  |
|
|
|
Day 1 (Friday, June 7) Introduction Paul DiMaggio opened the conference with several broad questions regarding the relationship between the arts and community life. He noted that we can think about the relationship in terms of what accounts for the health of artistic communities in geographically defined areas (e.g., what makes an arts community healthy for creativity, diversity, and innovation). Or, we can examine the impacts that artists, arts activities, and arts institutions have on their communities. Drawing on Josh Guetzkow’s paper, How the Arts Impact Communities: An introduction to the literature on arts impact studies, prepared for this conference, DiMaggio pointed out that any attempt to measure the “effects” of the arts on communities requires a careful and precise definition of both terms (“art” and “community”). Second, it is difficult to isolate the effects of the arts in a complex system (e.g. a community). Different types of art activities or organizations might have distinct effects on different citizens. Given this heterogeneity, it is unrealistic to expect large effects on most community-level outcomes from arts-related (or any other specialized) interventions. How do we study the impact of the arts when effects might be non-linear and complex, challenging many of the assumptions that underlie typical models used in the social sciences to measure cause and effect relationships?
Session 1: Mature Research Fields in the Arts: Economic Impact and Schooling Effects In the first session, Ellen Winner presented her research on the relation between arts education and non-arts, academic performance. Despite the many claims that the arts increase children’s verbal and math test scores and their self-esteem, few studies have actually demonstrated a causal relationship. The Reviewing Education and the Arts Project (REAP), directed by Ellen Winner and Lois Hetland, carried out met-analyses of more than one hundred studies that examined the link between arts and academics. They found a strong correlation: students who elect to take arts courses in school or after school get better grades and higher scores on standardized tests than those who do not involve themselves in the arts. However, no studies showed that studying the arts is the actual cause of increased academic performance. Winner noted that the correlation between the arts and academic skills might reflect the fact that high achieving students in the US choose to participate in the arts. They might do this because of family values, or even to gain a competitive advantage in college admissions. In the U.K., students who took many arts courses actually performed worse on national exams than those who focused only on academic courses. One possible explanation is that in the U.K., academically weak students are guided into the arts, while strong ones are advised to focus on academics. Thus, context is important and selection effects may explain the positive correlations in the US between arts study and academic performance. The only causal links that Winner and Hetland’s meta-analyses uncovered were between music listening (in adults) and spatial test scores (an effect that lasts only minutes), music performance and spatial test scores (an effect that is not short-lived), and classroom drama and verbal skills. Professor Winner concluded that claims about the impact of arts education are problematic. First, researchers who attempt to demonstrate the transfer of arts learning to other domains (science, math, reading) need to clearly explain the process by which this happens. And, they must be able to measure the relevant arts learning in the first place. For example, certain types of art education might help students develop a capacity for “close observation” that could transfer to science; or they might learn perseverance; or discipline; or use the arts as an entry point into other subjects. But, these outcomes must be measured explicitly; control group and comparison cases must be selected carefully; and teachers must be prepared to teach explicitly with the desired transfer in mind (it most likely will not happen accidentally). Finally, Winner felt that arts advocates should not rely as heavily as they do on utilitarian arguments, but rather should endorse arts education for its own sake. Rarely, if ever, has research demonstrated that learning in any subject matter transfers to learning in any other. We should not require more of the arts than we do of other subjects. In the session’s second presentation, Bruce Seaman offered a critique of economic impact studies. While economists have recently become more sophisticated about measuring the impact of spending, in the past typical measurement errors in impact studies included: spending diversion errors (the failure to exclude spending by local residents who would have spent their money locally even had they not spent it on the arts); ancillary spending errors (erroneous attribution of all hotel, restaurant, or retail spending by non-local arts consumers, many of whom may have come to town for other reasons); and use of unrealistically high “multipliers” to estimate aggregate effects of arts spending. Seaman also reported some additional lessons from existing research on the economic impact of sports. In particular, he noted the supply constraint, or crowding-out error – sporting events often compete with other planned events for the same fixed resources (hotel, restaurants, taxis) – leading to a substitution of spending, rather than additional spending. (For example, Atlanta pulled out of two bids to host the Superbowl because the hotels were already full.) Similarly, World Cup hosting countries tend to lose money because World Cup activities substitute for other economic activities. He also noted the ex-post verification error where unforeseen events may change previously reported impact assessments (e.g., when sports teams leave or go on strike, for example, they have no economic impact). Finally, economic impact studies must contend with the problem of aggregation – i.e., spending that takes place in one city might pull spending away from its neighbor. From the standpoint of the first city, this is good, but from the standpoint of state or regional or national policy, there is no benefit. By focusing on an inter-regional impact analysis, the intra-regional component is missed. Seaman reported that the total economic impact of arts events and organizations should be calculated by taking the sum of consumption benefits to users and non users (e.g., those local citizens who do not attend, but who still value the existence of the event/institution), long run economic growth benefits, and short run spending impact. Economic impact studies, by design, rule out consumption benefits, and neglect long term growth, focusing almost exclusively instead on short run spending. Seaman suggested that contingent valuation was one method that could capture some of the local consumption benefits. This method relies on a survey of what residents would pay to expand, create or sustain existing cultural facilities (or natural resources). Contingent valuation methods are controversial, however, because of biases that are built into the approach. An example of this is “anchoring bias,” whereby respondents report being unwilling to pay much more than what they are told is currently spent on the arts. When the anchor (i.e., what is reported as current expenditures) changes, respondents’ willingness to pay changes as well. In a few cases, researchers have tried to use both traditional economic impact studies and contingent valuation methods and have come up with two different sets of values. But the relationship between these two figures is poorly understood. Can we say anything positive about economic impact studies? Seaman made the point that while typical economic impact studies might exaggerate spending impacts (because of many of the errors described above) they also fail to capture local consumption benefits and long term growth. Therefore, it is quite possible that their conclusions actually approximate the true economic impact (even as they draw their conclusions from flawed methods) because the biases tend to even one another out. Discussion: Following the presentations, participants raised questions about how to measure what children are actually learning through participation in the arts. There was discussion about the need to focus on outcomes or skills that can be traced more directly to such participation – discipline, motivation, study habits, team work, presentation skills – rather than the impact of the arts on other academic subjects (math, science, reading). Arts educators should focus on these outcomes and the intrinsic benefits of arts learning, rather than just spill-over effects in different subjects. It was also suggested that researchers distinguish effects for different types of students, and that they be more explicit about both the precise learning environment or activity involved, and the nature of the effects that supposedly result from that activity. Regarding economic impact studies, participants raised the concern that such studies fail to measure intangible impacts, like enhancing city and regional reputation or improving the quality of life -- factors which may, for example help determine where firms move their headquarters or where knowledge workers want to live. Another participant recognized that in complex urban-political systems (which most cities are), economic impact studies probably play only a minor role in influencing decisions about public investments
Session 2: Research on Communities Mario Small presented ethnographic research on the Jorge Hernandez Cultural Center in the predominantly Puerto Rican Villa Victoria housing project in the South End of Boston. The Center was founded in 1986 in an old church with two primary objectives: 1) to generate income and 2) to use culture as a means to celebrate Puerto Rican heritage and to create social ties between the neighborhood's Puerto Rican residents and non-residents from the surrounding areas. The Center pursued these goals by hosting social events and concerts highlighting Latino art and culture. Its programming attracted well known artists, large crowds, and ticket revenue for the cultural center and housing project. However, no cross-cultural social capital – what Robert Putnam has called “bridging” capital -- was created because the residents of Villa Victoria almost never attended the Center’s events. In addition to high ticket prices, Small argued that most residents did not participate because the programming did not appeal to them. The Center sponsored concerts and dance events that appealed to New Englanders’ ideas of “Latino music”—Cuban Salsa, Chicano Rock, Flamenco, Tango, Latin Jazz. However, this Pan-Latin American repertoire is very different from the music styles that Villa Victorianos know and enjoy, such as Trova, Puerto Rican guitar trios, Spanish Reggae (Reggaeton), Jibaro and Spanish Hip Hop. One lesson from Small’s work is that ethnographic research
in a community can help reveal the barriers to certain programmatic
objectives – in this case the use of music and culture to
build social capital. In the case described above, Small found
that differing cultural perceptions and musical tastes prevented
the Jorge Hernandez Cultural Center from creating the types of
social bonds that the program was intended to foster in the first
place, even as the programs increased the Center’s visibility
and provided a useful source of revenue. Claude Fischer, author of America Calling, a social history of the telephone, discussed the social consequences of technology as a cognate example for thinking about the social consequences of art and culture. First, he suggested that “consequences” was a better description than “impact” because the latter term suggests a “billiard ball” model of cause and effect – an external force (technology or the arts) supposedly “strikes” existing attitudes, social relations and institutions and causes them to move in new directions. In contrast, “consequences: suggests that there can be multiple forces at work (internal and external) on a variety of individuals and institutions, resulting in numerous unexpected, overlapping and often contradictory outcomes. Also, the introduction of new technology (or art and culture) can create not only new opportunities for citizens, but also new constraints. For example, the introduction of the car permitted a growing population of workers to buy affordable housing in the suburbs and commute to the city. But, its popularity also became a constraint because it became hard to live a middle class life without an automobile. And not all new technology pushes social life in the same direction. For example, when movies were introduced on a mass scale they became social destinations and brought more people to downtowns – by contrast, the introduction of streetcars helped cities spread out in a radial fashion. Sometimes the introduction of a new technology might not alter social life, but instead accentuate or facilitate existing patterns of behavior. For example, many people assumed that the telephone would increase communications between people or alter social relationships. Instead, it changed the efficiency of communication (i.e., it let people who regularly communicate do so more easily and it allowed people to move farther from one another and still maintain regular correspondence). In this case, the telephone had an unanticipated consequence on localism. Fischer concluded that to gauge accurately the consequences of any intervention (technology or the arts), one needs research over a long enough period to capture multiple, heterogeneous consequences. It is also important to have strong comparison cases. For example, Fischer’s own research compared the different consequences of the telephone and automobile. Discussion: Participants acknowledged the importance of conceptual precision in attempting to measure the consequences of art and culture for community life. For example, when people talk about the impact of the arts on social capital, the terms art and social capital are typically defined too broadly (or not defined at all). Some arts activities (like a community theatre project) might lead to bridging social capital (more ties between different groups of people); others (like a weaving club) might lead to bonding social capital (stronger ties within existing groups). Similarly, participating as an audience member (one type of art experience), might have very different consequences than participating in an event that requires back-and-forth engagement (a different type of experience all together). Some arts activities are primarily about self-expression; others about promoting community interests. These two types are unlikely to have similar consequences. One person pointed out that Small’s work in the Villa Victoria housing project demonstrated that policy (and research) often focuses too much on form rather than content of artistic programs. In this case, creating a venue for dance and music (a successful concert series) was the desired outcome for the project. However, by ignoring how the “content” of the program (Pan Latin music and dance) corresponded to the lives and interests of residents, policy makers failed to recognize that building “social capital” (a stated goal of the series) was not possible unless the content could effectively engage both residents of Villa Victoria and surrounding New Englanders. Finally, there was a recognition that Fischer’s work on
the social consequences of technology provided important lessons
for micro-economic theory. In particular, micro-economic theory
assumes that people’s preferences for consuming art, media
and other goods and services are fairly stable. However, as Fischer
has demonstrated, as social environments change (e.g., with the
introduction of new technology), preference might change in unexpected
ways. We need better theories and empirical research about how
individual preferences come into being in the first place and
how these preference adapt to changing economic and social conditions.
Session 3: Perspective from the Natural Sciences Henry Horn discussed different approaches to thinking about biological diversity. First, how do we determine the optimal or minimal level of diversity for any given community? For example, what is the minimal number of different kinds of species that are necessary for a viable ecosystem (a similar question could be asked about cultural diversity). Horn pointed out that any measure of diversity will greatly depend on scale – are you looking at one square acre or one thousand square acres? What is the appropriate scale for examining diversity? Levels of diversity or heterogeneity plotted against one scale might be entirely different when examined at a different scale. It is also important to look not only at diversity in terms of overall numbers of different species, but also in terms of patterns and distributions. That is, what are the relationships between different species in any given area? What do the patterns look like? Horn suggested that discrete biological patterns might result from certain self-organizing principles or be the outcome of an historical process of growth and development. Some patterns are simply historical accidents. And, certain processes might be really good at creating new patterns, and really bad at erasing existing patterns. The point is that certain biological arrangements might be very difficult to undo (again, analogous processes might be at work in cultural ecosystems). Horn also raised a series of questions about the underlying motivation and justifications for wanting to conserve or foster biological diversity. Is it to protect aesthetic novelties? Irreplaceable species? Is it to protect the functioning of an ecological system (utilitarian claims similar to those made on behalf of the arts)? Is conservation a moral debt to the past or future? And, assuming that you can arrive at a consensus about the importance of biodiversity, how do you actually measure diversity? Do you count the number of different species? Or, do you count categories of species? Do you measure diversity by function (not by the intrinsic qualities of a species, but by the function it performs in a larger system). How do you count? Is rarity a plus or minus? Or in other words, do we want a community with high numbers of relatively rare species (high variance)? Or is it better to have more robust populations of fewer species (lower variance)? How do you assign weights or prioritize (which species should be more abundant and which less)? Finally, the more details we can collect about the unique characteristics of entities in an ecosystem, the richer are the analytical possibilities. However, too much information can lead to data narcosis – making it difficult to find any sensible patterns or to process the information in meaningful ways. So, there are tradeoffs in measuring and describing diverse ecosystems – and many meaningful distinctions might be lost in an effort to gain analytical efficiency. The discussion began with a comparison between environmental conservation and biodiversity and cultural heritage and cultural diversity. Research in the latter area has been focused on techniques of preservation and conservation and has neglected the social consequences of maintaining heritage sites, buildings and practices. And, unlike environmental conservation, the preservation field has not developed methods or theories to measure diversity or to trace the interrelationships between sites and structures or to study the community-level characteristics of places that have a rich, well preserved cultural heritage. How do different part of the cultural system interact? How many types of artists or organizations do you need to have a viable community? What is the maximum or ideal number that can co-exist in any one space? What is the role of competition between art forms and institutions for audiences and resources? (note: sociologists have borrowed ecological models to study competition between organizations). It was noted that humans have the capacity to purposefully cooperate to improve the cultural system. Perhaps heterogeneity and diversity of cultural content would be enhanced in a environment of cartel-like cooperation. There is general agreement that biodiversity is a good thing, but there is still no consensus on whether cultural diversity is equally important. Perhaps we need to develop a set of agreed-upon indicators of “cultural diversity” (or heterogeneity); and then communities could evaluate whether they meet certain baseline levels and whether the cultural life of a community is getting more or less diverse over time. And, policy makers could better evaluate whether interventions lead to more or less diversity. Perhaps arts researchers can learn from the natural sciences in developing approaches to measuring diversity. Do we look at diversity of form (music, dance, theatre, craft) ? Genre? Diversity of function (participatory arts, community arts, arts schools, museums, etc.)? Institutional and organizational diversity (age, size, location)? And at what scale should we look for diversity, given the fact that what appears to be diverse at one level (e.g., a city) may mask much homogeneity at more local levels of analysis (for example, neighborhoods within that city). Drawing on the ecological case, the question was raised as to whether “new growth” is more likely in a diverse ecosystem. Similarly, perhaps “innovation” or new cultural forms are more likely to emerge in a diverse and heterogeneous cultural milieu than in less diverse places. And perhaps latent interests in the arts might be more likely to become manifest in environments where there are more different stimuli and cultural options available. The discussion highlighted the fact that in systems that have tight interdependency among system elements, those elements will interact in ways that yield nonlinear relationships between inputs and outputs. Perhaps under some conditions, conventional ways of talking about and measuring causality become inoperative. Biologists have not yet developed useful methods for analyzing communities with high numbers of species each of which influences the vital rates (births, deaths, and so on) of every other. In so far as human communities are complex systems characterized by nonlinear relationships among complexly interacting components, how can one develop deliberate and meaningful interventions? Are there lessons from environmental science and policy that inform how one can assess interventions? What can cultural policy learn from the natural sciences in this respect? The upside of nonlinear effects is an adaptive landscape. If you are in a valley, a ridge may not be far away. A small push in the right direction can have a very large effect. Also, it was noted that many non-linearities follow an S-curve – where exposure to some intervention of interest (e.g., arts participation) will have no effect at first; then, at a point of inflection, one observes a narrow range of significant effect, followed by an asymptotic leveling of effect at higher levels. Researchers need to do a better job of identifying the inflection point, because this is where small interventions can have the greatest impact.
First Day Recap The recap began with a discussion of the challenges that funders face in trying to support research or influence social outcomes. Funders are generalists and pragmatists and typically in a hurry. This raises several questions. First, given the richness of the conversation so far and the daunting research tasks ahead, how much information and research is needed before making a decision? Second, one major conclusion from the meeting thus far is that because the cultural system is so complex, it is important to disaggregate vague categories or definitions and achieve much greater levels of specificity. However, when will we know when we have reached a level of specificity we can trust to make decisions, and will we have the patience to get there? Regarding scale, how far will we have to back up to see the picture? Third, how much should we be trying to influence taste preferences and behaviors? Should we even try to engineer cultural systems? Others mentioned that, for both methodological and operational purposes, we need more clarity and precision about both the inputs (for example, types of artistic practice or institution) and the outputs (the specific community or individual level impact) in which we are interested. . One important step would be to identify a common outcome that the funding, policy and arts research communities agreed is important. Is there consensus around the importance of measuring diversity and heterogeneity? What other outcomes? Perhaps in education there is growing consensus around the importance of measuring learning, rather than achievement (e.g., test scores). The discussion has assumed that to understand “impact” we need to identify causal linkages. Is there any role for studies that show correlation and interdependence without demonstrating causality? There are at least three possible approaches to examining causal relationships between arts activities and outcomes at the community level. First, one can use linear models: researchers attempt to extract five or six variables from a system and regress them on some outcome of interest (e.g., examining the relationship between spending on orchestral music over time and the number of new compositions performed by symphony orchestras). Second is the systems approach of which SimCity is the popular epitome, where one identifies the key components of a system (artists, arts organizations, funders, and so on) and tries to identify enough parameters linking parts of the system to model the relationships between these parts. Sometimes the SimCity approach is too limited, reducing the dynamics of system to too few parameters. In such cases, a third approach may be necessary – a complex systems model. Finally, one participant pointed out that the field needs to
take into account who will undertake the type of complex analysis
described above? What are the incentives to encourage scholars
to do work in cultural policy and on the cultural sector more
generally? Does the tenure system support work in this field? Day 2 (Saturday, June 8)
In the first presentation, Blair Wheaton discussed useful lessons for cultural-policy researchers from research on stress and mental health. Early research on mental health (1950 to 1970) was poorly integrated. Advocates of many different theories and schools of thought and different approaches to diagnosing disorders tended to talk past one another. In the 1970s a decision was made to standardize the different definitions of “psychiatric disorder,” stating specific criteria for each. Even though the early taxonomies were far from perfect, the very process of standardization led to national surveys of mental disorders (using agreed upon boundaries and concepts) and to an explosion of research and findings about the distribution of mental health among people and groups. One lesson is that reaching consensus about cultural indicators might be necessary for certain types of arts research to reach adequate scale and scope. Three other general lessons for arts research include these:
Research on stress has called attention to the importance of both social and individual context. Wheaton demonstrated how measuring only individual-level effects can have misleading results. He also provided evidence that community-level factors can have a strong influence on individual levels of health and well being. For example, neighborhood wealth and community organization have significant effects on birth weights, above and beyond the effects of income, race or the marital status of the child’s mother (individual effects). Or, the effects of income on depression (typically lower incomes correspond to greater stress levels) depend on the poverty level of a neighborhood. Poor people are more depressed when living in poor neighborhoods then they are when living in more wealthy communities. Hierarchical Linear Models (HLM) can be used to distinguish the effects of a neighborhood as distinct from the unique experiences and traits of individuals. Not just social context, but individual context as well, has important conditioning effects on the relationship between stressful life events and mental-health outcomes. For example, the average effects of divorce on mental health across all divorced persons is weak. However, if you take into account the context of the marriage before divorce, you find very strong effects. Specifically, partners who experience their marriages as good suffer great distress when they end; by contrast, when bad marriages end in divorce, stress levels may actually be reduced. It is also important to realize that contexts change over time, and that past contexts may continue to exert an influence on present day mental health. Using a technique known as cluster-based hierarchical models, researchers have shown that for some adults, characteristics of their childhood neighborhood might exert a greater influence on their mental health then where they reside currently. One possible research strategy would be to 1) use GIS geocodes to create cultural indicators (context) at the neighborhood level; 2) collect national aggregate information over time (or look at existing national data sets that have a dependent variable of interest – crime, stress, civic participation, or so on; and 3) link the two with geocodes and then estimate the effects. Discussion: Wheaton’s talk demonstrates that the simplest methodological approach often leads us to the wrong explanation. For example, education levels of spouses might predict the chances of divorce. But, looking at only that factor in isolation may miss the mediating impact of social networks or the background characteristics of marriages that lead to divorce. Likewise, scholars who study arts participation have tended to look exclusively at a few unchanging demographic characteristics to explain participation – education, race, income, residency – without studying the social or cultural context within which participation (or non-participation) takes places. Such research is also not very useful to funders and policy makers who are looking for levers with which to intervene to increase participation, because a foundation cannot change someone’s demographic profile. Instead, scholars should examine demographics not as a sole determinant of participation, but rather as a factor that influences different responses to treatments or interventions. So for example, higher income might generally predict higher levels of participation. However, a more interesting finding would be that people at certain levels of income respond differently to different types of art education – for some types of arts education, for example, low income participants might actually respond more favorably than people at higher income levels. Second, Wheaton’s talk points out that scholars and policy makers should pay more attention to the “when” and “where” of interventions. Again, context makes a difference. Regarding arts participation, at what stage in a life course is it most effective to intervene? What environment is most conducive to different types of intervention – home, school, clubs, individual-based learning, or group-oriented participation? Finally, we should recognize that many of the relationships linking the arts and culture to community-level and individual-level outcomes are nonlinear. The challenge is to model that nonlinearity, taking into account context and life course factors, and then find the right time and the right place to intervene. In the second presentation, Bruce Western, discussed the advantages of using Bayesian analysis for studying community-level outcomes. Modeling community-level effects can be complex (e.g., using hierarchical models) and often requires a great deal of data to capture nonlinearities and the contextual effects. Complex models can be very costly (both in terms of the data collection, but also in terms of the simplicity of an argument or explanation). Another way to analyze community-level effects is to do comparative research across communities, cities, states or nations. Rather than study individuals within communities, you can look at communities themselves. However, such comparative analysis often suffers from weak data (small samples, data that are not based on random samples, multicollinearity, and so on) as well as weak theory (e.g., there are typically many different possible explanations for an observed outcome). Under such conditions researchers tend to try out different models to see which fit the data best, a practice referred to as data mining. Data mining can yield spurious relationships – that is, observations of statistical relationships in the data (i.e., one factor appears to predict or explain another) that are the result of chance, rather than reflections of real causal processes. Bayesian analysis is particularly well suited for overcoming these challenges and explaining variation across a limited number of communities. The Bayesian approach has two implications: 1) things that we often treat as certain can be made more uncertain because we treat them probabilistically; and 2) uncertain things can be made more certain because we can introduce information from outside the sample (from contextual knowledge, historical knowledge, or case study research) and have this reflected in our results. In the first case, we essentially test different models and determine a probability for each model that takes our uncertainty into account. In the second case, scholars can include a range of relevant information that is typically not captured in normal quantitative models. Thus, Bayesian analysis can be used to understand differences between a limited number of communities. It is a powerful tool with small data sets because it provides an opportunity to combine quantitative and qualitative information. Discussion: In response to the presentation, there was discussion about the role of qualitative research in studying communities. Under what conditions is qualitative research appropriate? And, how can qualitative and quantitative research be integrated better? How can the two methods work side by side? Second, a question was raised regarding the extent to which research, generally, is policy relevant. For example, hierarchical models may help us to learn that neighborhood poverty is associated with low birth weights (or low arts participation); but then what role can policy makers play in affecting change? Other than ending neighborhood poverty, what other options are available for intervention? Regarding the first issue, the group discussed many ways in which qualitative research can be designed to gain maximum leverage on understanding community-level outcomes. First, qualitative research can be very effective when comparing a few cases (at least 2), that are similar in almost every way but also differ on a few important dimensions --- for example, countries that share almost identical histories and economies, but have very different political institutions. Qualitative case studies also help reveal underlying mechanisms that explain the relationship between some outcome and a community characteristic. Case studies also can be very powerful for looking at unusual cases – e.g., trying to understand a poor community that has an unusually high investment in the arts. Additionally, if a quantitative model reveals that there is an inflection point (or tipping point) at which the probability of a certain outcome (e.g., experimenting with a new art form) rises sharply, then qualitative case study work may help reveal how to intervene most effectively at that point. With regard to effective policy intervention, it was suggested that research must first reveal underlying, perhaps immutable, conditions (e.g. the effect of poverty on arts participation) in order to identify policy levers that might make a difference. Three other challenges face policy makers. First, the issue of latency is a problem: many impacts may not show up until 5 to10 years into the funding cycle. Second, it is crucial to understand what statisticians call “suppressor effects” – factors that lead to the ineffectiveness of an intervention that would otherwise have yielded the desired effect on a community or individual. Finally, policy makers must pay attention to “interaction effects,” when certain interventions work differently on different group of people.
Session 5: Robust Institutions and Communities Ann Swidler gave the conference’s final presentation on the general theme of robust and well functioning communities and institutions. Her presentation focused on three topics: 1) How would we recognize a healthy and robust community, and what contribution might the arts make to such a community? 2) What types of cultural or artistic “outcomes” might a researcher want to measure at the community level? And, 3) What types of interventions are possible? First, Swidler laid out three things that robust institutions
should do: sustain individual, family, and group functions; create
leadership responsive to their members; and sustain movement toward
collective goals. Studies of Native American tribal communities
have revealed that these communities are robust when governance
institutions are aligned with the mythologies and narratives of
their citizens. Community members derive identity from rich cultural
narratives, which provide legitimacy for government and social
institutions and shape the norms within which community leaders
operate and serve the citizenry. Perhaps researchers should pay
more attention to how communities construct narratives –
through stories, photographs, buildings, songs, festivals –
and how these narratives render collective meaning and identity.
Second, Swidler discussed the role of art and culture in creating
or sustaining prestige systems. She cited work in the sociology
of culture that demonstrates that art institutions and art consumption
have served to mark and sustain boundaries between high and low
status persons. But, policy makers can also use the arts to democratize
existing prestige systems. As an example, when a curator says
that graffiti is beautiful and museum-worthy, he or she has added
legitimacy to a form of cultural expression and a subculture that
was not previously validated by the larger culture. Should policy
makers invest more in institutions that will display or archive
lower-status cultural traditions (e.g., music of youth subcultures,
or folkways) in ways that enhance the status/prestige of these
groups? Finally, Swidler emphasized the need to focus on “cultural stuff” itself, rather than the instrumental role that culture plays in producing other socially desirable outcomes. She asked, “Why not see if arts programs increase the amount of art on parent’s refrigerators?” Perhaps policy makers and community leaders can agree that more creative expression – more cultural work –- is a good thing in itself. Innovation and creativity are legitimate and valuable outcomes. Others have found that competitive market conditions lead to more innovation and more diverse cultural forms. Perhaps, arts funders could intervene, when appropriate, in markets in order to encourage more competition. Likewise, research has found that where audiences and producers interact on a regular basis (e.g., jazz, which is performed in clubs where listeners are often also knowledgeable aficionados) innovative flourishes. Perhaps there are ways to intervene to instill in popular culture circuits some of the same organizational features that characterize high culture circuits: to provide spaces, for example, where critics, producers and audiences interact intensely, providing more avenues for feedback and more room for innovation. Discussion: One challenge for the cultural community is making the case that the outcomes we are interested in (more and better art and creativity) are important. In other fields, like mental health or crime, there already exists a consensus around a desired outcome. Swidler’s presentation encourages us to think about a range of indicators that we might measure – community health or robustness; creativity and innovation as a pattern of collective behavior; creative work in non artistic occupations; shared narratives and collective identity. And, we might develop methods to study any one of these indicators. The challenge is to create an aggregate picture: do we have a synchronized way to bring all this together? Additionally, we need to continue to remind ourselves that research must find an audience among policy makers. So, for example, research on prestige structures might prove revealing, but this will need to be translated into terms that make sense to school leaders, policy makers, and parents. To what extent is the particular content of art important, and to what extent should we focus on art’s functions (e.g., to foster social relations, provide avenues for self expression)? For example, art activities may bring people together to build friendships, but other activities (e.g., sports) do the same thing, and might do it better. We need to think about what is distinctive about art and culture. Swidler makes clear that one distinctive characteristic of art is that it produces cultural meaning for a community and fosters collective memory. Some narratives serve not only to build solidarity and understanding within a community but also serve to build bridges between different groups and communities (or to maintain boundaries). Thus it is important to recognize that people might share a common framework or overarching story, but differ in the way that framework is interpreted. A robust community may depend on shared cultural narratives that are open to reinterpretation and evolution, and which can serve to link different communities together, not just strengthen existing structures.
Concluding Session: Taking Stock and Implications for Research Grantmakers began the final session by discussing the type of research on community cultural impacts they would like to see. Given what was learned at the meeting, one person called for “a much deeper conversation and an acknowledgement of the complexity of the cultural sector,” recognizing “that the sector has changed in the last 20 years and that we lack the methodologies and concepts to explain the change to the greater American public.” This deeper understanding should include research into the causal processes that lead to systemic effects. As an example, what is the relationship between innovation (at the community level) and the resilience or robustness of a community, neighborhood, or urban area? Or, can we better link how the arts connect to quality of life? And, in addition to tracing real effects, it would be useful to measure people’s perceptions of how art and culture contribute to community life. In addition to demonstrating how cultural practices or institutions benefit communities, others felt that it is essential to deal with the question of opportunity costs. Are investments in cultural practices more effective than investments in other solutions when faced with pressing community problems? There was considerable discussion about a systems approach to understanding the cultural sector. Can we better explain the patterns and interconnections among cultural organizations and artists? What institutions and resources, and what relationships among them, are most strongly related to innovation? How can we explain the diversity or heterogeneity of the system? And, what are the consequences for communities of differing degrees of diversity in a cultural system? We also need to know more about the “policy system” – the actors and institutions that shape the environments within which artists work – in order for funders and policy makers to know if they are pulling the right levers in their efforts to induce change. One suggested research strategy is to create a macro model that could trace the supply and demand of arts and cultural programming (using data on artists, arts and cultural organizations, and audiences) over two or three decades. Such a model might also include information about community health, economic growth and demographic change. (Much economic and demographic data now can be mapped easily to local levels from nationally collected data sets using GIS.) . We need good comparative research (field experiments or studies of several communities over time) would make it possible to understand better the causal connection between cultural activity and other outcomes. . Field experiments should focus on communities where it seems plausible that a small intervention can have a discernible effect. Following this idea, it was suggested that it may be more practical to start by studying urban systems of modern scale – perhaps four or five smaller cities like Trenton and Sacramento – than to attempt studies of the largest metropolitan areas (e.g., New York or Chicago). Collaborating research teams in each city would collect data on a range of important dependent variables (indicators of community cultural health) such as creativity and innovation; heterogeneity and diversity of cultural practices and organizations; arts participation; economic health, and social capital. Next, researchers would identify and enumerate the parts of the arts system that might be influential. These would include (1) artists (2) arts organizations (including nonprofits, unincorporated and embedded institutions, sources of support and infrastructure, and commercial cultural enterprises); and (3) the public (with information on demographics, arts participation, and social and cultural attitudes). In order to identify causal effects over a time frame adequate to discern them, such a ambitious project would collect data every year (or every 2 years) in each of the sites for at least 10 years. Using these data one could employ a variety of modeling strategies. Such strategies would include
This type of study would get beyond the current piecemeal efforts to understand the cultural sector. Although the resources to sustain this large a study may not be available, a beginning strategy would be to undertake a piece of it in one location, with the hope that the results will be of such striking value that they will catalyze additional efforts. Many important lines of research in the social sciences begin with a paradigmatic study that generates publicity and stakes out important methodological and theoretical ground. (Robert Putnam’s early work on the decline of civil society is an excellent example.). These critical studies, which may be modest in size, lead to larger efforts because they set an agenda for a research community, with scholars drawing on similar data, methods and theories to critique, refine and advance important ideas. Concern was expressed that too few talented scholars are taking up these issues. Academics and funders must provide incentives for young scholars to study cultural policy, and also work hard to package good research so that policy makers will draw upon the most rigorous studies, rather than the more superficial studies that are easier to digest. Perhaps the National Arts Journalism Program (NAJP) and the Cultural Policy and the Arts National Data Archive (C-PANDA) should collaborate to sponsor a presentation once a year to introduce journalists to the most important new research findings that relate to the cultural sector. It was noted that research on metro areas should be alert to the fact that different groups within a community employ the arts and culture in different ways. The “impact” of art and culture will be different in immigrant communities than in non-immigrant communities, and may differ as well between urban/suburban residents, and by race and ethnicity. It is important to be careful in defining “communities” in order to examine how cultural support systems work for different populations. Finally, several participants felt that we still had not adequately defined the problems that these models and methods are intended to answer, and suggested the importance of paring down the number of research agendas and questions. They argued that the field needs to come to preliminary agreement on a researchable problem, establish common vocabulary or indicators, identify a small number of research questions and then marshal the funding and scholarship to mount a paradigmatic study. Finally, there was discussion about how to use mapping software (like GIS – Geographical Information Systems) to create data sets that link the arts and culture to community-level indicators. For example, at Berkeley, scholars are creating the Electronic Cultural Atlas Initiative to map all cultural activities and organizations in Berkeley. Doing this on a national scale, across multiple communities, would produce a powerful data set. For example, one might be able to identify an “innovation index” for each community and link this to community demographics and organizational, political, and economic characteristics. The GIS allows you to link cultural information with other community level information in a particular space or geocode. The Unified Database of Arts Organizations produced by the Urban Institute and the National Assembly of State Arts Agencies has been equipped to support geocodes. |
||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
