in a mixed-gender sample and reported both a protective and risk association for father–daughter communication only. Females who reported being very comfortable talking about sex with their fathers were less likely to report early sexual debut compared with those who were uncomfortable or very uncomfortable; however, females who had some comfort were at increased risk. In contrast, McNeely et al. [19] reported a protective association for females using a maternal communication measure focused on disapproval of sex. However, they reported no association for males or females regarding other topics of communication. Only one longitudinal study reported bivariate analyses, indicating a protective association for females only [115].

There were too few longitudinal studies to draw conclusions about the relationship between parent–adolescent sexual communication and other ASRH outcomes.

For six outcomes (ever had sex, contraceptive use, condom use, number of sexual partners, frequency of sex, pregnancy/birth), there was some evidence of a risk association; however, these results did not meet our standards of evidence, with consistent findings from at least two longitudinal studies for a given outcome. In one longitudinal study, Pearson et al. [114] found risk associations for ever had sex among female, white, and Latino youth when assessing frequency of communication. In another longitudinal study, Wight et al. reported a risk association for contraceptive use among males related to ease of communication with their father. However, no association was found among males related to ease of communication with their mother, nor among females communicating with their father or mother [115]. The remaining studies indicating risk association were cross-sectional studies.

Regarding generalizability of findings for sub-groups, a key finding was that parent–adolescent sexual communication seemed to be more protective for females than males. Four of eight longitudinal findings [19], [81], [131], [137] found a protective association for females as compared with no longitudinal protective findings for males. Almost half of the cross-sectional findings also reinforced the protective association for females. Too few longitudinal studies stratified findings by race/ethnicity to draw conclusions about generalizabilty; however, cross-sectional studies showed mostly protective results, especially for African Americans. With regard to age, there was sufficient longitudinal evidence for a protective association among high school–aged youth only [81], [115].

The type of measure of communication also appeared to be important–protective influence may be dependent on the quality of communication and the content. For example, “ever” communicating with youth was protective for condom use, frequency of sex, and pregnancy/birth. Quality of communication was protective for intent to have sex for females but not for males [137]. However, ease and frequency of communication [18], [114], [116] were related to risk associations in certain sub-groups.

Parental monitoring 

Sixty-one studies (27 longitudinal and 34 cross-sectional) examined the direct association between parental monitoring and an ASRH outcome; seven studies (five longitudinal and two cross-sectional) also examined an indirect effect.

Most studies used measures assessing the parent's knowledge of their child's whereabouts, friends, or activities (indirect monitoring), or the amount of time that their child spent at home or in public without adult supervision (direct monitoring). Some studies assessed parental strictness or rules about dating and other activities, such as curfews, and a smaller number of studies assessed factors such as parental overcontrol or parental–adolescent decision-making. Fourteen studies used a combination of measures. The majority of studies (n = 45) assessed parental monitoring from the adolescent perspective; however, seven used parental report only, and seven used both youth and parental report. One study [106] reported satisfactory correlation between youth and parental reports (r = .25, p < .01) and combined both sets of items into one scale (Cronbach's α = .71).

Ten studies used single-item measures to assess parental monitoring; the remainder used scales ranging from 2 to 24 items. Of these studies, 30 presented a Cronbach's alpha; the range of scores was .45–.90; 24 (67%) of the alphas presented were greater than or equal to .70. Six studies conducted factor analysis with parental monitoring items. Thirty studies used scales for which validity had been previously established.

Table 1 shows the findings of direct association between ASRH outcomes and parental monitoring. Overall, there was sufficient evidence that parental monitoring can be a protective factor for ASRH outcomes, with at least two longitudinal studies demonstrating a protective association with four outcomes (early sexual debut, contraceptive use, condom use, and contraction of an STI). Findings for ever had sex were mixed, with sufficient findings from two or more longitudinal studies reporting both a protective and a risk association. Furthermore, when findings were broken down by specific behaviors and subgroups, results were inconsistent with some evidence for a risk association.

Findings regarding ever had sex were mixed, with eight findings from five longitudinal studies [27], [116], [145], [146], [147] reporting protective findings for sexual initiation, 13 from 10 longitudinal studies [20], [29], [54], [60], [145], [146], [149], [150], [151], [152] reporting no association, and two from two longitudinal studies [29], [148] reporting a risk association. Hope and Chapple [27] found that parental monitoring at ages 11–13 was protective against sexual initiation 4 years later for males and females in a multi-ethnic sample. Similarly, using data from Add Health, Roche et al. [116] reported that greater parental decision-making (i.e., whether the parent makes decisions for the youth, such as who the youth spends time with, how often, and what they are allowed to watch on TV, curfews, choice of clothing and diet) at ages 12–13 was protective of later sexual initiation for males and females. In contrast, five longitudinal studies conducted in multi-ethnic samples reported no association between parental monitoring and sexual initiation for males or females [29], [54], [145], [149], [150]. One longitudinal study of 12–15-year-old African American and white females found that the type of parental monitoring was influential; direct monitoring of unsupervised time at home was protective against being sexually experienced but no association was found between direct monitoring of time with peers, indirect parental monitoring (i.e., knowledge of youth's whereabouts), and sexual experience [146]. Only two of the 10 longitudinal studies reporting no association reported bivariate analyses. Both reported a protective association at the bivariate level [20], [149], indicating that other factors may influence the effect of parental monitoring on sexual initiation.

Regarding evidence of risk association, two findings from two longitudinal studies indicated that perceived parental overcontrol was a consistent risk factor for ever having sex among males and females [29], whereas perceived stricter parental rules was a risk factor for males [148].

There was sufficient evidence indicating that parental monitoring can be a protective factor against early sexual debut. Six findings from six longitudinal studies [19], [28], [61], [159], [160], [161] indicated a protective association; however, a similar number of findings indicated no association. No studies reported a risk association. Although Longmore et al. [28] reported that parental monitoring was protective of early debut for males and females aged ≥13 years in a predominantly white sample, findings from two longitudinal studies among multi-ethnic samples reported a protective association for females only [19], [159], and one longitudinal study reported a protective association for males only [61]. In a German study, Meschke et al. [161] examined several types of parental monitoring and found that not being alone on a date was protective for females but not for males and that having a curfew was nonsignificant for males and females. Only one of the four longitudinal studies [161] reporting findings of no association reported bivariate analyses. Similar to the multivariate findings from this study, not being alone on a date was protective at the bivariate level but having a curfew was not, adding little information regarding the association between parental monitoring and early sexual debut.

Regarding contraceptive use, there was sufficient evidence to indicate that parental monitoring can be a protective factor among females, predominantly African American youth, with two findings from two longitudinal studies [152], [164] reporting a protective association. However, longitudinal data from the National Survey of Children conducted among African American and white youth indicated no association [149]. Part of the discrepancy may relate to how the dependent variable was measured–the longitudinal studies reporting protective associations assessed use of hormonal or dual methods versus no contraceptive use [152], and contraceptive use at last intercourse [164], whereas the longitudinal study reporting no association assessed contraceptive use during the last 4 weeks [149]. However, the latter study indicated a protective association between parental monitoring and contraceptive use at the bivariate level, indicating the potential presence of mediating factors.

For condom use, there was sufficient evidence to indicate that parental monitoring can be a protective factor, although findings were mixed. DiClemente et al. [164] reported a protective association between indirect parental monitoring and condom use in a longitudinal study of African American females. Similarly, Stanton et al. [102] reported that indirect parental monitoring was a protective factor for condom use at 6- and 18-months later in a longitudinal study of African American, low-income youth aged 9–15 years; however, no association was found at 12-, 24-, 36-, and 48-months, perhaps because of less consistency in monitoring over a period. This same study reported a protective association between indirect parental monitoring and condom use at the bivariate level, indicating that other factors may mediate or moderate the association over time [102].

Regarding contraction of an STI, four findings from two longitudinal studies provided sufficient evidence to support parental monitoring as a protective factor; however, three findings from two longitudinal studies also reported no association. Crosby et al. [170] and DiClemente et al. [164] both reported a protective association between indirect parental monitoring and contraction of certain STIs among African American females aged 14–18 years. However, Crosby et al. [170] also reported no association for gonorrhea and for multiple STIs, both in bivariate and multivariate analyses. Similarly, a longitudinal study of older adolescent white males in Oregon found no association between indirect monitoring and STI acquisition [167].

With regard to recent sexual activity, number of sexual partners, frequency of sex, sexual risk indices, experience of pregnancy or birth, and sexual intentions, there were too few longitudinal studies to draw conclusions about the relationship between parental monitoring and these outcomes.

In addition to the direct associations presented in Table 1, the indirect effects of parental monitoring on sexual health outcomes were also reviewed. Among a predominantly white sample of youth in the Pacific Northwest, Ary et al. [173] reported that the effect of parental monitoring on later problem behavior was mediated through the youth's association with deviant peers. In a longitudinal study of white and non-white youth, Hope and Chapple [27] reported that self-control mediated the effect of parental monitoring on adolescent sexual behavior. These findings provide clues to the potentially more complex effect of parental monitoring on sexual and reproductive health that needs further investigation.

We also considered the generalizability of findings across sub-groups. Substantially greater evidence supported both indirect and direct parental monitoring as a protective factor for females than for males, with 11 longitudinal studies reporting a protective association for females [19], [61], [145], [146], [150], [152], [159], [160], [161], [164], [170] and two longitudinal studies indicating a protective association for males [61], [150]. In addition, two longitudinal studies indicated that parental overcontrol (measured as too many rules or overly strict rules) was a risk factor for males [29], [148]. With regard to race/ethnicity, greater evidence exists to support a protective association between parental monitoring and sexual outcomes among African American youth (10 findings from three longitudinal studies [102], [164], [170]) than among white youth (one protective finding from one longitudinal study [161]). No longitudinal studies presented stratified results for Latino, Asian, or Native American youth, indicating the paucity of longitudinal research within these sub-groups. With regard to age, there was longitudinal evidence for a protective association between parental monitoring and sexual outcomes among middle school youth, with six protective findings from three longitudinal studies [27], [116]. For high school youth, the findings were less consistent. Although eight findings from three longitudinal studies indicated a protective association [19], [150], [164], 10 findings from four longitudinal studies indicated no association. Further, two cross-sectional studies indicated a risk association [68], [157]. No studies reported on youth aged < 20 years who are out of high school.

Peer connectedness 

Nine studies (four longitudinal and five cross-sectional) examined the association between peer connectedness and ASRH outcomes. Of these studies, eight examined direct associations, whereas one cross-sectional study examined indirect effects only. The studies used a variety of measures to assess the quality [51], [57], [74], [94], [175] and frequency of friends' support [81], or the amount of time spent with peers [76]. Six studies used scales to assess peer connectedness; four studies used previously validated instruments or items adapted from established instruments. Four studies provided Cronbach's alphas, ranging from .66 to .82; four (67%) of the alphas presented were greater than or equal to .70.

Table 2 shows the findings of direct associations between ASRH outcomes and peer connectedness. Overall, there was insufficient evidence to evaluate the association because of the limited number of longitudinal studies. However, when results were broken down by specific behaviors and sub-groups, they revealed some evidence of both a risk and protective association.

Table 2. Distribution of reviewed studies' findings related to the association between peer connectedness, partner connectedness, school connectedness, community connectedness and adolescents' sexual and reproductive health outcomes

	Sexual behaviors by sub-construct	Nature of association
		Protective association	Risk association	No association
	Peer connectedness (four longitudinal and five cross-sectional studies)
	Recent sex/current sexual activity			2a[51]
	Early sexual debut		1b[57]	1b[57]
	Frequency of sex		2b[76]	3b[76], [175]
				1a[74]
	No. sexual partners	1b[175]	1a[51], 1b[175]	3a[51], [74]
	Sexual risk index			2b[81]
	Pregnancy/birth	2a[94], [176]
	Subtotal	1b	3b	6b
		2a	1a	6a
	Partner connectedness (four longitudinal and eight cross-sectional studies)
	Ever had sex	2b, c[177], [178]	3b, c[178]	1b, c[177]
			3a, c[16], [179]	1a, c[43]
	Recent sex/current sexual activity		1a[51]	1a[51]
	Use of contraception	2b, c[180], [181]		3b, c[180]
		2a, c[182], [183]		1a, c[43]
	Use of condom	1b[181]	2a[183], [184]	1a[182]
		1a[182]
	No. sexual partners		1a[51]	1a[51]
	Intentions		1a[16]	2a[16], [185]
	Subtotal	5b	3b	4b
		3a	8a	7a
	School connectedness (11 longitudinal and 7 cross-sectional studies)
	Ever had sex	4b, c[149], [178], [186]		7b, c[55], [178], [186], [187]
		5a, c[119], [33], [188]		4a, c[33], [34], [188]
	Early sexual debut	3b, c[13], [59]		1a, c[52]
		1a, c[52]
	Use of condom			2b[149], [189]
	No. sexual partners	1b[189]		1b[149]
		1a[52]		1a[52]
	Frequency of sex	3b, c[76], [77]		6b, c[76], [77], [149]
		1a, c[52]		2a, c[52], [78]
	Sexual risk index	1a[82]
	Pregnancy/birth	1b[172]		4b[77]
	Subtotal	12b		20b
		9a		8a
	Community connectedness (one longitudinal and three cross-sectional studies)
	Ever had sex	3a[33], [190]		2a[35], [190]
	Recent sex/current sexual activity	1b[66]
	Use of contraception	1b[66]		1b[66]
	Use of condom			2b[66]
	No. sexual partners	2b[66]
	Pregnancy/birth			1a[35]
	Subtotal	4b		3b
		3a		3a

a Indicates the studies were cross-sectional in design. b Indicates the studies were longitudinal in design. c Indicates that it met the standard of evidence (i.e., findings from at least two longitudinal studies provided evidence for a protective or risk association).

Regarding frequency of sex, one longitudinal study among a nationally representative sample of 11–17 year olds reported a risk association for African American females and white males related to greater time spent with friends [76]. Similarly, one longitudinal study among white 7th–9th graders (12–15 year olds) reported a risk association regarding early sexual debut for girls and boys; youth who reported better peer relations were more likely to initiate sex before age 16 [57].

Regarding number of sexual partners, one longitudinal study reported a protective association among African American high school students [175]; however, one cross-sectional study conducted in a multi-ethnic rural sample reported an increased likelihood of multiple sex partners in youth aged < 15 years and no association among 16–17 year olds [51].

Regarding pregnancy/birth, two cross-sectional studies conducted among Latino females reported a protective association with peer connectedness [94], [176]. Of the six studies that reported findings of no association, three reported bivariate analyses; two indicated a protective association at the bivariate level [81], [74], and 1 indicated no association [57]. It is possible that some of these associations would have retained significance if another analytic method that considered indirect effects had been used.

Given the limited number of stratified findings by gender and racial/ethnic group, it was not possible to draw conclusions about the association between peer connectedness and ASRH outcomes for sub-groups.

Partner connectedness 

Twelve studies (four longitudinal and eight cross-sectional) examined the association between partner connectedness and an ASRH outcome. All studies investigated the direct association of partner connectedness on ASRH outcomes; one study examined an indirect effect.

A variety of measures were used to assess partner connectedness. All assessed the emotional and social connection and commitment between sexual partners. Two studies used measures that examined whether youth felt that having a child would bring them closer to their partner [51], [43]. Other studies included measures looking at how satisfied the adolescent was with their current boy/girlfriend [185], whether they had a relationship involving caring, closeness, and physical contact [16], or whether they were in love with their partner or needed to feel like they were in love with their partner before being sexually intimate [177], [182]. Additional studies asked youth whether they considered their relationship to be romantic [180], [181], or asked about characteristics of the relationship to determine the level of connectedness [178], [183], [184]. Only two studies included information about the psychometrics of their measures, reporting Cronbach's alphas of 0.60 and 0.92.

Table 2 shows the findings of direct association between ASRH outcomes and partner connectedness. Overall, there was sufficient evidence indicating that partner connectedness can be a protective factor for ASRH outcomes, with at least 2 longitudinal studies demonstrating a protective association with two outcomes (ever had sex and contraceptive use). However, when findings were broken out by sub-constructs and specific outcomes, results were less consistent with some evidence of a risk association, although insufficient to meet the a priori standard of evidence.

There was sufficient evidence to support a protective association between partner connectedness and ever had sex for females; however, differential findings were reported by racial/ethnic group and by type of measure. Bearman and Bruckner [178] reported a protective association for emotional commitment to a partner among African American females; however, it represented a risk factor for ever having sex among male and female non–African American youth, and among African American males. In an English study, Taris and Semin [177] reported a protective association for emotional commitment among females but not in males. However, the authors did report indirect effects for males. The level of emotional commitment that males felt had an effect on the types of courtship strategies (progressing through holding hands, kissing, necking, light and heavy petting, and sexual intercourse) that they felt were permissible, which in turn had a significant effect on whether they had sex.

Findings from two longitudinal studies provided sufficient evidence to support a protective association between partner connectedness and contraceptive use for females; the evidence for males was mixed [180], [181]. Manlove et al. [180] found that the effect of partner connectedness, as measured by whether the participant described their relationship as romantic or nonromantic, on contraceptive use depended on the way the outcome was coded, and by gender. Being in a romantic relationship was not a significant predictor of ever versus never using contraceptives for males or females; however, being in a romantic relationship was protective for always versus no or inconsistent contraceptive use for females. Ford et al. [181] found that adolescents who reported being in a romantic relationship were more likely to use contraception than those who did not categorize their relationship as romantic.

For additional outcomes (recent sexual activity, condom use, number of sexual partners, and intentions), only cross-sectional research findings were available which limits the conclusions that may be drawn.

Considering the generalizability of findings, there was sufficient evidence to support a protective association between partner connectedness and sexual health outcomes among females only [177], [180]. There was insufficient evidence to assess the influence of partner connectedness among racial/ethnic sub-groups or among sub-groups related to age. These studies emphasize the need to better understand under what circumstances, and for whom, partner connectedness may have a protective or risk effect.

School connectedness 

Eighteen studies (11 longitudinal and 7 cross-sectional) examined the direct association of school connectedness on ASRH outcomes. The studies used a variety of measures. Five studies used single items to assess the level of involvement in school activities [52], [76], [77], [187], [189], five used scales to assess whether youth liked school and felt it was important [33], [78], [119], [149], [172], and seven used scales to assess variables such as bonding with peers and teachers, and perceptions of fair treatment and safety at school [13], [34], [55], [82], [178], [186], [188]. Eight studies provided Cronbach's alphas, ranging from 0.57 to 0.89; 5 studies had alphas greater than 0.70.

Table 2 shows the findings of direct association between ASRH outcomes and school connectedness. Overall, there was sufficient evidence that school connectedness can be a protective factor for ASRH outcomes, with at least 2 longitudinal studies demonstrating a protective association with three outcomes (ever had sex, early sexual debut, and frequency of sex). No studies demonstrated a risk association.

Four findings from 3 longitudinal studies [149], [178], [186] found school connectedness to be protective for ever had sex; however, 7 findings from 4 longitudinal studies [55], [178], [186], [187] found no association. Bearman and Bruckner [178] assessed feelings of closeness, inclusion, fair treatment, and happiness at school, reporting a protective association for non–African American adolescents only. McNeely and Falci 186] assessed students' report of teacher support and feelings of social belonging using data from the National Longitudinal Study of Adolescent Health; across the sample, only teacher support was protective for ever had sex. Among a sample of white and African American youth, Baumer and South found that self-reported “school attachment” was protective of sexual initiation [149]. In contrast, Ohannessian and Crockett (1993) [187] assessed participation in academic activities and reported a protective association for white males and females at the bivariate level; however, the multivariate association was nonsignificant. Similarly, Cleveland [55] reported a bivariate level protective association between school attachment and virginity status, which did not retain significance in multivariate analyses, indicating that other factors may mediate or moderate these associations.

Regarding early sexual debut, three findings from two longitudinal studies [13], [59] reported protective effects for school connectedness among males and females; no longitudinal studies reported results of risk or no association.

There was sufficient evidence to support a protective association between school connectedness and frequency of sex with three protective findings from two longitudinal studies [76], [77]; however, findings were inconsistent across sub-groups and by type of school connectedness. Lauritsen [76] reported a protective association for African American females but not for white females, nor for African American or white males. Sabo et al. [77] reported a protective association related to sports involvement for females but not for males, whereas involvement in arts was protective for males but not for females. Baumer and South [149] reported a protective association at the bivariate level only, indicating that other variables may have mediated the association.

There are very few longitudinal studies to draw conclusions about the relationship between school connectedness and other ASRH outcomes (condom use, number of sexual partners, sexual risk index, and pregnancy/birth).

Regarding generalizability of findings across sub-groups, there was sufficient evidence for a protective association among males [59], [77], [178] and among females [59], [76], [77], [172], [178], [189]. However, there was insufficient evidence to draw conclusions regarding the influence of school connectedness by race/ethnicity or by age.

Community connectedness 

Four studies (one longitudinal and three cross-sectional) investigated the direct association of community connectedness on an ASRH outcome. A variety of measures were used to assess community connectedness. Two studies used measures that examined perceptions of neighborhood social support (Quality of Neighborhood Scale) [191] and neighborhood monitoring [33], [35]. These scales had Cronbach's α, .63 and .69, respectively. One study used a single-item measure that assessed youth participation in community-based African American organizations or social groups, one study used a scale from the Youth Asset Survey (α = .78), and one study used a count of the number of community activities in which the youth is involved but provided no information on the reliability of the measure.

Table 2 shows the findings of direct association between ASRH outcomes and community connectedness. Based on the small number of studies, there was insufficient evidence to conclude that community connectedness is a protective factor. However, no studies demonstrated a risk association, and the stratified results revealed areas for additional research.

One longitudinal study [66] found community connectedness to be protective for recent sex, contraceptive use, and number of sexual partners among African American female adolescents. This study showed that youth involved in African American organizations were less likely to report recent sex and more likely to report fewer sexual partners and greater contraceptive use with their steady partner. However, the study also found no association between involvement in African American organizations and contraceptive use with a casual sexual partner and consistent condom use with either a steady or casual partner. Two cross-sectional studies found community connectedness to be protective for ever had sex [33], [190]. Oman et al. [190] found community connectedness to be protective for youth living in a two-parent family but not for those living in a one-parent family. Small and Luster [33] found community connectedness to be protective for both males and females for ever had sex, but another cross-sectional study [35] found no association between community connectedness and ever had sex among African American females.

Given the small number of identified studies, it was not possible to assess the generalizability of findings. The one longitudinal study [66] used a sample of African American females, as did one of the cross-sectional studies [35]. All of the studies included youth of high school age, so it was unclear whether the influence of community connectedness is differential by age.