|
Article Excerpt
Background and study aims
The project involved an interdisciplinary study of the U.S. national alcohol policy development process (Greenfield et al., 2004a) (this issue). The researchers studied various alcohol policies, including availability and consumer information measures, such as beverage alcohol excise taxes, health warnings, and advertising regulation. Researchers at Berkeley's Alcohol Research Group (ARG) and the former Addiction Research Foundation (ARF) in Toronto conducted in-depth semistructured interviews of key policymakers and stakeholders, including alcohol producers and distributors, advertisers and broadcasters, national organizations, researchers, consultants, government officials, lobbyists, Congressional staff, journalists, and consumer advocates. Working with a political scientist consultant, Dr. Bruce Bimber, we examined existing theory on policy formation, especially the conceptualizations of John Kingdon (1995), to assess the relevance of these conceptualizations to the alcohol policy case and qualitatively gauge their fit with the key informants' perspectives.
To examine the role played by mostly written research-related materials in policy development, selective documentary analysis also traced the use of such materials in legislative debates, in hearings, and in trade and public health newsletters (Anglin et al., 2000). Other research has considered the role of public opinion in the policy process (Greenfield et al., 2004b, this issue) and in national news media (Lemmens et al., 1999). In the marketplace of policy ideas vying for priority and legislative action, many promising policy proposals are sidelined before they can be implemented and their effects assessed. Studying how policy proposals and bills are brought forward, moved, marked up, and enacted or tabled may yield means for improving this process. In addition to studying the policymaking process and use of science, the three-year project's aims included dissemination of findings in hope of informing the public-health-policy community of ways to promote more effective, balanced, and scientifically supported policies in the practical world of U.S. federal politics (Greenfield et al., 1999).
As described in greater detail by Greenfield et al. (2004a) (this issue), our analysis of alcohol control policy development process in the U.S. began with political scientist John Kingdon's (1984, reissued in 1995) empirically based theory of the federal policy formation process developed from both prior theories and interviews he conducted with policy communities involved in federal health and transportation domains. To summarize, Kingdon proposed three distinct but interactive "process streams" simultaneously running through the policy development process. He termed these three (a) problem recognition, (b) the formation and refining of policy proposals, and (c) politics (abbreviated from problems, policies and politics). The first involves agenda setting, fluctuations in attention given to changing issues; the second focuses on actions of the policy community, specialists in and around government who craft policy alternatives; and the third is the political stream, encompassing such elements as swings in national mood, shifting public opinion, changes of administration, changes in Congressional representation with elections, and interest-group campaigns (Kingdon, 1984). Kingdon argued that when streams become "coupled," "policy windows" on a given topic briefly open. When opening, legislation has a chance of being moved or even enacted, but when the window closes, there is frequently a long period of latency when little further can occur on related policies. Within each stream are actors, individuals and groups, who may participate in several streams simultaneously. The actors possess, or are viewed by other actors as having, varying resources like political expertise, constituency strength, or ability to secure votes. The essence of Kingdon's theory is that favorable circumstances in each of the three streams are the necessary but not sufficient condition for policies to be moved forward and enacted. He says that if coupling does not occur quickly, and at the opportune time, the chance for action is forfeited.
The present paper addresses the larger policy study's first aim. Simply put, we asked a systematic sample of alcohol policy community members in and around federal policymaking their views of policy dynamics in order to explore the fit between these key informants' observations on alcohol policy dynamics and Kingdon's conceptualizations based on his studying policymaking in the general health domain. We wished to learn from these accounts specific strategies (good or bad) and process insights that could be useful to future alcohol policy efforts, findings that might ultimately improve the likelihood of the passage of sound, scientifically supported federal alcohol policies. Topics covered in the companion papers--specifically the roles of scientific information, public opinion, and alliance dynamics in the policy process--are dealt with only tangentially here.
Methods
This paper draws on qualitative data gathered for the study of the federal alcohol policy development process in the United States. An extensive description of the methods for this study is found in Greenfield et al. (2004a) (this issue). The project selected for scrutiny four primary and several secondary alcohol policy issues: the warning-labels legislation passed in 1988, the Sensible Advertising and Family Education (SAFE) Act, excise taxes on alcohol, and reauthorization legislation for the Substance Abuse and Mental Health Services Administration (SAMHSA) and, more specifically, actions regarding one of its component agencies, the Center for Substance Abuse Prevention (CSAP), which had come under fire from the alcohol industry.
Data include 64 in-depth interviews conducted with key policy community members and archival material relevant to the selected policy issues. Key informants were chosen using a two-phase sampling strategy. The research team identified a core of prominent policy community members. Main groups tapped included legislative and executive branch personnel, public health advocacy and business interest group representatives, and researchers and the media, whose stands overall were deemed to represent a sufficient spectrum of positions. When a primary subject was unavailable or refused an interview, a secondary subject was chosen from among names collected in the course of locating and contacting the primary choice. The second stage of sampling consisted of identifying gaps in interview data as indicated by analysis of archival and preliminary interview data.
A total of 25 federal (15 legislative and 10 executive branch), 32 interest group (16 business and 16 advocacy), five research, and two media sector respondents were interviewed. The legislative branch sector was defined as Congress and its offices; the executive, as the Presidency and agencies of the administration; business interests were anyone who works for an alcoholic beverage producer, distributor, or retailer, in advertising or broadcasting, or for a trade group or other industry organization; public health advocates and activists included independent consultants or lobbyists who work for activist organizations; researchers were those who engaged in ongoing collection and analysis of data at educational or research institutions and publishing results, with membership in professional associations; and media were those who cover alcohol-related stories for TV, newspapers, or radio, excluding those working for the advertising or broadcast industries as lobbyists.
Completed interviews were transcribed, checked for accuracy, rendered anonymous, and entered into a text management database (NUD*IST[R]; Richards & Richards, 1991). A codebook was developed for indexing transcripts, and codes were applied either mechanically or electronically (by searching for and coding specific text strings) to database documents. Team members read interview transcripts and wrote theoretical memoranda on interview process and emergent findings, with memoranda also included in the text database. Further analysis of data involved iterative reading and summarizing of text coded at the nodes defined below (one document is one interview or set of theoretical memos written on an interview (Denzin & Lincoln, 1994); the text unit is the paragraph, as transcribed). For this paper on policymaking, the following nodes were used to select text for analysis (with the "yield" of the search noted parenthetically):
* Politics: Talk about the role of formal decision-making and political power as an element in policymaking, one of Kingdon's policymaking streams; public policy proposal and refinement. From Kingdon's model: turnover in Congress, staff, administration as it affects policy process; national mood; R's characterization of the general dynamics of the policymaking process; back-and-forth compromise processes with respect to legislation. Steps and conditions to make policy a success; policymaking process, dynamics, conditions for success or failure of a policy; political credibility, prestige....
|
