What Is Community? An Evidence-Based Definition for Participatory Public Health

Abstract

Increased emphasis on community collaboration indicates the need for consensus regarding the definition of community within public health. This study examined whether members of diverse US communities described community in similar ways.

To identify strategies to support community collaboration in HIV vaccine trials, qualitative interviews were conducted with 25 African Americans in Durham, NC; 26 gay men in San Francisco, Calif; 25 injection drug users in Philadelphia, Pa; and 42 HIV vaccine researchers across the United States. Verbatim responses to the question “What does the word community mean to you?” were analyzed. Cluster analysis was used to identify similarities in the way community was described.

A common definition of community emerged as a group of people with diverse characteristics who are linked by social ties, share common perspectives, and engage in joint action in geographical locations or settings. The participants differed in the emphasis they placed on particular elements of the definition. Community was defined similarly but experienced differently by people with diverse backgrounds. These results parallel similar social science findings and confirm the viability of a common definition for participatory public health.

PUBLIC HEALTH PROGRAMS and policy are often defined at regional and national levels, but community is, literally, where prevention and intervention take place. Community context has been identified as an important determinant of health outcomes.1 Recognition of these facts has led to increased calls for community collaboration as an important strategy for successful public health research and programs.2–9 Reviews of the effectiveness of collaborations for improving community health indicate that they can be effective9–11 but that there are many potential obstacles to realizing the benefits of a participatory approach in both public health research and programs.8,11–15 In particular, the lack of an accepted definition of community can result in different collaborators forming contradictory or incompatible assumptions about community and can undermine our ability to evaluate the contribution of community collaborations to achievement of public health objectives.9,16 We are more likely to validate the effectiveness, or ineffectiveness, of collaborative models if we can identify core dimensions of community that have external validity across communities, are consistent with measures that have internal validity within diverse communities, and have predictive value for community-level health outcomes.

Efforts to develop and test HIV vaccines have highlighted social challenges that make community support, trust, and involvement critical for the long-term success of preventive HIV vaccines.17 Project LinCS (Linking Communities and Scientists) was undertaken to identify effective strategies for meeting these challenges. HIV vaccine efficacy trials and, ultimately, immunization programs require national coordination across diverse communities. The primary Project LinCS research objective centered on identifying ways to make efforts such as large-scale efficacy trials and immunization programs as participatory as possible. A related objective was to demonstrate the value added to research when community members actively participate in the design, implementation, and interpretation of the research.

For this commentary, we analyzed data from 1 component of Project LinCS that asked people how they defined community. The interviews were conducted with diverse groups, thus allowing us to look at the potential effect of local and historic experience on the way people defined community and to determine whether a single definition of community could effectively encompass the diversity of local experience.

METHODS

Data Collection

Project LinCS participants included African Americans in Durham, NC; gay men in San Francisco, Calif; injection drug users (IDUs) in Philadelphia, Pa; and HIV vaccine researchers in locations across the United States. The 3 local research communities were selected through a competitive funding process. Through a study design collaboratively developed by researchers and local community advisory boards convened for the project, a total of 76 face-to-face, in-depth, open-ended interviews were conducted in Philadelphia, Durham, and San Francisco to identify broad issues, including how people defined community. The interviews were conducted between December 1995 and May 1996. Another 42 in-depth interviews were conducted between September 1997 and September 1998 via telephone with HIV vaccine researchers across the United States; these interviews included a parallel question on defining community. Study protocols and interview guides were approved by local and Centers for Disease Control and Prevention institutional review boards; written informed consent was obtained from all participants (the vaccine researchers mailed back signed consent forms before their interviews).

Interviews were tape-recorded, transcribed into computer text files, and coded by interview question. Verbatim responses to the following question were then extracted for analysis: “The word ‘community’ means different things to different people. What does the word community mean to you? What is a community?”

Data Analysis

The analysis team developed a structured codebook through a standardized iterative process.18 The extracted text was read and, on the basis of the text content, a set of initial codes and definitions was proposed. None of the codes were specified before the text was read. Two people then independently coded randomly selected text segments from each participant group and compared results. Coding discrepancies were discussed with the analysis team and the codebook was revised accordingly. This process was repeated until the coders reached a satisfactory level of agreement.

To minimize the possibility that relevant text was missed, all text was double-coded. Intercoder agreement for text associated with each participant group was then assessed for each code by κ scores.19 For each code with a κ of less than 0.8, discrepancies in coding were reviewed by this commentary's first author, who then decided whether to apply the code to the text in question. For codes with a k of 0.8 or greater, discrepancies were reviewed by the 2 coders, who then decided which code applications to retain. Text coding and intercoder agreement assessments were done with a developmental version of the software program AnSWR.20

For each participant, numeric listings summarized whether or not each code was applied to that person's text. Numeric matrices were then generated to summarize which codes occurred together in the text of all persons within each participant group. From the matrices, complete-link Johnson's hierarchical clusters were generated in ANTHROPAC21 to identify core elements used to define community. The cluster analysis helped identify similarities in the way people defined community and the extent to which those similarities cut across participant groups.

RESULTS

The 4 participant groups varied in terms of a number of sociodemographic characteristics beyond those immediately reflective of the intentional targeted sampling. Participants who were scientists or from San Francisco tended to have higher levels of education and higher incomes than participants in Philadelphia and Durham. None of the scientists and only 1 of the Philadelphia participants lacked health insurance, although they differed in terms of the primary source of insurance (employer based for scientists and government based for IDUs). Approximately one third of San Francisco and Durham participants lacked health insurance. The mean number of years lived in the current neighborhood of residence was longest for IDUs in Philadelphia (16.5 years), followed by scientists (7.2), African Americans in Durham (5.7), and gay men in San Francisco (2.6).

Of the 118 participants interviewed, 113 provided definitions of community (the question was omitted in interviews with 3 participants in Durham and 2 in San Francisco). Coding identified 17 distinct themes or elements that appeared in the definitions of 2 or more respondents (Table 1 ▶). Hierarchical cluster analysis identified 4 clusters among the themes. A core cluster contained 5 elements: locus, sharing, joint action, social ties, and diversity. Each core element reflected some aspect of face-to-face interaction. A second cluster centered on group-based elements of community: divisiveness, leverage, pluralism, and responsibility. These elements reflected social cohesion and community involvement and often acted as boundary-setting or -maintaining mechanisms. Each element of these 2 clusters was cited by at least 4 members of each participant group.

TABLE 1—

Elements Coded From Definitions of Community From All Sources Within Project LinCS (n = 113)

Definitional Element	Brief Definition	No. of Participants Citing Element (%)
Locus	Physical location; place with people	87 (77)
Sharing	Shared perspective; common interests	65 (58)
Action	Joint action or activities	57 (50)
Ties	Social ties, relationships	56 (50)
Diversity	Differences or diversity (e.g., age, race, income, behavior)	27 (24)
Divisiveness	Fragmentation, division into disputing factions	17 (15)
Leverage	Effectiveness; ability to influence resource availability	17 (15)
Responsibility	Importance of accepting consequences of individual action	17 (15)
Pluralism	Coexistence of 2 or more distinct cultural traditions	15 (13)
Criminality	Impact of criminal activities	12 (11)
Unity	Community fellowship	15 (13)
Drug use	Impact of drug use and addiction	11 (10)
Nonexistent	No community; concept holds no meaning	9 (8)
AIDS	Impact of HIV/AIDS	7 (6)
Services	Availability of social services and programs	6 (5)
Religion	Religious or spiritual focus	5 (4)
Survival	Adaptability, resourcefulness	4 (4)
Other elements		10 (9)

The third and fourth clusters centered on elements that reflected stresses experienced by communities or their members. The first stress cluster included the elements of criminality and drug use. The second centered on the elements of AIDS and unity.

Core Elements of Community

Each of the 5 core elements—locus, sharing, joint action, social ties, and diversity—was cited by 20% or more of respondents (Table 1 ▶). Locus was included in 77% and sharing was included in 58% of all definitions. Both locus and sharing were included in 42% of responses, and at least 1 of the 2 was included in 93% of them. Locus and sharing were each cited alone (i.e., without discussion of other core elements) in 16% of responses. Neither joint action nor diversity was discussed alone, and social ties were discussed alone by 2 people (2%). Sharing and locus were discussed by some participants as alternative ways to define community, while others described them as closely interconnected. Definitions that included all 5 elements were elicited by 8 participants (7%); another 13 definitions (11%) included all of the core elements except diversity. These 2 combinatorial groups account for 27% of all definitions offered (Table 2 ▶). Each of the core elements is described below; the quotations in the box ▶ on p 1933 illustrate how these elements were woven into actual responses.

TABLE 2—

Co-Occurrence of Core Elements in Definitions of Community From All Sources Within Project LinCS

Co-Occurring Elements	No. of Respondents (%) (n = 113)
All 5 (locus, sharing, joint action, social ties diversity)	8 (7)
Any 4	22 (20)
Diversity, locus, sharing, social ties	2 (2)
Joint action, locus, sharing, social ties	13 (12)
Joint action, diversity, sharing, social ties	3 (3)
Joint action, diversity, locus, social ties	2 (2)
Joint action, diversity, locus, sharing	2 (2)
Any 3	26 (23)
Locus, sharing, social ties	5 (4)
Diversity, sharing, social ties	1 (1)
Diversity, locus, sharing	3 (3)
Joint action, locus, social ties	6 (5)
Joint action, locus, sharing	8 (7)
Joint action, diversity, sharing	1 (1)
Joint action, diversity, locus	2 (2)
Any 2	33 (29)
Sharing, social ties	1 (1)
Locus, social ties	11 (10)
Locus, sharing	6 (5)
Diversity, sharing	3 (3)
Joint action, social ties	2 (2)
Joint action, sharing	3 (3)
Joint action, locus	7 (6)
Only 1	20 (18)
Locus	12 (11)
Sharing	6 (5)
Social ties	2 (2)
No core elements cited	4 (4)

Table t3.

Core Elements of Community

DURHAM AFRICAN AMERICAN: I think community can be defined in 2 different ways. There's a community that you define as such because you are forced by where you live, by your upbringing to be around those people. This isn't a voluntary type of thing. This is your community because you live there. The place you call your hometown is your community because you grew up there, you knew people there. You didn't really have a choice as to whether that would be your community or not, it just was. And so you were molded and informed by that surrounding, by that society, but it wasn't a voluntary type. And that's how I would describe my work community. It is my community because I have to work there and it is my workplace. That's not to say that I don't choose at times to include these people in other aspects of my life. But certainly, the people I choose to be sociable with most aren't people I share the same job with.

SCIENTIST: A community is a fairly broad term in my mind that encompasses groups of people working together toward the same goal. . . . I would say I identify with the HIV research community, the HIV care community, my own personal community with my family, certainly my regional community.

INTERVIEWER: Okay, when you say family community, is that just your immediate family?

SCIENTIST: I mean school communities related to my children's schooling, and the communities here in [place name] that are involved in providing support, such as the [AIDS foundation name]. And that's what I mean by community. . . . I mean, there's my neighborhood. That's what I mean by geographical locale. Political community, I suppose I would say that as a Democrat I'm a participant in the political groups here in town and nationally. . . . I think that our communities are less dependent these days on actual physical adjacency, if you will, that the Internet has brought many people together, and when I say I'm part of the AIDS research community, I think of that as a worldwide community.

SAN FRANCISCO GAY MAN: I lived alone for a while after [my partner] died and I really hated it. I really felt very lonely and now I live in a situation with a good friend. . . . There's a sense of comfort in that. Because, you know, he's single and I'm single and we have a group of friends and there's a lot of connection, that we kind of create community. I think we as gay and lesbian people create family too in a lot of ways that are not biological and I think [in] some ways that sense of creating family is creating community, that's what we support ourselves and surround ourselves with.

PHILADELPHIA IDU: Well, in the drug culture, I wouldn't call that a community, you know. I would just call that a part of the community that's just tryin' to survive but [what] community means to me is a way people look out for one another and they do things together, insofar as socializing together, praying together. You know, they have a mutual bond but see, you know, and some of that goes on in the community, you know, and that drug culture can be right here but you still have a group of people that tries to keep the neighborhood together and try to set the right values for the children.

Locus: a sense of place.

Locus encompassed the idea of community as something that could be located and described, denoting a sense of place, locale, or boundaries. One can be “in” a community physically whether or not one identified as being a member of the community. People referred to locus in terms of specific areas (neighborhood, corner, block; street, road, highway; zip code area; village, city, county), with reference to specific settings (home, household; workplace; local taproom or bar, corner grocery store, newsstand, sandwich shop, bookstore; community building, swimming pool, recreation center; church, school), and in terms of general locations (an area or place where people live together; environment or surroundings).

Sharing: common interests and perspectives.

Sharing referred to the existence of shared perspectives and common interests that contributed to a sense of community. Community members were described as sharing the following:

• Values, norms, mind-set, viewpoint, ideology, beliefs, visions

• Passions, obsessions, interests, likes and dislikes, opinions, concerns

• Activities, goals, objectives

• Symbols, jargon

• Skin color, sexual identity

• Tribulations, oppression, repression, history

Sharing contributed to a sense of community through the following:

• Common issues, threads, beliefs, factors

• Being in tune with each other

• Comfort, familiarity, togetherness, identity, recognition

Joint action: a source of cohesion and identity.

Joint action was described as a source of community cohesion and identity. A conscious intent to generate community through action was not viewed as necessary; rather, joint action was seen as leading naturally to the creation of community. Community was described as emerging from the joint actions of people who did the following:

• Socialize, hang out, converse, intermingle, gossip, “shoot the shit”

• Work at the polls, volunteer at the library, run phone banks, train people, work on projects

• Keep people informed about resources, services, and what's happening

• Paint houses; paint the street; push brooms; shovel snow; keep up the area; clean up the block, neighborhood, yard, or house; have block parties

• Give food, share resources, provide for neighbors in need

• Watch over, check up on, look out for, keep an eye on each other

• Set values and goals for the children, have their butts kicked a little bit if they're slacking off, push for the betterment of everyone, do something positive, improve the neighborhood

• Get together, do things together, work together, act together, participate, plan, get things done, get inspired, engage in activities, give input, accomplish goals

• Write, speak, educate, encourage, pray

Social ties: the foundation for community.

Social ties were described in terms of interpersonal relationships that formed the foundation for community. In some instances, such relationships were described as requiring little, if any, effort or ongoing acknowledgment on the part of the individual. The types of relationships cited included the following: family, parents, siblings, cousins; roommates, household; lovers, partners; friends, neighbors, associates, coworkers, acquaintances; role models, support groups.

In addition, participants often described characteristics that they associated with community-based relationships or people. Community, participants said, meant ties with people

• Whom they can trust

• With whom they feel comfortable

• Who care about each other

• With whom they interact, hang out, choose to be sociable, spend time, connect

• Who are known to them

• Whom they always see in the background or around them

• With whom they grew up

Diversity: social complexity within communities.

Diversity emerged in discussions of social complexity (e.g., communities within communities, stratification, interwoven groups, hidden communities, or multiple levels of community). As used here, diversity excluded culturally based ethnic distinctions (see the discussion of pluralism below). Discussions of diversity focused on a larger societal view of community and made reference to differences in interpersonal interaction that resulted from the following:

• Different levels of interaction between people, from the intimate to the superficial

• Demographic and social diversity in the form of race, ethnic origin, socioeconomic status, sexuality, drug use, profession

• The presence of specialized groups that performed needed tasks, such as activists and service providers

• The presence of groups that identified with overlapping or multiple communities

• The presence of groups that were disowned, stigmatized, stereotyped, or distrusted within communities

As seen in Figure 1A ▶, the saliency of each of these elements for the core definition of community varied by participant group. Action, locus, and social ties were the most consistently discussed elements across all groups. Sharing predominated in interviews among scientists and gay men in San Francisco, while diversity was discussed relatively infrequently by all groups except the San Francisco participants. Thus, the relative saliency of sharing and diversity appears to be an important distinguishing characteristic of communities.

FIGURE 1—

Saliency of definitional elements of community by participant group: (A) core elements; (B) group-based elements; (C) stress elements.

Note. San Francisco = gay men in San Francisco, Calif; Philadelphia = injection drug users in Philadelphia, Pa; Durham = African Americans in Durham, NC; HIV vaccine researchers = HIV vaccine researchers across the United States.

Despite group differences in the saliency, frequency, and co-occurrence of the 5 core elements, the overall response pattern indicates that these 5 elements were universally recognized within each community as a whole, if not by every community member. In anthropologic terms, the elements constitute a common cultural domain.22,23 Together, they suggest a full definition of community as a group of people with diverse characteristics who are linked by social ties, share common perspectives, and engage in joint action in geographic locations or settings.

Group-Based Elements of Community

Like the core elements, the group-based elements of community—divisiveness, leverage, pluralism, and responsibility—had meaning across all of the participant groups but were less frequently cited than the core elements (Figure 1B ▶). Overall, at least 1 of the group-based elements was cited by 36 (32%) of the participants, including 18 (75%) of the participants from San Francisco, 8 (32%) from Philadelphia, 4 (18%) from Durham, and 6 (14%) of the scientist participants. Nineteen participants cited 2 or more group-based elements. Discussions centered on the implications of individual-level behavior for the community as a whole and on the relationship of the community to larger society (see box ▶ p 1934).

Table t4.

Group-Based Elements of Community

Divisiveness and Responsibility:

DURHAM AFRICAN AMERICAN: Now it's, like, everybody is on their own. Nobody cares about what happens, and you don't even have people participating in community-based projects.

PHILADELPHIA IDU: You got somebody over here with a big car, plenty money, and the kids see this. They're influenced by this, but by the same token, there's his grandmother, his mother, his aunt trying to say, “Well, no, that's not the way, come on, we goin' to church, or we goin' to the Center,” you know.

Leverage and Responsibility:

SAN FRANCISCO GAY MAN: Things are gonna start to grow, and if you wanted those little businesses to survive, why weren't you giving them your business? Why weren't you going to lunch at their place? Why weren't you buying books at their store? Oh, you weren't? But yet . . . you want to choose who comes in there? I said that's not life, that's not business.

PHILADELPHIA IDU: It was like community because people were concerned. I can go down the street and break out somebody's window playin' baseball, right, and before I got home, my mother would know about it and everybody that was involved. You know. It's because everybody was concerned about what was going on in the neighborhood. Just like, on the same note, if we didn't have any food, you know, and somebody would get word of it, friends would bring us food over.

Pluralism:

PHILADELPHIA IDU: When I was growin' up there it was a White neighborhood, now it's all Spanish and all Black and all everything, Chinese, and Koreans and so it's kind of not a community anymore, it's so mixed up that there's no nothing there anymore. Koreans messin' with Koreans, Whites messin' with Whites, Blacks messin' with Blacks and it's all mixed up there and nobody bothers nobody no more.

Divisiveness referred to descriptions of community fragmentation or a lack of unity, often expressed as an overemphasis on individualism and self-interest, or as attitudes that hindered unity and cooperation. Divisiveness was often discussed in the context of the need for or a perceived lack of responsibility on the part of community members.

Discussions of leverage centered on the potential ways that groups or individuals can bring about positive or negative consequences for the community as a whole. Pluralism referred to discussions of the maintenance of distinctions between coexisting ethnic groups. Unlike the core element of diversity, which focused on variability in a wide range of individual-level characteristics, pluralism implied ethnic and cultural distinctions among people living in the same area. Cultural pluralism is an anthropologic concept defined as “social and political interaction within the same society of people with different ways of living and thinking.”24^(p658) The challenges of living in a pluralistic community were described primarily by African American, Latino, and Asian/Pacific Islander gay men in San Francisco who attempted to navigate simultaneously among problems related to their ethnicity (e.g., racism, restrictive immigration laws) and those related to their sexual identity (e.g., homophobia, rejection by family members).

Responsibility was discussed in terms of the way people were or should be responsible for their own behavior, including how their behavior reflected on or affected the community as a whole. As such, leverage and responsibility were often discussed together.

The Impact of Stress on a Community

Two clusters emerged that centered on stresses affecting community. The first centered on the negative effects of criminality and drug use. As seen in Figure 1C ▶, although these were more frequently discussed by participants in Durham and Philadelphia, they were described as elements that undermine community in San Francisco as well. In contrast to the stresses of drug use and criminality, AIDS was more likely to be described as something that brought people together in a common struggle, increasing a sense of unity. For one Durham participant, the violent death of a child prompted a similar response, motivating her to work to improve circumstances in her community (see box ▶ p 1935).

Table t5.

Impact of Stress on a Community

PHILADELPHIA IDU: Do you know how many people was tryin' to rob me outta that coat right there? [Points to her coat on hanger.] I mean, these are people that's in the community, okay, and that this is supposed to be a community—why is everybody doin' what they doin' to one another?

SAN FRANCISCO GAY MAN: Crystal meth has really bothered me, just 'cause I see it as really damaging to my community. . . . I see people whose lives are getting all messed up, and I see that it's everywhere. I mean, it's almost epidemic in the gay community.

DURHAM AFRICAN AMERICAN: I never mingled or associated with anyone. I didn't want to be a part, you know, but I think the thing that really brought me out was when the [child] got killed over here. . . . The child was sittin' out on the stoop, other people were sittin' out on their stoop and stuff, and it was someone shootin' at another person, and one of the shots hit the child and she got killed over here in our neighborhood, and I think that is what basically brought me out to want to be a part of the community and get something done about what's goin' on over here.

SAN FRANCISCO GAY MAN: . . . as I get older, I have a stronger sense of [community], or a stronger sense of commitment to the community. I mean, you know, that's why I'm getting increasingly involved with AIDS activities.

SAN FRANCISCO GAY MAN: I don't think we have much of a gay community, unfortunately. I wish we did. But we don't because of what we are. We're everything. We're Black and we're White and we're poor and we're rich, so how can we have a gay community—I think that's a really silly word, the “gay community,” because it doesn't really exist. That's like the White, male, heterosexual community. I think that's silly.

PHILADELPHIA IDU: You ask me what the community is. Nothin', zilch. To me there is no damn community.

Statements that community was nonexistent were made by 6 respondents from Philadelphia and 3 from San Francisco. These discussions often included statements about diversity or pluralism as an obstacle to the development of community, by contributing to a lack of common identity or undermining a sense of responsibility to the group. For the Philadelphia drug users, the lack of community was sometimes seen as pervasive, or the drug culture was viewed as a noncommunity that was both separate from and surrounded by a functional local community that included the elements of locus, action, ties, and sharing.

DISCUSSION

The Elements of Community

Collectively, the Project LinCS participants described community by using a limited set of elements that reflect concepts previously noted in the social science literature. Four of the core community elements identified here through empiric means—locus, sharing, joint action, and social ties—are commonly found in social science definitions of community. In an early literature review of 94 definitions, Hillery25 found that two thirds cited social interaction, geographic area, and common ties as essential elements of community life, and almost three fourths cited area and social interaction. A review of an additional 60 definitions subsequently published in the social science literature found little change beyond a slight increase in emphasis on “people with common ties residing in a common geographic area.”26

Taking a different approach, McKeown and colleagues27 analyzed the way community was conceptualized in 4 classic ethnographic studies conducted by 2 anthropologists at different stages in their careers. They noted overall agreement in the use of 4 basic attributes to describe community: locality, biological and social membership, common institutions, and shared actions. From a psychological perspective and using an empiric approach that parallels our own, Chavis and colleagues28 identified 4 elements composing a sense of community: membership, influence, integration and fulfillment of needs, and shared emotional connection.

Using data from a study in a suburb of Toronto, Wellman and Wortley29 argued that locus was of decreasing importance for urban communities and that these were best described in terms of “personal community networks” that are socially diverse in composition, spatially dispersed, and sparsely knit. Others have suggested that the decreasing importance of locus actually leads to a sense of the loss of community. For example, Glynn30 evaluated the relationship between people's ideal sense of community and their perception of their actual community in 3 diverse settings (an Israeli kibbutz and 2 dissimilar cities in Maryland) and found that neighborhood identification was important for the development of an actual sense of community.

Patrick and Wickizer9 reviewed social science definitions of community with an eye toward developing and implementing effective community-level health interventions. They identified 3 broad conceptual approaches to the definition of community: those that defined community as place, as social interaction, and as social and political responsibility. The concept of social and political responsibility is similar to our core element of joint action, combined with our group-based elements. Patrick and Wickizer9^(p51) offered a working definition of community as “the entire complex of social relationships in a given locale, and their dynamic interaction and evolution in working toward [the] solution of health problems.”

The importance of local diversity has not been previously articulated in definitions of community, although the effect of such diversity on health measures has been noted. For example, Sampson and colleagues31 pointed to the need to explore the meaning and sources of variation within neighborhoods or local communities for collective efficacy for children. Zakus and Lysack16 noted that communities are rarely, if ever, a homogeneous whole and that this represents a major challenge for successful community participation in setting health policy. The fact that diversity emerged as a core element in our empiric exploration of definitions of community was driven to a great extent by the experiences of the gay men who were interviewed. San Francisco is a national and international meeting ground for gay men. The interviews we conducted suggest that many of them are consciously seeking to build a community based on a positive valuing of unity, diversity, and cultural pluralism. With increased mobility and immigration throughout the United States, the importance of diversity for community structure and function is likely to increase in other locations and for other populations. The challenges presented by local diversity, in turn, are likely to become increasingly important for public health efforts as well.

Implications of the Experience of Community for Collaboration

The saliency of the different elements of community for each of our 4 participant groups had implications for the ways in which our collaborations developed. For gay men in San Francisco, a strong sense of shared history and perspective was a dominant theme, followed by a sense of identity with a specific location, the creation of strong and lasting social ties, established avenues for joint action, and the role of diversity. This profile is superficially similar to the one elicited from the vaccine researchers; however, particular elements were discussed less frequently than in San Francisco. Significant differences also existed with regard to how the elements were discussed. Most of the San Francisco participants had thought about community, and many were struggling to reconcile their need for community with a sense of marginalization from society at large. In contrast, the scientists tended to describe themselves as well grounded in multiple communities.

The profiles for Durham and Philadelphia also had a surface resemblance to each other. In contrast with San Francisco participants and the scientists, Durham and Philadelphia participants viewed locus as the principal element of community. This was especially true for IDUs in Philadelphia. Both groups emphasized the importance of joint action and social ties, while minimally discussing the role of diversity. African American participants in Durham included more college-educated and nonheterosexual individuals than did Philadelphia participants and, perhaps as a result, were somewhat more likely to discuss the role of shared perspectives for community than were Philadelphia participants. In both Durham and Philadelphia, most people described community as a “given” in their immediate environment. However, for IDUs, the given community environment was less likely to be described as supportive than for African Americans in Durham.

These differences in the way people perceived and talked about the core elements of community suggest the need for multiple models of collaboration for public health research and programs. In fact, the collaborations we established in San Francisco, Durham, and Philadelphia illustrate this need. At each site, researchers worked with a community advisory board (CAB), but in different ways.

The San Francisco collaboration.

In San Francisco, community advocates and activists were collaborating with HIV vaccine researchers before Project LinCS was funded. Consistent with the emphasis on sharing seen in Figure 1A ▶, the CAB members placed a high value on opportunities for in-depth discussion with both local and nonlocal Project LinCS collaborators. Similarly, they recognized the importance of existing social ties and activities and were careful not to allow Project LinCS to draw energy away from other HIV vaccine work. As a result of their long involvement in treatment and prevention activities, many San Francisco CAB members had a sophisticated understanding of research. Reflecting the saliency of joint action for the community, the CAB used this knowledge to take an active role in developing the protocol and interview guide for the scientists. Additionally, half of the scientist interviews were conducted by a San Francisco CAB member (with full support from the Philadelphia and Durham CABs). The CAB also collaborated with local investigators in the development of a complex targeted sampling plan to ensure maximal diversity of Project LinCS participants, reflecting their awareness of the important role of diversity and pluralism in their community.

The Durham collaboration.

In Durham, the African American community was beginning to mobilize around AIDS when Project LinCS began. Here, the local investigators had to reach out widely to people and organizations with links to the African American community. A socioeconomically diverse group of representatives came together and volunteered to work with the researchers, meeting at a historically African American university campus in Durham. Consistent with our analysis that showed an emphasis in Durham on locus and joint action, the CAB focused its efforts on making sure that the project provided tangible benefits to the local community. These efforts resulted in a brochure on questions to ask when volunteers were invited to participate in research and a local newspaper insert on lessons learned from the project, which was distributed to more than 11 000 houses in predominantly African American neighborhoods.

The Philadelphia collaboration.

The Philadelphia investigators had a long-standing relationship with IDUs that centered on a storefront research program. About a year before Project LinCS began, the investigators invited study participants to form a CAB. The meetings were initially chaired by the principal investigator, but later the CAB members established their own set of rules and took on increasing responsibility for the functioning of the board. Philadelphia LinCS participants emphasized locus, action, and social ties in their definitions of community. Similarly, the CAB defined its primary role as one of maintaining and building linkages between the research staff and the IDUs in the surrounding neighborhoods, a community that functioned largely through informal structures and at the margins of society.

The Role of Community Representation for Collaboration

Israel and colleagues13 and Zakus and Lysack16 noted that participatory approaches such as ours that rely on representation can lead to conflicts with regard to how community is defined and who may legitimately represent the community. Our experiences, and our empiric data, suggest that an important element for success may be ensuring that CAB representatives are actively connected to diverse people in their local communities and empowered to function in ways that are meaningful to their community base. Other research supports this view. Conway and colleagues32 compared perceptions of health priorities among local District Health Council members and among a random sample of household residents in Chicago and Cook County, Illinois. The results showed substantial agreement in priorities, indicating that advisory boards can effectively represent community perspectives regarding health priorities. Jewkes and Murcott33 presented results of a qualitative assessment of the uses, meanings, and interpretations of community participation in the context of the World Health Organization's Healthy Cities Project as implemented in the United Kingdom. In interviews with 50 participants drawn from health, local government, and voluntary sectors, they found that “being known” was the most fundamental requirement of an effective representative. Data from a case study by Bond and Keys34^(p37) support the feasibility of empowering multiple community groups simultaneously through a single advisory board “when the board culture promoted inclusionary group processes and the activation of member resources.”

If collaboration is to be an effective component of public health research and programs, it will require a greater understanding of the way people interact individually and as groups. The definition of community provided in this commentary supplies a potential framework for investigating such interactions. Each of the core elements composing community (locus, sharing, joint action, social ties, and diversity) can be evaluated relative to public health outcomes through existing social science models, including social network analysis,35–37 sense of community,38–42 social capital,43,44 cultural domain analysis,22,23,45,46 and geographic information systems.47 Such models provide a solid foundation for a systematic approach to community-level and community-based public health research and programs.

CONCLUSIONS

The results of our analysis point to a core definition of community as a group of people with diverse characteristics who are linked by social ties, share common perspectives, and engage in joint action in geographical locations or settings. Our results further suggest that a cookbook approach to participatory programs and research will not work because the experience of community differs from one setting to another. Rather, each research collaboration, and each level of collaboration from the local to the national and the international, must reconcile the differences and similarities among the participating communities.

Community collaboration in public health programs and research presents many challenges, in part because community has been defined in ambiguous and contradictory ways. Despite important differences in the experience of community, our study suggests that people largely agree about what community is. The empiric evidence, in turn, is bolstered by established social science theory. Additionally, existing social science tools provide a strong foundation for measuring and evaluating the contributions of community collaboration to the achievement of public health objectives.

Thus, a sound empiric and theoretic basis exists for achieving consensus on a definition of community for public health. Consensus will facilitate the systematic comparison of local populations by directing attention to a set of core elements for measurement. Systematic comparison, in turn, will facilitate hypothesis testing and strengthen the scientific study of the role of community in public health. For example, it could help us identify functional thresholds for the core elements, such that groups above the threshold are significantly more likely to experience beneficial health outcomes than those below the threshold. In other words, it can help us understand how to build and support “good” communities that enhance the health of their members. It can help us understand which characteristics or combinations of characteristics are necessary or sufficient for supporting intermediate goals such as the sustainability of prevention programs or the diffusion of beneficial health practices. And it can provide a sound theoretic basis for building successful community collaborations in public health through the systematic evaluation of who participates, why they participate, what they share, what they do, and how participants are connected to each other and to their constituencies.