Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

APPDEV-00774; No of Pages 9

Contents lists available at ScienceDirect

Journal of Applied Developmental Psychology

Alignment of perceived needs across levels of a community

Jonathan F. Zaff a,⁎, Elizabeth Pufall Jones b, Katie Aasland b, Alice E. Donlan b, Emily S. Lin b,Jennifer Elise Prescott b, Alexandra Baker b

a America's Promise Allianceb Tufts University

⁎ Corresponding author.E-mail address: jonz@americaspromise.org (J.F. Zaff).

http://dx.doi.org/10.1016/j.appdev.2015.02.0020193-3973/© 2015 Published by Elsevier Inc.

Please cite this article as: Zaff, J.F., et al., Align(2015), http://dx.doi.org/10.1016/j.appdev.2

a b s t r a c t

a r t i c l e i n f o

Available online xxxx

Keywords:Comprehensive community initiativesAlignmentCommunity needsFamily needs

Youth-focused comprehensive community initiatives (CCIs) exist to create the conditions so all young peoplemay have the developmental supports needed to thrive. Research shows alignment is a key ingredient formeaningful change in a community. The current study discusses the theoretical basis for the importance ofalignment, and provides a method to measure alignment of perceived needs in the community using semi-structured interview data. Our results suggest a method of using the perceptions of multiple stakeholders toreveal that there are alignments and misalignments across the levels of a community. Direct service providers(DSP) and families had the most alignment, while the least alignment was between the CCI leadership andfamilies. Further, DSP and families stressed basic needs (such as needing to pay bills and buy food, or needingtransportation), while CCI leadership stressed the need for the creation and/or implementation of academicprogrammatic efforts to ensure positive developmental outcomes.

© 2015 Published by Elsevier Inc.

Comprehensive community initiatives (CCIs) have been promotedover the past 30 years as entities that can bring together sectors and in-dividuals to address complex problems that are believed to be immuneto single interventions (e.g., Wolff, 2001). A key assumption of CCIs isthat all facets of the community are interconnected and important toachieving positive change within the community (Kubisch, 2010).CCIs assess, design, and implement policies and programs that leveragethe capacity of the community, and have traditionally targeted issuesrelated to public health, social welfare, housing, education, and commu-nity development (Kubisch, 2010). In particular, CCIs focused on youngpeople are positioned to coordinate efforts, resources, and fundingacross organizations to cultivate human, institutional, and social capital(together considered community capacity) that a community wouldneed to resolve the needs of their young people (Chaskin, 2001).Community capacity has been shown to improve the available develop-mental supports and positive developmental outcomes for youngpeople (Leventhal & Brooks-Gunn, 2003). Unfortunately, relativelyfew youth-focused CCIs have been effective at attaining community-level impacts (see Foster-Fishman, Berkowitz, Lounsbury, Jacobson, &Allen, 2001; Kubisch, 2010; Roussos & Fawcett, 2000 for reviews).

However, community-level impacts have been found when CCIs arealigned across the levels of a community, with the vision and program-matic strategies of each level aligning with the strategies and imple-mentation of actions and programs; and in turn, those actions and

ment of perceived needs acro015.02.002

programs aligning with the children and youth who are the focus ofthe efforts (e.g., Auspos, 2010; Hawkins et al., 2008; Spoth, Guyll,Redmond, Greenberg, & Feinberg, 2011). In this paper, we explorethe degree towhich three CCI's visions and perceived needs of the com-munity are alignedwith the needs identified by direct service providers,and youth and their families in the communities. To complement andextend the current literature, we present an in-depth examination ofalignment of vision and perceived needs across levels of a communitysystem. In our introduction, we discuss theory that supports the notionthat misalignment creates roadblocks for success whereas alignmentfacilitates opportunities. Then we use cross-case qualitative analyses(Miles & Huberman, 1994) to explore alignment and misalignmentwithin three communities, and finally propose ways to improve align-ment within CCIs.

Consistent with Relational Developmental Systems Theories thatemphasize the importance of person ← →multiple-context relations(RDST; Bronfenbrenner & Morris, 2006; Lerner, 2012; Overton, 2013),alignment within and across levels of a community system shouldlead to a higher likelihood of achieving positive developmental out-comes (Brandtstädter, 1998; Zaff & Smerdon, 2009). However, CCIs donot often meet the ideal of alignment within and across levels, and in-stead show a lack of alignment or even misalignment across the levelsof the community (e.g., among decision-makers, practitioners, and thefamilies and youth at the center of the CCI's work; Auspos, 2010).Thus, we propose that one reason for modest or null effects of CCIs onyouth outcomes is a lack of alignment. For the purposes of this study,we define alignment as sharing the same or complementary perceived

ss levels of a community, Journal of Applied Developmental Psychology

2 J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

community needs and how these needs will be met, across multiplelevels of the community (e.g., community leaders, business leaders,parents, and youth).

The link between alignment of vision and action within and acrosslevels of a community and impact has been explicit (e.g., Auspos,2010; Nowell, 2010; White & Wehlage, 1995) or implicit in numerousstudies of CCIs (e.g., Fagan, Hanson, Hawkins, & Arthur, 2009).Auspos (2010) has discussed alignment in vision among members ofcollaborative units in numerous CCIs, and White and Wehlage (1995)found that there is often a lack of alignment between the decisionsbeingmade by the collaborative unit and the implementers of those de-cisions. In particular, Nowell (2010) found that misalignment acrosscommunity collaborative stakeholders negatively impacted acollaboration's efforts to combat domestic violence. However, little re-search has been conducted on how to assess alignment. Therefore, wepresent an in-depth example of one way to assess alignment withinthree different communities.

Our proposition that alignment is an essential ingredient foreffective CCIs is based on RDST (e.g., Overton, 2013), previous researchand evaluations of CCIs (e.g., Auspos, 2010; Nowell, 2010), and frame-works that have been developed for the functioning of CCIs(e.g., Foster-Fishman et al., 2001; Lasker, Weiss, & Miller, 2001). Westart with the recognition that development is defined as the relationbetween an individual and her or his context (Lerner, 2012) and that in-dividuals are embedded within multiple contexts or layers within theirecology (Bronfenbrenner & Morris, 2006). Youths' ecologies extendfrom proximal relationships with other people (e.g., parents, teachers,youth workers, other adults in their communities), and organizationswithin which young people learn and grow (e.g., schools, faith-basedinstitutions, youth development organizations), to more distal factors,such as public policies, economic conditions, and cultural norms(Bronfenbrenner & Morris, 2006). Importantly, consistent with the te-nets of RDST, all of these factors are part of an interconnected system,such that each piece influences other pieces and simultaneously influ-ences the individual. In turn, the individual also exerts influence onthe layers of the ecology that surrounds her (Bronfenbrenner &Morris, 2006).

When applied to the lived experiences of youth,we call this system ayouth system. When there is alignment among the contexts within thesystem and the individual strengths and needs of the individual (anideal which we call a supportive youth system), the hypothesized resultsare adaptive developmental regulations for the individual and for theecology around the individual (Brandtstädter, 1998); that is, a support-ive youth system leading to benefits for the young person as well asbenefits for the surrounding community. Extending this idea one stepfurther, we propose that alignment within and across levels of theeco-system will facilitate the optimization of the youth system(Bronfenbrenner & Morris, 2006).

Alignment within a given level of a system is necessary but not suf-ficient for alignment across the entire system. For example, research hascontinually shown that it is important to have member agreement andalignment on vision and how to reach that vision within a collaborativebody (e.g., Auspos, 2010). However, less work has focused on theimportance of alignment across levels of the system (e.g., shared visionand goals among collaborative members, direct service providers, andfamilies).

Two CCI models provide illustrative examples of how to encouragecross-level alignment: Communities that Care (CTC) and PROSPER.The CTC community change system is a community collaborativemodel that focuses on identifying community needs, aligning relation-ships and contexts, choosing proven prevention programs, andimplementing those programs with fidelity (Hawkins, Catalano, &Arthur, 2002). In addition to on-the-ground efforts to create alignmentacross contexts, the CTC process has explicit protocols in place to em-phasize alignment. For instance, a core component of the CTC processis that key partners learn about (and ultimately internalize) the social

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

development model. This model informs the community's selection andimplementation of prevention programs and provides a frame throughwhich to consider the creation of supportive conditions for youth in thecommunity. In addition, the CTC staff conducts a needs assessment thatis used to target the needs of the youth in the community. That is,instead of assuming that the council knows intuitively what issuesyouth in the community confront, the council surveys a representativesample of youth and asks them directly. CTC also explicitly encouragesalignment across contexts through facilitating and emphasizing strongschool-community relationships (Fagan et al., 2009). These core CTCcomponents (coupled with choosing “proven” programs that addresscommunity needs and implementing the programs with fidelity) haveresulted in significant and substantive impacts on reductions inincidence of alcohol, cigarette, smokeless tobacco, and overall acts ofdelinquency (Hawkins et al., 2008).

The mechanisms for alignment for CTC are similar in the PROSPERmodel, a university–researcher–community partnership model de-signed to facilitate the effective delivery of prevention and interventionprograms in schools and within families. A core component of thismodel is the Prevention Coordinator, who aligns the work of a commu-nity council with the needs of the community. A long-term impact studyof PROSPER has shown substantive impacts on child, parent, and familyoutcomes, including drug and alcohol use, among others (Redmondet al., 2009). Otherswhohave studied single-site CCIs have documentedsimilar processes for understanding the needs (and strengths) of a com-munity, developing strategies, and obtaining buy-in from program pro-viders, practitioners, parents, and young people (e.g., Anderson-Butcheret al., 2008; Bringle, Officer, Grim, & Hatcher, 2009).

These examples of successful collaborations impacting youthoutcomes suggest alignment is integral. Researchers can also lookto examples of unsuccessful initiatives to learn about the dangersof misalignment. For example, in the late 1980s and early 1990sthe Annie E. Casey Foundation funded the New Futures Initiative, afive-city effort designed to improve the conditions within a commu-nity and support the community's young people. However, the ini-tiative did not result in positive outcomes for youth, and teenpregnancy rates (a primary focus of the effort) increased during theprogram's tenure. In addition, it is questionable whether any struc-tural changes (e.g., to policy or organizational capacity) implement-ed by the program were sustained over time (Annie E. CaseyFoundation, 1995). White and Wehlage (1995) concluded thatthere was misalignment between the vision of policymakers andthe experiences of on-the-ground practitioners who worked withyoung people every day. They posited that desired change wouldoccur if areas of misalignments became aligned. For example, a cityagency that incorporates the lived experience of the youth in theircommunity into policy changes would be more likely to have an im-pact on youth (White & Wehlage, 1995). In the case of The New Fu-tures Initiative, those making the decisions and planning solutionstended to be senior-level officials in agencies and organizations, in-stead of a group of individuals who represented all aspects of acommunity.

We propose that a major reason for the lack of youth-level effectsfrom most collaborative work is the misalignment across levels of theCCI. Specifically, we suggest that a lack of alignment across collaborativeentities, agencies, practitioners (e.g., executive-level and direct serviceproviders), youth, and their familieswill limit the impact on communitycapacity and on social outcomes. The focus of the present study is to de-scribe amethod for assessing alignment of perceived needs across levelsof a community, including the CCI, those working directly with youth(i.e., direct service providers; DSP), and the youth and their familiesthemselves. We focus on CCIs that have already been formed, have thesame goal (all youth graduating from high school ready for collegeand/or career), and have agreed, as a community, to work together.We do not assess the connection of alignment to specific outcomes,but rather examine the alignment itself.

ss levels of a community, Journal of Applied Developmental Psychology

3J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

Method

This paper uses data from a larger, longitudinal, mixed-methodsstudy of how communities come together to support young people.Data were collected at three levels of a community using multiplemethods designed to elicit the most information from each level of thecommunity: (1) CCI leadership (two sets of annual key respondentinterviews with members of the CCI leadership); (2) direct service pro-viders (two rounds of focus groups); and (3) youth and their caregivers(three semi-annual interviews with caregiver-youth dyads, in eachcommunity).

Participants

Participants were drawn from a longitudinal study focused on threeurban communities that have an existing CCI focused on increasing theeducational attainment of their children and youth. The communitiesare similar in terms of having a high rate of poverty and a large racialand ethnic minority population. However, there were differences(which we describe below) across the communities in terms of theirdemographic composition, as well as how each defined their area ofintervention. Similar methods for recruitment were used in each com-munity, with each level of the CCI requiring a different method usedfor recruitment (i.e., CCI leadership, direct service providers and Youthand Caregiver dyads). Here, we describe each community context andthe participants from that community.

Community ACommunity A is a mid-sized southern city, where the CCI defined a

120 contiguous block target area for intervention. Within these 120blocks, 51.3% of families live below the federal poverty level, and 69%are families with children 18 years-old or younger living below thepoverty level (United States Census Bureau, 2010; see Table 1). In thiscommunity, 60.3% are unemployed or not in the labor force, and 31.5%are only high school graduates or equivalent (United States CensusBureau, 2010). In this city as a whole, violent crime is 725.9 instancesper 100,000 people in the city in 2012 (United States Department ofJustice, 2012). The majority of youth in the schools located within this

Table 1Community demographics.

Demographics Community

A B C

Personal Familystructure

Single male withown children

8.10% 1.90% 3.00%

Single female withown children

16.50% 11.40% 16.20%

Education Less than 9th grade 18.60% 5.00% 11.50%No diploma (9–12) 20.40% 10.30% 8.40%High schoolgraduate/equivalent

31.50% 25.10% 25.70%

Place of birth Foreign born 12.70% 5.60% 38.00%Language English only 71.80% 90.20% 42.80%

School Sex Male 50.03% 51.04% 49.87%Female 49.97% 49.72% 50.13%

Race/ethnicity Hispanic 27.30% 10.11% 45.03%Black 65.01% 73.07% 13.81%White 5.03% 18.18% 32.07%Other 2.66% 5.37% 9.09%

Free and reduced-price luncheligible

76.07% 80.23% 74.09%

Economic Income b poverty level 51.30% 21.80% 22.70%With children under18 years

69.00% 32.90% 35.50%

Employment Population16+

Unemployed 12.40% 7.30% 5.30%Not in labor force 47.90% 37.40% 27.50%

ViolentCrime

Per 100,000 2010 706.90 754.20 324.702011 738.20 792.00 370.002012 725.90 815.20 348.40

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

120 block area that the CCI targets are AfricanAmerican, and themajorityof the student population receives free or reduced-price lunch (between68.4% and 89.5%) (National Center for Education Statistics, 2012).

In Community A, we interviewed 18 caregiver and youth dyads; themajority of caregivers (n = 17) and youth (n = 10) were female (seeTable 2). The mean caregiver age was 42.12 (SD = 11.23) and themean age for youth was 12.37 (SD = .92) at the time of our first inter-view. Fifteen families described themselves as Black and the other threeLatino. Dyads were recruited using a variety of methods including post-ing flyers within the 120-block area and outreach from CCI members.Through our relationship and regular meetings with the CCI leadership,we were able to conduct interviews with 7 members of the leadershipand focus groups with 36 DSPs.

Community BCommunity B is a large urban city in the south. The CCI in this com-

munity focuses on the entire city as its targeted area of intervention. Inthis city, 21.8% of families live below the federal poverty level, and 32.9%are families with children 18 years-old or younger living below the pov-erty level (United States Census Bureau, 2010; see Table 1). In this city44.8% are unemployed or not in the labor force, and 25.1% are onlyhigh school graduates or equivalent (United States Census Bureau,2010). Further, violent crime was 815.2 instances per 100,000 peoplein the city during 2012 (United States Department of Justice, 2012).The majority of youth in the schools in this city are African American(73.07%), and the majority of the student population receives freeor reduced-priced lunch (80.23%) (National Center for EducationStatistics, 2012).

In Community B, we interviewed 15 caregiver and youth dyads; themajority of caregivers (n = 14) and youth (n = 9) were female (seeTable 2). The mean caregiver age was 40 (SD = 6.47) and the meanage for youth was 12.1 (SD = 1.06) at the time of our first interview.The majority of families (n = 13) described themselves as Black.Dyads were recruited using a variety of methods including postingflyers in schools and through the outreach of CCI members. Throughour relationship and regular meetings with the CCI leadership, wewere able to conduct interviews with 10 members of the leadershipand focus groups with 31 DSPs.

Community CCommunity C is a small urban community in the northeast. The CCI

chose a specific public housing development within the city as theirtargeted intervention area. This housing development lies within a cen-sus tract where 22.7% of families live below the federal poverty level,and 35.5% are families with children 18 years-old or younger (UnitedStates Census Bureau, 2010; see Table 1). In this area 32.8% are unem-ployed or not in the labor force, and only 25.7% are high school gradu-ates or equivalent (United States Census Bureau, 2010). In this censustract area, only 42.8% of the population use English as the primarylanguage at home (United States Census Bureau, 2010). Further, violentcrime is 348.4 instances per 100,000 people in the city overall during2012 (United States Department of Justice, 2012). The majority ofyouth attending schools central to the housing development areHispanic (45.03%), followed by White (32.07%), and Black (13.81%).The majority of the student population received free or reduced-priced lunch (74.09%).

In Community C, over the course of the longitudinal study weinterviewed 19 caregiver and youth dyads; all of the caregivers were fe-male and the majority of youth were male (n = 9) (see Table 2). Themean caregiver age was 41.64 (SD = 6.97) and the mean age foryouth was 12.54 (SD = 1.50) at the time of our first interview. Six fam-ilies described themselves as Black, 7 as Latino, and 5 as White. Dyadswere recruited using a variety of methods including posting flyers inschools and throughout the housing development, CCI member out-reach, and with assistance from community informants. Through ourrelationship, and regular meetings with the CCI leadership, we were

ss levels of a community, Journal of Applied Developmental Psychology

Table 2Participant demographics.

Participant type Demographic Community

A B C

Youth/caregiver dyads Sex (F/M) Youth 18 15 1910/8 9/6 8/9

Caregiver 17/1 14/1 19/0Age (SD) Youth 12.37 (.92) 12.1 (1.06) 12.54 (1.5)

Caregiver 42.12 (11.23) 40 (6.47) 41.64 (6.97)Marital STATUS Married 2 7 7

Single 4 5 3Divorced 1 3 1Cohabitation 1 – 1Widowed 1 – 1Separated – – 1

Education High school or less 4 2 6Some college 4 4 3Associates degree – 1 1Technical school – – 1Bachelors degree – 3 2Graduate school – 5 –

Employment Employed 4 11 4Unemployed — looking 2 3 2Unemployed — not looking 1 1 8Retired 1 – –

Housing status Rent 7 7 19Own 1 8 –

Race/ethnicity African-American/Black 15 13 6Latino/Hispanic 3 1 7White – – 5Other – 1 1

Grade 5th – – 26th 14 5 37th 3 6 88th 1 4 39th – 1 –

Collaborative leadershipa 7 10 8Direct service providersa 36 31 28

a Demographic data were not collected for collaborative leadership or direct service providers.

4 J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

able to recruit and conduct interviewswith 8members of the leadershipand focus groups with 28 DSPs.

Materials

To examine alignment across levels of the CCIs, and to exploresimilarities and differences in alignment among the respective CCIs,we conducted a cross-case analysis (Miles & Huberman, 1994). Forthis analysis we completed meta-matrices (Miles & Huberman, 1994),which are complex visualizations of qualitative data that allow aresearcher to infer meaning and analyze themes. To complete thesemeta-matrices enabling cross-case comparisons, we condensed data(Miles &Huberman, 1994) froma variety ofmeasures across the variouslevels of the CCI (i.e., the CCI leadership, direct service provider, andyouth and caregivers). Below we describe the measures we used tocollect the data, the procedures we used to collect data, and the specificselection of data we used for our current analysis.

CCI leadership interviewThe questions for the key respondent interviews with members of

the CCI leadership focused on their perceptions and attitudes aboutthe youth and families they serve, their organization, and the structuresand processes of the CCI. The interviewswere semi-structured, sowhilethere was a list of questions developed to understand these aspects ofthe CCI, interviewers were given the flexibility of asking follow upprobes, or skipping over questions, dependent upon the respondent'sanswers.

Direct service provider (DSP) Focus GroupsThe DSP focus group questions explored the direct service provider's

experiences and work with families in their service community.

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

Questions focused on their perceptions and attitudes about the youthand families they serve, the structures and processes at play in theirvarious organizations, their specific work with youth and families,their perceived impact of this work, and what they believed was neces-sary to truly have an impact on the lives of families in their service com-munity. We also asked questions so that we could understand theirknowledge of the CCI in the community and the impact that the CCIhad on their direct service work.

Youth and caregiver interviewsThe focus of these interviews was to assess youth and caregivers'

abilities to navigate and negotiate their communities to access theopportunities and resources they need to thrive academically, socially,and civically. The interviews were semi-structured so that while wedeveloped a set group of questions to understand these aspects oftheir life experiences, interviewers were given the flexibility of askingfollow up probes, or skipping over questions, dependent upon therespondent's answers.

Procedure

Procedure for CCI leadership interviewsSemi-structured individual interviews were conducted with execu-

tivemembers of the CCI boards/leadership committees in the respectivecommunities at the beginning of the first and second year of our three-year longitudinal study (Winter 2011 and Winter 2012 respectively).Members of the boards were recruited through the lead conveners,andwe had samples of at least four boardmembers in each community.The size and composition of the boards changed over the course of theyear, and as a result we were not always able to follow-up with thesame board members. However, the sample was refreshed with new

ss levels of a community, Journal of Applied Developmental Psychology

1 The coding scheme is available from the corresponding author by request.

5J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

members of the board (see Table 2 for sample sizes). Interviews lasted45min to 1 h andwere conducted in a private setting of theparticipant'schoosing.

For the current analysis, we engaged in an iterative coding processusing the data from both executive member interviews with regards towhat they identified as the needs and concerns of families in the commu-nity, and how and if the CCI was working to address the needs and con-cerns in the community. Given the iterative nature of the coding process,we describe the coding process fully in our analysis and results section.

Procedure for direct service provider focus groupsThrough a partnershipwith the lead convener in each of the commu-

nities,we identified a set of organizations thatwould bemost relevant tohave their staff represented in the focus groups (i.e. the organizationshad some form of affiliationwith the CCI in the community). Once iden-tified, we emailed the directors at each of the organizations indicatingthe purpose of the study, and asking them to nominate at least three ex-emplar DSPs from their organization to participate. We then contactedthese DSPs and asked them to participate in a 45 minute focus groupto discuss their work with families in their respective communities.Focus groups were conducted in public venues that were often in a cen-tral location and offered convenient parking and accessibility to publictransportation (e.g., a school or community center located in the CCIfootprint). Focus groups were audio taped, and the audio files werethen coded in Atlas.ti 6. Focus groups were conducted twice during thecourse of our three-year study, at the beginning of the first year of thestudy (Winter 2011) and the beginning of the second year in the study(Winter 2012).

For the current analysis, we engaged in an iterative coding processusing the data from both focus groups with regards to the needs andconcerns of families in the community as identified by DSPs. The itera-tive coding process is explained in our analysis.

Procedure for youth and caregiver interviewsInterviews were conducted semi-annually for three years (Spring of

2011–Spring of 2014) by graduate students for a total of six interviewsper family. Whenever possible, the same graduate student conductedthe interviews with the same caregiver/youth dyads in subsequentrounds, so that a rapport could be built between the participants andthe researcher. Youth and caregiver interviews were conducted sepa-rately, and in private whenever possible. Interviews were conductedin the participants' home, or any other locationwhere they felt comfort-able and able to speak candidly about their life experiences. Youth inter-views typically lasted 30 to 45 min, and caregiver interviews typicallylasted 45 min to 1 h. All interviews were audio recorded and thentranscribed for coding.

For the current analysis, we engaged in an iterative coding processusing rounds 1, 2, and 3 interviews (i.e., the Spring of 2011–Winter2012, which was the same time period during which the collaborativeleadership interviews and DSP focus groupswere collected,) examiningthe needs and concerns identified by the dyads. The iterative codingprocess is explained in our analysis.

Analysis and results

We conducted a cross-case analysis (Miles & Huberman, 1994) toexamine alignment across levels of the CCIs, and to explore similaritiesand differences between the respective CCIs. For this analysis, each levelof the CCI was considered a case (e.g., CCI leadership in Community A,DSP Community A, and Families Community A), and the variables ana-lyzed were the needs of the community identified at each level(e.g., safety). To explore commonalities across levels within each CCI,as well as across levels between the CCIs, we developed meta-matrices allowing us to stack cases so that we could look for commonthemes among the needs variables. We then partitioned (by exact

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

need) and clustered (type of need and by community) the data regard-ing participants' perceived needs (Miles & Huberman, 1994).

Alignment of identified needs

We used an iterative process for data analysis whereby codingcontinually informed our analysis and findings; each iteration of codinginformedhowwepartitioned and clustered our codes for further analysisand the refinement of our findings. Each round of data (CCI leadershipinterviews, and DSP focus groups, and youth and caregiver interviews)was coded and analyzed by one post-doctoral fellow and five graduatestudents. Although our initial analyses focused on emerging themeswithin each level of the CCI, we noticed several prevalent themes thatillustrated issues of alignment among the different levels.

For example, in these initial stages of our analysis, we noticed veryfew DSPs knew about the CCIs in their communities. Looking closer,we noticed differing theories of change across levels of the systems(e.g., therewas no shared vision forwhat different actors in the commu-nity wanted for their young people). Indeed, the theory of change foreach participant group was informed by their varying backgrounds,and by what they thought would be best for accomplishing levers ofsuccess. Further, these conversations inevitably revolved around theperceived pervasive needs in the communities (e.g., financial assis-tance). Thus, we returned to our data in order to partition the needs ofcommunity members identified by our participants, and then clusterthem by types, looking at how and if the perceived needs at each levelof the CCI aligned with the others.

Given that coding schemes and interview structures for these initialanalyses varied across our participant groups, we developed a consis-tent method by which to extract needs from the interviews. The post-doctoral scholar and one graduate student from the research teamreturned to the interviews, coding schemes, codingmemos, andwrittenanalyses, combining and condensing these components to derive a sin-gle structure by which to understand the identified needs of youth andfamilies at each level of the community.1 Using this structure, we con-ducted our analysis examining the needs identified by each level in agiven community for the first 1.5 years of data (i.e., 2 annual interviewswith CCI leadership, 2 annual focus groupswith direct service providers,and 3 semi-annual interviews with youth/caregiver dyads), constantlyreturning to the data to understand how each level (the CCI leadership,direct service providers, and the youth and caregivers) defined and ad-dressed (or proposed to address) the need.We then tabulated the needsin meta-matrices, which allowed us to stack cases so that we could lookfor common themes among the needs variables, partitioning (by need)and clustering (by type of need and community) the data (Miles &Huberman, 1994) to illustrate how the needs aligned within each com-munity, and if we noticed similarities across communities.

To show how we arrived at the summary of alignment, we providean illustrative case, depicting the overarching needs indicated for Com-munity A, and how each level of the community discussed or definedthe need (see Table 3). Exact needs could be defined in several differentways within any given type. For example, a financial need may be de-fined as a need for money to pay for a basic need, the need for adequateemployment, and/or the need to fund enrichment opportunities forthe youth. For this analysis, each definition of a need was regarded asan instance for possible alignment. Using the example above, financialwas a type of need that was partitioned in the three above-definedways, each of which is considered one instance. For example, instance14 in Community A is agreed upon by all three levels of the CCI, thus in-dicating that the CCI leadership, DSPs and Families all agree that there isa need for a centralized source of information about resources in thecommunity. This process of analysis showed that, overall, our commu-nities lacked alignment, which we believe critical to understandingthe impact of collaborative efforts on child and youth outcomes. After

ss levels of a community, Journal of Applied Developmental Psychology

Table 3Alignment meta-matrix for community A.

Need Instance Level

CCI DSP Caregiver

Parental Involvement 1 Need to educate parents on parenting— parenting training Need to help the parents be able to meet the needs of theirchildren on their own — need to educate the parents so thatthey can break the cycle of poverty

2 One off involvement techniques such as block parties andKaboom builds, community dinners

Need more than one off involvement like just coming to seeyour child in a play

3 Also more sustained efforts such as parent advocates Need parent advocates to help parents navigate the schoolsystem

4 Need more proactive involvement in the school rather thanreactive

Academic 5 A need for early education that isn't siloed from otherprogramming/schooling

Children need the same academic resources and buildingblocks as their peers

6 Need for tutoring — building out tutoring programs bothin and outside of the school

Need tutoring/help after school

Safety 7 Need structurally safe homes8 Need less violence in the community — drive by shootings

Financial 9 No effective/affordable transportation Need transportation to programs (cannot afford/provide owntransportation due to financial hardship)

10 Struggle with meeting the day to day needs – paying theelectrical bill – which inhibits involvement and long termplanning

Need free or help paying for programs because we cannotafford day to day needs much less enrichment opportunities

11 Violence stems from poverty/we need to get rid of povertyEffective Communication 12 Need to build trust among the families and providers

13 Mobility hinders ability to get in touch with youth andfamilies effectively and consistently

14 Need a one stop shop for opportunities — Benefit Bank tofind out about financial, medical and enrichment resources

Need to centralize information about resources at the school Need better communication regardingopportunities/resources in the community

15 Need for interpreters because of language barriersEnrichment 16 A pipeline of services

17 Need a community center18 Need for more playgrounds/safe outdoor recreation space19 Need after school enrichment opportunities like art, music,

sports and dramaFood 20 Need to be able to put food on the table

21 Need to feed the children before you can teach the children22 Need more healthy food options in the community and at school23 School needs to provide food packs for the youth

6J.F.Zaffetal./JournalofA

ppliedDevelopm

entalPsychologyxxx

(2015)xxx–xxx

Pleasecite

thisarticle

as:Zaff,J.F.,etal.,Alignm

entofperceivedneeds

acrosslevels

ofacom

munity,JournalofA

ppliedDevelopm

entalPsychology(2015),http://dx.doi.org/10.1016/j.appdev.2015.02.002

Table 4Frequency of agreement between levels of the community.

Agreement Community Total

A B C

CCI-DSP only 4/23 (17.39) 0/17 (0.00) 1/15 (6.66) 5/55 (9.09)DSP-Caregiver only 2/23 (8.69) 6/17 (35.29) 4/15 (26.66) 12/55 (21.81)CCI-Caregiver only 1/23 (4.35) 2/17 (11.76) 1/15 (6.66) 4/55 (7.27)CCI-DSP-Caregiver 1/23 (4.35) 1/17 (5.88) 0/15 (0.0) 2/55 (3.63)

7J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

developing an overall picture of alignment, we returned to the tables tofurther examine experiences of alignment within each community.

Adding up all of the definitions across the three communities, thereare a total of 55 instances on which levels of the CCI could potentiallyagree (Community A = 23, Community B = 17, and CommunityC = 15). Reading across each individual table indicates where andhow CCI levels agreed on any given need definition (see Table 4).From our analysis, overall we found alignment in 23 out of the totalpossible 55 instances between at least two levels of the community.Two out of the 23 instances were aligned across all three levels (CCI,DSP and Youth-Caregiver). The most alignment occurred between DSPand Caregivers, which might be expected since the DSP in our studywork directly with families in the communities. Agreement was lowestbetween CCI leadership and Caregivers, showing a possible disconnectbetween what families say they need to support their children andwhat the CCI eventually implements. Even if the CCI leadership focuseson a need that they uncover through careful analysis of the community,they might not have buy-in from Caregivers; thus reducing the likeli-hood that families will engage with the CCI-led initiatives.

We found that Community B had the greatest alignment, with atleast two levels agreeing on 9 out of a possible 17 instances of need def-initions (see Table 4). This agreement occurred primarily between theDSPs and Caregivers (n = 6) and secondarily between the CCI leader-ship and Caregivers (n = 2); these included one instance where aneedwas defined in the samemanner across all three levels of the com-munity (i.e., parents need to be involved in the education process to en-sure academic success). This was also the only instancewhere the CCI inCommunity B agreed with how the DSPs defined a need. Alternatively,while Community A had 8 agreements out of a possible 23 instances,the majority of agreements were between CCI executives and DSPs.

From our analysis, we also were able to see which level of the com-munity had the highest frequency of agreement with the other levels.Across the communities, we found that the DSP (19 out of 55 instances)and Caregiver levels (18 out of 55 instances) had more agreementwith another level than the CCI leadership (11 out of 55). Frequencyof agreement also varied by community. Community B had agreementamong more than half of the instances (9 out of 17), while CommunityC (6 out of 15) and Community A (8 out of 23) had rates of agreementthat were at 40% or below. We also found that 40% of needs expressedwere unique needs; that is, needs that were expressed by only one ofthe levels of the community (see Table 5). The DSP level had the highestrate of unique needs, with fully one half of their expressed needs beingunique. Caregivers and CCI leadership not only expressed fewer needsoverall, but also expressed a lower rate of unique needs.

Further, the ways needs were discussed varied by level. DSPs andFamilies were aligned in how they spoke regarding needs, focusing

Table 5Frequency (%) of “unique” needs by party.

Party Community Total

A B C

CCI 1/7 (14.28) 2/5 (40.00) 3/5 (60.00) 6/17 (35.29)DSP 11/18 (61.11) 5/12 (41.66) 3/8 (37.50) 19/38 (50.00)Caregiver 3/7 (42.85) 1/10 (10.00) 3/8 (37.50) 7/25 (28.00)Total 15/32 (46.87) 8/27 (29.62) 9/21 (42.85) 32/80 (40.00)

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

primarily on the daily challenges that youth face and act as barriers tosuccess (e.g., not having a consistent caring adult in the home, notenough food on the table, violence outside their front door, and parentsunemployed, incarcerated, or having to work multiple jobs in order tomake endsmeet). On the other hand, CCI leadershipmembers discussedthe structures the CCI needs, or the actions they need to take, in order tomeet the needs of youth and families (e.g., how the CCI struggles tobuild the internal structures necessary to facilitate collaborationamong members).

For example, in Communities A and C, both the DSPs and the Fami-lies discussed how there is a need for affordable transportation intheir cities. Both indicated that if youth and parents had reliablemodes of transportation theywould be able to engage inmore of the ac-tivities and afterschool programs in the community. Both Caregiversand DSPs indicated that transportation would facilitate parents acquir-ing and retaining employment. However, facilitating or funding trans-portation in the community was not discussed by any of the CCIleadership in these two communities. Similarly, CCI leadership in bothCommunity A and C mention a need for “wraparound” services or a“pipeline” of services. However, both the DSPs and families in thesecommunities felt that meeting the basic needs of the families is neces-sary before creating a wraparound service structure.

Discussion

Youth-focused comprehensive community initiatives (CCIs) havethe potential to achieve community-level impacts that individual orga-nizations would not be able to accomplish alone (Lasker &Weiss, 2003;Wolff, 2001). Because they are typically comprised of organizationsfrommultiple sectors within a community, CCIs are positioned to createthe conditions so that all young people have the developmental sup-ports they need to thrive; whether academically, socioemotionally, vo-cationally, or civically. However, since CCIs were promoted 30 yearsago as the “answer” to social and economic problems in communities,fewCCIs have successfully achieved the impacts that theywere purport-ed to achieve (Auspos, 2010; Foster-Fishman et al., 2001).

One reason for the lack of success stories can be traced to a lack ofalignment within and across levels of a community (e.g., the vision forthe CCI; the needs perceived by the collaborative unit, direct serviceproviders, and families; the actions that are conceived, implemented,and experienced). Studies of successful (Hawkins et al., 2008;Redmond et al., 2009) and unsuccessful CCIs (Auspos, 2010; White &Wehlage, 1995) have found alignment to be a key ingredient. Althoughsome insights into the expression of alignment have been derived fromprevious studies (e.g., Fagan et al., 2009), much still needs to be learnedaboutwhere alignments andmisalignments can occurwithin a commu-nity. The current study was designed to look deeply into the inevitablenuances in perceived needs for each level of the community (i.e., CCI,direct service providers, and young people and their families) toprovide a method for examining alignment.

Our results suggest that there are alignments and misalignmentsacross the levels of a community collaborative. We found that DSPsand families had the most alignment in perceived needs. Because ofthe scope of their work, DSPs usually interact with young people andtheir families multiple days per week and often are skilled at under-standing what is occurring in their young people's lives (Rhodes,2004). However, youth developmentworkers are not necessarily skilledat navigating the decision-making roles of a CCI (e.g., Borden & Perkins,2006). Therefore, if the collaborative entity does not intentionally createopportunities for DSPs to be a part of the strategic planning anddecision-making processes, their perspectives (and as our results sug-gest, the perspectives of the young people and families who theyserve) will be missing from the collaboration's discussions. The poten-tial of youthwhen they are included in community decisions has prece-dence;when they are considered to be assets and given opportunities tobe included in the workings of the community, young people can create

ss levels of a community, Journal of Applied Developmental Psychology

8 J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

positive change (Checkoway & Richards-Schuster, 2006; Zeldin &Topitzes, 2002). In the case of the three CCIs that were the focus ofthis study, none included DSPs or families in the decision-makingprocess; at least during the first year of their work.

Another way to consider alignment is by the expression of uniqueneeds; that is, needs that are only expressed by one level of the commu-nity. Havingmany uniqueneeds could be an indication ofmisalignment,since these would be needs that no one else across the other levelsrecognizes as important. In addition, one could suppose that an increasein the number of unique needs would be related to an increase in thefrustration of the level of the community with the unique needs.For example, if the DSP continually voice concern about a set of needsto which no one else pays attention, then the DSP would most likelybecome disengaged from the CCI process. There is an indication thatnot including the voices of DSPs and families in the planning mayhave contributed to the misalignment across the levels of the systemsin the three communities. No matter the strategies being developedby the collaborative entity or the programs to be implemented, an out-standing question is whether parents and their children will engagewith programs and services that they do not necessarily believe are insync with what their perceived needs are (e.g., a CCI opening a newtutoring center in the areawhen parentsmight say that they need bettertransportation to access social services).

Limitations and future directions

This study used the first wave of data from a longitudinal study ofCCIs. The current study presents only a snapshot into the work ofyouth-focused CCIs; in this case, CCIs that were in the early stages oftheir existence. Because the current study is not longitudinal, we wereunable to assess whether alignment persisted over time. Understandingwhether there was more cross-level alignment over time and examin-ing why there was more (or less) alignment would be an importantnext step. Initial analyses of Wave 2 from our study suggested that thecollaborative entities were being more responsive to the perceivedneeds of community members.2 The reason for this responsiveness ap-peared to be the result of the CCIs gaining their footing over the courseof a year, having a better system in place to elicit feedback from thecommunity, and simply being based in the community for anotheryear. This preliminary work suggests longitudinal analyses would pro-vide insights into how CCIs develop.

In addition, althoughwewere able to assess perceived needs at eachlevel of the community, our sampling did not include every member ofthe collaborative entity, nor a representative sample of DSPs (we reliedon nominations by the executive directors of the organizations fromwhere the DSPs came). Considering the diverse paths of DSPs, includingmore than full-time, full-time, part-time, and volunteers, and the diver-sity of roles that DSPs have within an organization, it is difficult toconstruct a representative sample. We believe that our nominationmethodology, though not perfect, still provides valid perspectivesfrom the DSP level of each community.

Similarly, we erred on depth over breadth with our sample of fami-lies. The families were recruited with assistance from the collaborativeentity and most likely represent the middle of the distribution offamilies livingwithin those communities. That is, the families expresseddifficulties that might be expected for any family, especially familiesliving in an economically disadvantaged community. However, theymost likely did not represent families at the highest risk for their chil-dren experiencing negative outcomes (e.g., poor academic outcomes),nor did they represent families that were exemplars for their ability tothrive. Future research that includes the perspectives of the two tailsof the distribution could provide additional insights into howall familiesperceive their needs and the needs of their children.

2 Analysis available by request from the corresponding author.

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

Finally, assessing the predicted link between alignment andeffecting change on desired developmental outcomes is a rich avenuefor future research. The current study provides a model for assessingalignment, and we hope that future researchers will use this model asa starting place to continue to understand the ways in which alignmentincreases the odds of CCI success over time.

Conclusion

Based on strong developmental theory, we believe that alignment ofvision, goals, and actions across levels of a CCI will improve the effec-tiveness of that CCI, and bring positive change to communities. Howev-er, in our analyses we found that in three established CCIs, there wassignificant room for improvement in alignment across levels. In partic-ular, we suggest that CCI leadership bring more families, DSPs andother community members into their planning sessions to ensuremore perspectives are heard. CCIs have the potential to enactcommunity-level change, but only if they are aligned to the meet thetrue needs of the community.

References

Anderson-Butcher, D., Lawson, H. A., Bean, J., Flaspohler, P., Boone, B., & Kwiatkowski, A.(2008). Community collaboration to improve schools: Introducing a new modelfrom Ohio. Children & Schools, 30(3), 161–172. http://dx.doi.org/10.1093/cs/30.3.161.

Annie E. Casey Foundation (1995). The eye of the storm: Ten years on the front lines of newfutures. Baltimore, MD: Author.

Auspos, P. (2010). Internal alignment in community change efforts. Voices from the field III:Lessons and challenges from two decades of community change efforts. Washington,DC: Aspen Institute, 52–62.

Borden, L. M., & Perkins, D. F. (2006). Community youth development professionals:Providing the necessary supports in the United States. Child & Youth Care Forum,35(2), 101–158. http://dx.doi.org/10.1007/s10566-005-9005-4.

Brandtstädter, J. (1998). Action perspectives on human development. In W. Damon, & R.M. Lerner (Eds.), Handbook of child psychology (5th ed.). Theoretical models of humandevelopment, Vol. 1. (pp. 807–863). New York: Wiley.

Bringle, R. G., Officer, S. D., Grim, J., & Hatcher, J. A. (2009). GeorgeWashington communityhigh school: Analysis of a partnership network. New Directions for Youth Development,122, 41–60. http://dx.doi.org/10.1002/yd.305.

Bronfenbrenner, U., &Morris, P. A. (2006). The bioecological model of human development.In R. M. Lerner, & W. Damon (Eds.), Handbook of child psychology (6th ed.). Theoreticalmodels of human development, Vol. 1. (pp. 79.3–79.828). Hoboken: Wiley.

Chaskin, R. J. (2001). Building community capacity: A definitional framework and casestudies from a comprehensive community initiative. Urban Affairs, 36, 291–323.http://dx.doi.org/10.1177/10780870122184876.

Checkoway, B., & Richards-Schuster, K. (2006). Youth participation for educational reformin low-income communities of color. In P. Noguera, J. Cammarota, & S. Ginwright(Eds.), Beyond resistance! Youth activism and community change: New democraticpossibilities for practice and policy for America's youth (pp. 319–332). London:Routledge.

Fagan, A. A., Hanson, K., Hawkins, J. D., & Arthur, M. W. (2009). Translational researchin action: Implementation of the communities that care prevention system in12 communities. Journal of Community Psychology, 37(7), 809–829. http://dx.doi.org/10.1002/jcop.20332.

Foster-Fishman, P. G., Berkowitz, S. L., Lounsbury, D.W., Jacobson, S., & Allen, N. A. (2001).Building collaborative capacity in community coalitions: A review and integrativeframework. American Journal of Community Psychology, 29(2), 214–261 (Retrievedfrom http://link.springer.com/article/10.1023/A%3A1010378613583).

Hawkins, J. D., Catalano, R. F., & Arthur, M. W. (2002). Promoting science-based preven-tion in communities. Addictive Behaviors, 27(6), 951–976. http://dx.doi.org/10.1016/S0306-4603(02)00298-8.

Hawkins, J. D., Catalano, R. F., Arthur, M. W., Egan, E., Brown, E. C., Abbott, R. D., et al.(2008). Testing communities that care: The rationale, design and behavioral baselineequivalence of the community youth development study. Prevention Science, 9(3),178–190. http://dx.doi.org/10.1007/s11121-008-0092-y.

Kubisch, A. (2010). Overview of community change efforts, 1990–2010. Voices from thefield III: Lessons and challenges from two decades of community change efforts.Washington, DC: Aspen Institute, 8–50.

Lasker, R. D., & Weiss, E. S. (2003). Creating partnership synergy: The critical role ofcommunity stakeholders. Journal of Health and Human Services Administration, 26(1/2),119–139 (Retrieved from http://search.proquest.com/docview/200027026/FBD87FEE969E4EEAPQ/4?accountid=14434).

Lasker, R. D., Weiss, E. S., & Miller, R. (2001). Partnership synergy: A practical frameworkfor studying and strengthening the collaborative advantage.Milbank Quarterly, 79(2),179–205. http://dx.doi.org/10.1111/1468-0009.00203.

Lerner, R. M. (2012). Developmental science: Past, present, and future. International Journalof Developmental Science, 6, 29–36. http://dx.doi.org/10.3233/DEV-2012-12102.

Leventhal, T., & Brooks-Gunn, J. (2003). Children and youth in neighborhood contexts.Current Directions in Psychological Science, 12, 27–31. http://dx.doi.org/10.1111/1467-8721.01216.

Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook.Thousand Oaks, CA: Sage.

ss levels of a community, Journal of Applied Developmental Psychology

9J.F. Zaff et al. / Journal of Applied Developmental Psychology xxx (2015) xxx–xxx

National Center for Education Statistics (2012). Search for public schools. Available fromhttp://nces.ed.gov/ccd/schoolsearch/index.asp

Nowell, B. (2010). Out of synch and unaware? Exploring the effects of problem frame align-ment and discordance in community collaboratives. Journal of Public AdministrationResearch and Theory, 20(1), 91–116. http://dx.doi.org/10.1093/jopart/mup006.

Overton, W. F. (2013). Competence and procedures: Constraints on the development oflogical reasoning. InW. F. Overton (Ed.), Reasoning, necessity, and logic: Developmentalperspectives (pp. 1–30). Psychology Press.

Redmond, C., Spoth, R. L., Shin, C., Schainker, L. M., Greenberg, M. T., & Feinberg, M.(2009). Long-term protective factor outcomes of evidence-based interventionsimplemented by community teams through a community–university partnership.The Journal of Primary Prevention, 30(5), 513–530. http://dx.doi.org/10.1007/s10935-009-0189-5.

Rhodes, J. E. (2004). The critical ingredient: Caring youth–staff relationships inafter-school settings. New Directions for Youth Development, 101, 145–161. http://dx.doi.org/10.1002/yd.75.

Roussos, E. T., & Fawcett, S. B. (2000). A review of collaborative partnerships as a strategyfor improving community health. Annual Review of Public Health, 21, 369–402. http://dx.doi.org/10.1146/annurev.publhealth.21.1.369.

Spoth, R., Guyll, M., Redmond, C., Greenberg, M., & Feinberg, M. (2011). Six-year sustain-ability of evidence-based intervention implementation quality by community–

Please cite this article as: Zaff, J.F., et al., Alignment of perceived needs acro(2015), http://dx.doi.org/10.1016/j.appdev.2015.02.002

university partnerships: The PROSPER study. American Journal of Community Psychol-ogy, 48(3–4), 412–425. http://dx.doi.org/10.1007/s10464-011-9430-5.

United States Census Bureau (2010). Profile of general population and housing character-istics: 2010. Available from http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=DEC_10_DP_DPDP1

United States Department of Justice (2012). Uniform crime reporting statistics. Availablefrom http://www.ucrdatatool.gov/

White, J. A., &Wehlage, G. (1995). Community collaboration: If it is such a good idea, whyis it so hard to do? Educational Evaluation and Policy Analysis, 17(1), 23–38 (StableURL: http://www.jstor.org/stable/1164268).

Wolff, T. (2001). A practitioner's guide to successful coalitions. American Journal ofCommunity Psychology, 29(2), 173–191. http://dx.doi.org/10.1023/A:1010366310857.

Zaff, J. F., & Smerdon, B. (2009). Putting children front and center: Building coordinatedsocial policy for America's children. Applied Developmental Science, 13(3), 105–118.http://dx.doi.org/10.1080/10888690903041469.

Zeldin, S., & Topitzes, D. (2002). Neighborhood experiences, community connection, andpositive beliefs about adolescents among urban adults and youth. Journal ofCommunity Psychology, 30(6), 647–669. http://dx.doi.org/10.1002/jcop.1002.

ss levels of a community, Journal of Applied Developmental Psychology