Influence of positive development opportunities on student well-being, depression and suicide risk: the New Zealand Youth Health and Well-being Survey 2012

ABSTRACT

This study explores the association between neighbourhood opportunities for Positive Youth Development (PYD) and adolescent depressive symptoms, well-being and suicide risk. A 2-stage random sample of 8500 students(years 9–13) from 91 high schools from throughout New Zealand was collected in 2012. Analyses were restricted to neighbourhoods with more than 10 students resulting in a final sample of 5191 adolescents within 266 neighbourhoods. Multilevel models linked data from neighbourhoods to individual student data to explore the association between neighbourhood opportunities for PYD and student depressive symptoms, well-being and suicide risk. Neighbourhoods with a high proportion of students involved in activities that help others, attend church groups, or participate in sports teams were associated with significantly increased well-being among students living in these neighbourhoods. No neighbourhood-level measures were found to significantly alter rates of depressive symptoms or suicide risk. Findings suggest that providing neighbourhood opportunities for PYD may enhance adolescent well-being.

KEYWORDS:

• Adolescent
• health
• well-being
• Positive Youth Development
• neighbourhood

Policy-makers increasingly recognise that a society with healthy, vibrant and contributing young people is essential for future economic and social well-being (Little & Green 2009). Identifying and investing in the factors that contribute to young people’s well-being remains a significant area of enquiry. Positive Youth Development (PYD) theory provides a framework for addressing and improving the well-being of young people while also reducing the harms (Ministry of Youth Affairs 2002). The PYD framework posits that development is a bi-directional (i.e. person–environment) process and external environmental assets such as families, schools and communities are important in supporting healthy outcomes for young people (Benson et al. 1998). These environments provide young people with belonging and connection, places for learning and recreation and chances to contribute (Blum 1998). PYD considers adolescents to inherently develop well if provided with environments that support and encourage pro-social norms, resilience and confidence in their identities and future, negating a ‘deficit perspective’ that focuses on isolated risk behaviours (Roth et al. 1998).

While recognising the crucial importance of family and school environments on PYD, the neighbourhood and communities in which young people live may provide additional opportunities for PYD. The optimal community settings for PYD are safe, structured and supervised, cultivating belonging, pro-social norms, a sense of mattering and skill-building (Eccles & Gootman 2002). Extracurricular activities act as a conduit for healthy development through their ability to provide these factors, integrating the efforts of families, schools and communities (Urban et al. 2010). Community-level factors have been examined in relation to supportive other adult relationships, caring neighbourhoods, communities that value young people, opportunities to serve others, and access to adult role models, creative activities, youth programmes and religious communities (Benson 2007).

Our hypothesis was that the availability of opportunities for PYD at the community level would be associated with fewer symptoms of depression and suicidality and higher levels of well-being. We conceptualise student participation in arts, drama, sports teams, volunteering and belonging to community organisations as measures of community opportunities for PYD in each neighbourhood. This study aims to examine the association between community-level opportunities supporting PYD and the mental health of young people.

Methods

Study design and population

This research draws upon data from the Youth’12 health survey; a nationally representative sample of high school students in New Zealand in 2012. Full details of the methodology and survey design are described elsewhere (Clark et al. 2013a, 2013b).

Students were randomly selected through a two-stage clustered sampling design. First, 125 schools were randomly selected. Of the 91 participating schools, 20% of the student body was randomly selected. In schools with fewer than 150 students, 30 students were selected to reduce risk of identification when data were reported back to schools. In total, 12,503 students were invited to participate in the survey. Of these, 8500 (68%) students took part representing 3.1% of Year 9–13 (equivalent to grades 8–12) students attending an eligible school and 3.0% of all Year 9–13 students in 2012. Written consent was required from each participating school and student, while parents could opt to have their child excluded from the survey. Ethics approval was gained from the University of Auckland Human Participants Ethics Committee (ref 2011/206).

The survey was administered using handheld tablets, allowing questions to be presented in an audio-visual form. The survey consisted of a 608 item branching questionnaire, ensuring that students answered only relevant questions.

Students’ addresses for their usual place of residence or the home where they spend most of their time were entered into a geo-coding programme to ascertain the students’ meshblock and their census area unit (CAU) number, a neighbourhood area of about 2000 people constructed by Statistics New Zealand (2010). These numbers were matched with those in a social indicators research concordance file (Atkinson et al. 2014a) allowing for each student’s data file to include a NZ Deprivation Index decile and score.

Measures

Well-being

The WHO-5 Well-being Index is a validated measure of well-being (Bech et al. 1996, 2003) based on the following five items: ‘I have felt cheerful and in good spirits’, ‘I have felt calm and relaxed’, ‘I have felt active and vigorous’, ‘I woke up feeling fresh and rested’, ‘My daily life has been filled with things that interest me’. The response alternatives were asked over the previous two weeks and included: ‘all of the time’, ‘most of the time’, ‘more than half of the time’, ‘less than half of the time’, ‘some of the time’, ‘at no time’. A mean value of all the responses to the five items (α = 0.89) constituted the WHO-5 Well-being Index score.

Depression

Depressive symptoms were measured by the Reynolds Adolescent Depression Scale Short Form (RADS-SF) (Reynolds 2004), consisting of a 10-item questionnaire. Students were asked ‘How do you usually feel’: ‘I feel happy’, ‘I feel lonely’, ‘I feel like hiding from people’, ‘I feel sad’, ‘I feel like hurting myself’, ‘I feel I am no good’, ‘I feel I am bad’, ‘I feel mad about things’, ‘I feel bored’ and ‘I feel like nothing I do helps anymore’. Depressive symptoms were calculated using a four point Likert response scale (‘almost never’, ‘hardly ever’, ‘sometimes’, ‘most of the time’), with reverse scoring for ‘I feel happy’. The RADS-SF shows high internal reliability (Cronbach’s alpha of 0.88) and high correlation with the full-length RADS, demonstrated in analyses of previous New Zealand Youth Health and Well-being surveys (Milfont et al. 2008).

Suicide risk

Suicidality was assessed with four questions: ‘During the last 12 months have you seriously thought about killing yourself (attempting suicide)?’; ‘During the last 12 months have made a plan about how you would kill yourself (attempt suicide)?’; ‘During the last 12 months have you tried to kill yourself (attempted suicide)’’ and ‘Did this ever result in an injury, poisoning, or overdose that had to be treated by a doctor or nurse?’ with response options 1 (‘not at all’), 2 (‘not in the last 12 months’), 3 (‘once or twice’), 4 (‘three or more times’). Students’ responses were standardised and combined in a suicide risk scale, with a range from 1 to 4 with a mean score of 1.21 (SD = 0.52) and Cronbach’s alpha of 0.81.

Neighbourhood-level variables

Students’ self-reported contribution and connectedness within neighbourhoods was aggregated to the CAU level as a measure of levels of participation and connectedness within students’ neighbourhoods.

Students were asked ‘Do you give your time to help others in your community (e.g. help out on the Marae or church, belong to a volunteer organisation such as Greenpeace)?’. Responses of ‘Yes, within the last 12 months’ or ‘Yes, but not within the last 12 months’ were dichotomised as having helped others.

Students were asked ‘Do you belong to a group, club or team which is not run by your school?’ and could select any number of

A church group; a sports team or group; a cultural group; an environment organisation (e.g. Greenpeace); a volunteer group who help people with disabilities or in hospital; a volunteer group involved with young people, e.g. Youthline; another type of group or club; or, none.

Involvement in environment and/or volunteer groups was grouped as ‘Volunteering’.

Students were asked ‘Over the last 12 months have you worked for money or had a paid job?’ and could select any number of ‘Yes, a regular part-time job (e.g. paper run); Yes, I worked during the school holidays; Yes, I sometimes worked during the school term’ or ‘No, I didn’t work for pay in the last year’. Students who indicated any form of paid employment were categorised as being employed.

Students were asked ‘How much time do you spend doing these activities each day?’. A range of response options were available, and students who spent some time (‘less than 1 hour; 1 to 2 hours; 3 hours; 4 hours; 5 hours or more’) doing ‘music, arts, dance or drama’ were considered to participate in ‘Arts’.

Students were asked ‘If you were having a serious problem is there an adult (who is not in your family) you would feel okay talking to?’ to assess connectedness. Options were ‘Yes’, ‘No’ or ‘Not sure’.

Demographics

Age, sex and ethnicity were self-reported by students. Using the Statistics New Zealand ethnicity prioritisation method, students were assigned to one of the following five ethnic groups: Asian, European, Māori, Pacific and Other ethnicity (Ministry of Health 2004).

Household deprivation was measured using nine socio-economic indicators: family ownership of a car, telephone and a computer/laptop; residential mobility; parental employment; perceptions of level of family worry about not having enough money to buy food; having family holidays; living room or garage used as a bedroom; and living in an over-crowded home. Students who reported two or more indicators of household deprivation were classified in the household deprivation group (Denny et al. 2016).

Neighbourhood deprivation was based on the NZDep2013, an area-based socio-economic Deprivation Index that assesses various dimensions of deprivation using 2013 census data including income, access to the internet, employment, qualifications, home ownership, single parent families, overcrowding and access to a car (Atkinson et al. 2014b). Students were sorted into quintiles according to their residential meshblock’s NZDep2013 deciles. Deprivation Quintile 1 indicates the least deprived and Quintile 5 indicates the most deprived neighbourhoods.

Analyses

All analyses were conducted using SAS software, Version 9.3 (2011; SAS Institute Inc., Cary, NC). Descriptive statistics were generated at the student level by using the survey procedures to account for the clustering and weighting of the dataset. To increase the reliability of neighbourhood measures, analyses were restricted to neighbourhoods with greater than 10 students (Raudenbush & Sampson 1999), resulting in a final sample of 5191 students living in 266 neighbourhoods.

Two-level models of students nested within neighbourhoods were used to estimate the association between neighbourhood opportunities for PYD activities and the mental health and well-being outcomes for students. These models controlled for individual level covariates: age, sex, ethnicity, and household and neighbourhood-level deprivation. Estimation techniques used residual maximum likelihood with ridge-stabilised Newton–Raphson algorithm using the MIXED procedure in SAS version 9.3. P values equal to or less than .05 were considered statistically significant.

To understand the degree of variation between neighbourhoods in the participation and contribution of PYD opportunities, intra-class correlation coefficients (ICCs) were calculated. ICC is the variance at the neighbourhood level divided by the total variance; that is, individual and neighbourhood variance. As these variables were aggregated by dichotomous individual responses, the individual variance was estimated as (Snijders & Bosker 1999).

Further analysis was performed on the significant association between well-being and the neighbourhood variables of helping others, church group attendance and sports team participation. Regression models were used to predict the well-being scores of students living in neighbourhoods with high or low participation in these activities.

Results

Table 1 describes the depressive symptoms, well-being and suicidality by demographic characteristics.

Table 1. Mean depressive symptoms, well-being and suicide risk by demographic characteristic.

	Depressive symptoms^a				Well-being^b				Suicide risk^c
	N	Mean	95% CI	P	N	Mean	95% CI	P	N	Mean	95% CI	P
Total	4992	19.4	(19.2–19.6)		5059	16.4	(16.0, 16.7)		5117	1.21	(1.19, 1.23)
Sex				<.001				<.001				<.001
Female	2654	20.3	(20.0, 20.6)		2680	15.6	(15.2, 16.0)		2703	1.27	(1.25, 1.30)
Male	2336	18.3	(18.0, 18.6)		2377	17.2	(16.9, 17.5)		2412	1.14	(1.13, 1.16)
Age				<.001				<.001				<.001
13 and under	1114	18.8	(18.3, 19.3)		1115	17.3	(16.9, 17.8)		1146	1.17	(1.14, 1.20)
14	1115	19.0	(18.6, 19.5)		1127	16.6	(16.2, 17.1)		1138	1.21	(1.18, 1.25)
15	980	19.6	(19.2, 20.1)		1008	16.2	(15.6, 16.8)		1017	1.24	(1.20, 1.28)
16	940	20.1	(19.7, 20.5)		956	15.7	(15.3, 16.1)		962	1.25	(1.22, 1.28)
17 and over	834	19.6	(19.1, 20.0)		844	15.8	(15.3, 16.2)		845	1.20	(1.17, 1.24)
Ethnicity				.003				<.001				<.001
Māori	985	19.6	(19.1, 20.1)		999	16.4	(15.9, 16.8)		1015	1.26	(1.22, 1.30)
Asian	639	20.2	(19.6, 20.9)		645	15.8	(15.2, 16.5)		653	1.16	(1.12, 1.20)
Pacific	873	19.2	(18.7, 19.6)		896	17.7	(17.0, 18.3)		915	1.28	(1.23, 1.32)
European	2206	19.0	(18.7, 19.3)		2232	16.1	(15.8, 16.3)		2241	1.18	(1.16, 1.20)
Other	284	20.3	(19.5, 21.1)		283	15.6	(14.9, 16.3)		288	1.22	(1.15, 1.30)
Household deprivation				<.001				.02				<.001
Deprivation	1039	21.2	(20.5, 21.9)		1073	15.6	(14.7, 16.5)		1093	1.39	(1.33, 1.45)
No deprivation	3953	18.9	(18.6, 19.2)		3986	16.6	(16.3, 16.9)		4024	1.17	(1.15, 1.18)
Neighbourhood deprivation^d				.71				.01				.01
1 – Least deprived	970	19.2	(18.8, 19.6)		978	15.9	(15.5, 16.3)		982	1.18	(1.15, 1.21)
2	834	19.3	(18.8, 19.9)		839	16.2	(15.8, 16.5)		846	1.19	(1.15, 1.23)
3	896	19.5	(19.0, 19.9)		904	15.9	(15.6,16.3)		915	1.19	(1.16, 1.23)
4	910	19.6	(19.1, 20.1)		917	16.3	(15.9, 16.7)		930	1.22	(1.18, 1.26)
5 – Most deprived	1338	19.4	(19.1, 19.7)		1374	17.2	(16.6, 17.8)		1398	1.26	(1.22, 1.29)

^aRADS-SF score, range 0–35. A score above 28 indicates significant depressive symptoms.

^bWHO-5 well-being score, range 0–25. A score below 13 indicates poor well-being.

^cSuicide risk scale, range 1–4. Highly skewed toward low risk.

^dNZDep2013 quintiles.

Female students reported significantly greater depressive symptoms (P < .001), higher mean suicidality (P < .001) and significantly lower WHO-5 well-being scores (P < .001) than males.

Depressive symptom scores tended to increase with age, peaking in 16-year-olds, compared to a low in those aged 13 years and under (P < .001). Well-being followed the same trend, with the lowest mean score in 16-year-olds, compared to the highest in those aged 13 years and under (P < .001). Suicide risk peaked at age 16 and was lowest amongst students aged 13 years and under (P < .001).

Depressive symptom scores were lowest in NZ European and Pacific Island students and highest in Asian and other ethnicities (P < .01). Pacific Island students showed the greatest well-being with mean score compared to lows in Other and Asian ethnicities (P < .001). Suicide risk was especially high amongst Māori and Pacific Island students compared with NZ European and Asian students (P < .001).

Depressive symptoms did not differ significantly between neighbourhood NZDep quintiles, but students from households experiencing socio-economic deprivation reported higher depressive symptoms (P < .001), lower well-being (P = .02) and higher levels of suicide risk (P < .001) than students from households not experiencing socio-economic deprivation. Conversely, well-being decreased and suicidality increased with neighbourhood deprivation quintile (P = .01).

Table 2 displays participation and contribution by demographic characteristics. Females were significantly more involved in arts, volunteering, church and cultural groups (P < .001) and helping others (P = .03) compared to males. A significantly greater percentage of males were involved in sports and employment than females (P < .001).

Table 2. Neighbourhood participation and contribution by demographic characteristic.

	Helping others		Church group		Sports team		Cultural group		Volunteering		Employment		Arts		Adult to talk to
	%	P	%	P	%	P	%	P	%	P	%	P	%	P	%	P
Total	43.6		24.3		43.9		10.8		19.0		46.8		56.6		58.0
Sex		.03		<.001		<.001		<.001		.004		<.001		<.001		.05
Female	45.1		27.4		38.6		12.8		20.5		43.6		65.1		59.2
Male	41.7		20.7		50.1		8.4		17.3		50.4		46.8		56.5
Age		<.001		.58		.002		.72		<.001		<.001		<.001		<.001
13 and under	39.4		24.7		46.1		10.4		15.3		37.2		58.6		53.6
14	40.4		24.2		47.4		11.4		17.3		40.0		59.6		58.4
15	42.2		25.7		43.7		10.5		19.4		46.8		57.8		54.7
16	45.1		24.2		40.5		9.9		21.5		55.3		55.1		60.2
17 and over	51.7		22.3		40.2		11.7		23.0		58.8		50.6		64.2
Ethnicity		<.001		<.001		<.001		<.001		<.001		<.001		<.001		<.001
Māori	55.1		17.6		44.7		13.5		15.5		48.7		56.7		59.8
Asian	44.1		24.4		31.1		13.1		19.6		31.1		59.4		47.0
Pacific	56.4		54.3		39.8		27.1		12.5		26.9		71.2		52.3
European	35.3		15.4		49.4		3.1		22.4		58.9		49.5		62.3
Other	39.1		24.3		39.9		6.8		22.1		41.7		60.9		59.0
Household deprivation		<.001		<.001		<.001		<.001		<.001		<.001		<.001		<.001
Deprivation	53.6		37.9		34.7		19.5		15.2		34.6		63.9		52.8
No deprivation	41.3		20.7		46.3		8.5		20.0		50.0		54.7		59.3
Neighbourhood deprivation^a		<.001		<.001		<.001		<.001		<.001		<.001		<.001		.09
1 – Least deprived	36.5		13.3		55.3		6.2		22.0		56.4		53.0		59.9
2	37.8		19.2		43.6		5.9		20.4		54.3		52.7		60.6
3	39.2		20.6		46.6		7.1		19.6		51.5		53.4		58.7
4	47.0		22.5		38.3		11.6		20.6		44.3		56.2		56.0
5 – Most deprived	55.4		39.6		37.2		19.4		14.5		33.2		64.3		55.8

^aNZDep2013 quintiles.

Helping others, volunteering, employment and having an adult to talk to all increased with age, and involvement in arts and sports teams decreased with age (all P < .01). Involvement in church and cultural groups remained similar.

There were ethnic differences in community opportunities with Pacific students more frequently reporting church group participation, European and Māori students more frequently reporting sports participation and Māori and Pacific students more frequently reporting helping others (all P < .001).

Students from both more deprived households and neighbourhoods reported they more frequently helped others, attended church, and participated in cultural activities and the arts (P < .001). Students from the least deprived households and neighbourhoods more frequently reported participation in sports, volunteering and employment opportunities (P < .001).

Table 3 shows the mean percentage of students participating in each area across the 266 neighbourhoods, and the range between neighbourhoods. There was high variability in participation between neighbourhoods, demonstrated by the range. For example, no students from some neighbourhoods were involved in church, and over 90% were in other neighbourhoods.

Table 3. Neighbourhood measures.

Measures of participation and contribution	Number of neighbourhoods	Mean percentage	Range	SEM	ICC^a (%)
Helping others	266	44.0	10.00–90.91	1.01	3.5
Church group	266	23.7	0–90.91	1.04	14.4
Sports team	266	44.2	9.52–88.24	0.94	4.3
Cultural group	266	10.2	0–45.45	0.61	13.6
Volunteering	266	19.4	0–69.23	0.70	4.0
Employment	266	48.1	5.00–100	1.15	10.2
Arts	266	56.1	20.00–100	0.87	2.7
Adult to talk to	266	58.3	23.08–100	0.82	2.7

^aICC, a measure of homogeneity.

The ICC represents homogeneity at the neighbourhood level. The greatest homogeneity was seen in participation in church and cultural groups (14.4% and 13.6%); and the least in arts and having an adult to talk to (both 2.7%).

Table 4 presents estimates for predictors of students’ RADS-SF depressive symptom scores, WHO-5 well-being scores and suicide risk. The independent association between each of the indicators of neighbourhood involvement and students’ symptoms were established in multilevel models, adjusted for individual level predictors, age, sex, ethnicity and household and neighbourhood socio-economic deprivation.

Table 4. Fixed effects estimates for models of the predictors of students’ RADS-SF scores, WHO-5 well-being score and suicide risk.

	Depressive symptoms		Well-being		Suicide risk
	Estimate (SEM)	P	Estimate (SEM)	P	Estimate (SEM)	P
Helping others	−0.67 (0.64)	.30	1.66 (0.59)	.005	−0.04 (0.05)	.45
Church group	0.21 (0.67)	.76	1.77 (0.62)	.005	0.02 (0.05)	.689
Sports team	−1.22 (0.67)	.07	1.25 (0.63)	.05	−0.07 (0.05)	.176
Cultural group	−0.87 (1.08)	.42	1.57 (1.01)	.12	−0.03 (0.09)	.72
Volunteering	0.02 (0.86)	.98	−0.28 (0.81)	.73	0.00 (0.07)	.95
Employment	−0.11 (0.58)	.86	0.08 (0.54)	.88	0.08 (0.05)	.07
Arts	−1.09 (0.72)	.13	0.94 (0.68)	.17	−0.07 (0.06)	.20
Adult to talk to	−0.89 (0.75)	.23	1.20 (0.70)	.09	−0.01 (0.06)	.86

Note: Adjusted for age, gender, ethnicity, household deprivation and neighbourhood deprivation quintile.

There is a significant increase in the self-reported well-being of students who live in neighbourhoods with high involvement in church (P = .005), sports teams (P = .05) and helping others in the community (P = .005)

No measure of involvement was found to significantly interact with depressive symptoms or suicidality.

The association between well-being and the statistically significant variables was explored. In the quartile of neighbourhoods with the lowest rates of helping others, self-reported well-being scores were estimated at 15.31 (95% CI 14.91–15.71), compared to 15.97 (95% CI 15.58–16.36) in the quartile with the highest participation (Figure 1). Similarly, in the quartile of neighbourhoods with the least participation in church groups, self-reported well-being was estimated at 15.48 (95% CI 15.08–15.89), compared to 16.15 (95% CI 15.75–16.55) in the quartile with the highest participation (Figure 2). In the quartile of neighbourhoods with the lowest rates of participation in sports team, self-reported well-being scores were estimated at 15.64 (95% CI 15.27–16.01), compared to 15.99 (95% CI 15.58–16.39) in the quartile with the highest participation (Figure 3).

Figure 1. WHO-5 well-being score estimate by neighbourhood quartile: helping others.

Figure 2. WHO-5 well-being score estimate by neighbourhood quartile: participation in church groups.

Figure 3. WHO-5 well-being score estimate by neighbourhood quartile: participation in sports teams.

Discussion

These analyses aimed to examine how community opportunities for PYD activities impact youth well-being and mental health. We found that neighbourhoods with higher levels of church group participation, sports team involvement and helping others, were associated with improved well-being of students.

Previous research demonstrates that a small but significant portion of the variance observed in PYD is explained by neighbourhood predictors, notably human resources such as mentors (Theokas & Lerner 2006). Research has consistently demonstrated that the more developmental assets, such as external support, empowerment, boundaries and expectations, and constructive use of time, that young people possess, the more likely they are to become happy, healthy and contributing members of society (Scales et al. 2000, 2006). Conversely, low neighbourhood involvement is associated with increased risk behaviours and poorer overall health (Morgan & Haglund 2009). Our findings expand on previous research, focusing on the opportunities for contribution and participation young people have in their communities (Phelps et al. 2007).

Individuals’ religious importance and church attendance has previously been shown to be protective against depression and suicidal behaviours in adolescents, particularly females (Rasic et al. 2011), and participation in faith-based activities is associated with positive developmental outcomes (Hansen et al. 2003). Likewise, participation in sports has been shown to be protective against depression and suicidal ideation among young people (Babiss & Gangwisch 2009). At a neighbourhood level, the availability of sporting opportunities and resources is correlated with adult residents’ physical activity and well-being (Sooman & Macintyre 1995). This research appears consistent with our finding that students in neighbourhoods with higher levels of church group participation and greater sports team participation have greater well-being.

The current study indicated an association between self-reported well-being of students and opportunities for involvement in altruistic activities, that is, helping others in the community. The neighbourhood variable of helping others included opportunities for volunteering, which on their own did not significantly impact well-being, suggesting a differing association between altruistic behaviours and well-being. Previous research has demonstrated involvement in altruistic behaviours is associated with greater levels of happiness and health in adults (Corral-Verdugo et al. 2011), and to some extent in adolescents (Schwartz et al. 2009). While opportunities for altruism may empower students and help teach empathy, it may also be related to neighbourhoods and communities that value young people’s contribution.

There were no significant associations between opportunities for PYD at the neighbourhood level with suicidality or depression symptoms among students. It is unclear why no associations were found for these mental health concerns, whereas we found a significant association between opportunities for PYD at the neighbourhood level and student well-being. For depression and suicidality, neighbourhood opportunities for PYD may be less relevant than more immediate factors such as family support and functioning (Kearns et al. 2012). These findings reinforce that well-being is a unique and significant construct to be examined in young people, negating a deficit-focused perspective. From a PYD perspective, well-being is not just the absence of mental health concerns, and for this measure of young peoples’ health, opportunities outside the home may be more critical.

Further analysis was performed to predict the effects of higher levels of church group and sports team participation and helping others on well-being scores. While these models showed small effects at the individual level, these findings have important implications at a population level, where net effects of community interventions may have a substantive impact.

While the significant predictors of well-being (i.e. neighbourhood levels of helping others, church groups and sports teams involvement) identified in our analyses were noted to be well-accessed by Māori and Pacific students, we found that Pacific young people had the lowest rates of volunteering and employment and Māori, Asian and Other ethnicity students had low levels of employment, sports team and volunteering participation compared to European students. This may reflect socio-economic disparities, as there is evidence that participation in extracurricular activities is influenced by socio-economic factors (White & Gager 2007).

Community opportunities for PYD may therefore act as a limitation as well as an asset. For example, growing up in low-socio-economic status communities may constrain a young person’s ability to participate in extracurricular activities due to diminished local resources and social capital. Conversely, the opportunities in their neighbourhood for PYD may act as a buffer for youth whose family environments are constrained in terms of developmental assets. Future analyses should examine whether opportunities for PYD can buffer the effects of socio-economic deprivation and discrimination.

Limitations

The Youth’12 data is cross-sectional and observational, so it is not possible to establish causal order between potential predictors and outcomes. It may be that the neighbourhood effects are a result of selection effects rather than causation (van Ham & Manley 2010). Selection effects have been demonstrated in outcomes such as employment, whereby residential selection is constrained by socio-economic factors which in turn influence employment prospects. Longitudinal data would help elucidate this relationship.

A further limitation is that we use self-report data from the student on both the dependent and independent variables. Students who are experiencing distress may withdraw from neighbourhood activities and therefore perceive fewer positive development opportunities. This possibility of correlated measurement error is somewhat mitigated by aggregating to the neighbourhood level but remains a significant limitation. Further studies could consider using split samples or data from independent sources to measure neighbourhood PYD opportunities.

These analyses have presented neighbourhood effects at the CAU level only, defining neighbourhoods as areas of around 3000–5000 residents. It is difficult to define a spatial area with appropriate respect to social networks and interaction. Further, neighbourhood effects may be confounded by travel to schools and other destinations. Further research on appropriate area units that fit with young people’s perspectives of their neighbourhoods would help clarify these concerns.

Lastly, this study population does not include youth who have permanently left secondary education or who attend infrequently. This group of youth are known to be at higher risk of poor emotional and behavioural concerns and their omission is likely to attenuate our findings (Clark et al. 2010).

Conclusion

We found that communities that provide opportunities for PYD are associated with better well-being among young people residing in those neighbourhoods. Community-level activities and supports for young people through existing institutions such as churches and sports clubs are important community assets for young people. Furthermore, opportunities for altruistic activities, such as helping others should be encouraged. Our findings support local neighbourhood initiatives to provide opportunities for PYD for young people to enhance their well-being.

Acknowledgement

The authors acknowledge Toshiba (Australia) Pty. Limited.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Youth’12 was funded by the Ministries of Youth Development, Social Development, Health, Education, Justice; the Department of Labour, Families Commission; and the Alcohol Advisory Council.