Policy diffusion and strategies for promoting policy change: Evidence from California local gun...

There is renewed interest in subnational policy change in light of the socalled Devolution Revolution. Scholars have examined both state responses to the devolution of mandated programs from the federal government and state innovations in health care and social welfare programs (Gray, 1994; Leichter, 1997; Staeheli, Kodras, & Flint, 1997; Thompson & Dilulio, 1998). Building upon previous research on innovative state governments (Clark, 1985; Eyestone, 1977; Gray, 1973; Savage, 1978, 1985; Walker, 1969), a number of recent studies also have attempted to identify the characteristics associated with innovative state governments and how interactions between neighboring states impact policy diffusion (Berry & Berry, 1999). The majority of these studies have analyzed morality policy innovations (e.g., Berry & Berry, 1990; Glick, 1992; Hayes & Glick, 1997; Mooney & Lee, 1999; Nice, 1994; Pierce & Miller, 1999).

There has been only limited analysis of innovations developed by local governments that later are adopted by state governments (Mintrom, 1997). Yet, agenda-setting and policy change theorists emphasize that policy ideas can be generated from a variety of actors and institutions within the American federal system (Cohen, March, & Olsen, 1972; Kingdon, 1995). Baumgartner and Jones (1993) make this point explicitly by portraying policy changes as resulting from shifts in policymaking to more receptive jurisdictions ("venue shopping"), coupled with the acceptance of new images of policy problems. All parts of the federal system-Congress, states, and local jurisdictions-can be policy venues; "the very existence of multiple venues means that cracks are likely to appear somewhere and these can be exploited by political activists" (Baumgartner & Jones, 1993, p. 234).

We take an initial step in this article toward filling this gap in the policy research literature by examining local gun control policymaking in California. We select a policy area that has had policymaking activity at the local, state, and federal government levels. In fact, there has been an extremely high level of conflict over the appropriate unit of government for policymaking. Gun rights organizations have attempted to restrict firearm policymaking to state governments by promoting state preemption of local gun regulations and sponsoring state-level legislation requiring issuance of concealed weapons permits (Bruce & Wilcox, 1998a; Kleck, 1997, p. 66). Historically, southern and noncoastal western states have had the fewest state restrictions on gun ownership and sales. Northeastern states, Washington, DC, and California have had the highest number of restrictions (Bruce & Wilcox, 1998b).

In addition, we have selected a policy area that typically is categorized as morality policy. Gun control includes all of the essential components of a morality policy issue: single-issue interest groups, partisan involvement, and constraints on national government action due to a history of state involvement (Bruce & Wilcox, 1998a; Spitzer, 1995, 1998). Although there have been a number of studies about state morality policy innovations, at least some morality policy innovations are likely to be developed and adopted first at the local government level. Morality policy proposals are relatively simple and easily understood, attracting higher levels of citizen participation and local activism than are usually found in other policy areas (Mooney, 1999). Disputes are characterized as conflicts between good and evil (Meier, 1994; Tatalovich & Daynes, 1998). Therefore, interest groups may find it preferable first to seek out receptive local policymaking bodies to build up pressure for state-level policy change rather than negotiate compromises.

We have chosen to concentrate on local gun control policymaking in California. The diversity of California's local governments provides a unique opportunity for in-depth analysis of the dynamics of interest group activity and policy diffusion. Additional reasons for choosing California include its large population, ethnic diversity, and historic role as one of the most innovative states (Gray, 1973; Savage, 1978, 1985; Walker, 1969). The California state government also emerged as a leader in gun control policymaking in 1999, providing a substantive policy concern about the extent to which state innovation has been the product of local policymaking activity.

Research Methodology

We gathered information from gun rights and gun control interest groups on their advocacy efforts within California. We then surveyed cities and counties that formally had considered or adopted gun control measures from 1994 through 1998, when, as we discuss later, state action could not be assured. We developed a list of these jurisdictions from tracking lists maintained by the Association of Bay Area Governments, California Rifle and Pistol Association, Legal Community Against Violence, and Women Against Gun Violence. Interestingly, we found that the California Rifle and Pistol Association had taken credit for deterring adoption of gun control measures in over 20 jurisdictions that had no written record of formally discussing the adoption of gun control ordinances.1 We compiled copies of gun ordinances and minutes of public testimony, supplemented by telephone interviews with elected officials and staff in some jurisdictions. We also contacted regional associations of government.

Fifty-five cities, including one consolidated city/county, considered gun control measures from 1994 through 1998, with 51 adopting more stringent controls. Together, adopting cities represented about 41% of the incorporated 1990 population. An additional six counties adopted controls; adopting counties represented about 51% of the statewide population. Bans on inexpensive handguns known as Saturday Night Specials were the most popular measure (37 adopting cities), followed by additional restrictions on gun dealers (26 cities), requirements for trigger safety locks (24 cities), record-keeping requirements for ammunition sales (9 cities), bans on large-capacity ammunition magazines (5 cities), and special taxes on gun shops (4 new taxes). Several cities also began in 1998 to consider the feasibility of joint legal action to sue gun manufacturers alleging violations of California's consumer protection laws.

In the remainder of this article we analyze factors contributing to local gun control policy innovation and diffusion. We discuss how focusing events and the acceptance of a new image of gun violence helped mobilize support for the adoption of local gun control measures during a period of state policy stalemate. We then consider how the interactions among interest groups, receptive urban governments, and regional associations explain policy diffusion patterns and the characteristics of the adopting cities. Finally, we discuss the most successful strategies and implications for the study of policy change.2

Focusing Events and Interest Group Activity

As defined by Birkland, a focusing event is: sudden, relatively uncommon; can be reasonably defined as harmful or revealing the possibility of potentially greater future harms; has harms that are concentrated in a particular geographical area or community of interest; and that is known to policy makers and the public simultaneously (Birkland, 1998, p. 54).

Focusing events frequently lead to policy change (Cobb & Elder, 1972; Kingdon, 1995) by providing pro-change groups and policy entrepreneurs with a causal story to illustrate that existing policy somehow has failed (Stone, 1989). Furthermore, the immediate media coverage prevents anti-change forces from framing a policy problem or suppressing action (Birkland, 1997, 1998). The impact of individual focusing events depends upon a host of factors including extent of media coverage, event magnitude, and ability to identify remedies. Therefore, focusing events often can be recognized only in retrospect. The units of government that respond and the ultimate policy response are at least partially dependent upon the nature of interest groups and coalitions. If coalitions are not well organized, policy change may not quickly follow events (Birkland, 1997, 1998).3

This was the situation in California at the beginning of the 1990s when Handgun Control, Inc. was the only statewide gun control lobbying group. Thus, focusing events tended to spark interest group formation and the development of policy alternatives rather than immediate, direct change. Only in the City of Pomona did violence lead immediately to action; the city council quickly adopted a previously tabled gun control ordinance after six people were killed in one weekend in 1997. The major gun-related focusing events in California and related outcomes are summarized in Table 1.

New gun control interest groups in California chose to devote a large portion of their time and resources toward the development of local ordinances, since local governments appeared to be more receptive to advocacy efforts than the state legislature. Following adoption of a statewide ban on assault weapons and a firearms purchase waiting period in 1989-90, gun control advocates and a few legislative entrepreneurs unsuccessfully tried to pass additional state legislation through most of the 1990s (Godwin & Schroedel, 1998). The activities that followed two focusing events illustrate how interest groups mobilized and then worked with local governments to develop gun control policy innovations during the mid- 1990s.

A citizens commission was formed in Pasadena after three children were killed on Halloween weekend in 1993. The City of Pasadena subsequently adopted an ordinance requiring registration of ammunition sales early in 1995. Although the ordinance eventually was repealed, several other communities in the Los Angeles area adopted similar ordinances shortly thereafter. The City of West Hollywood passed its own version in March 1995, and its contract attorneys, upon direction of the city council, then teamed with the Legal Community Against Violence to draft the first ban on inexpensive handguns.

The Legal Community Against Violence (LCAV) itself was founded as the result of another 1993 focusing event, the 101 California Street Massacre, named after the address of a San Francisco law firm that was the scene of a shooting. The LCAV, with funding from several foundations, provided pro bono legal services to sue manufacturers of weapons used in the California Street shootings. While dismissal of the lawsuit was being appealed,4 the LCAV offered its services to local governments that risked litigation for potential violations of an ambiguous 1969 state law preempting local firearms regulations. The LCAV successfully defended a 1994 gun dealer ordinance enacted by Lafayette, a city located in the San Francisco Bay Area, and assisted West Hollywood with its defense against a lawsuit from the California Rifle and Pistol Association and other gun-rights groups.5

The LCAV has directly advocated the adoption of local gun control ordinances throughout California by publishing an annual Legal Resource Manual that contains sample ordinances, presents arguments on the validity of local ordinances, and highlights pro bono legal services. Over 2,000 copies of the first 1995 edition were distributed free to local officials. LCAV's involvement in local policymaking has continued through biannual surveys of local government ordinance activity, sponsorship of legal workshops, and member testimony before local governing bodies (Legal Community Against Violence, 1995, 1997, 1998, 1999).

Additional Interest Group Responses to Gun Violence

In addition to focusing events, the cumulative effect of gun violence contributed to growth in the number of pro-gun control interest groups. Law enforcement officials, especially chiefs of police in urban areas, more aggressively promoted state and local gun control measures in the 1990s (California Police Chiefs Association, 1995). Their increased activism primarily was due to the continued high number of police officers killed in the line of duty in California in spite of declining crime rates (State of California, Department of Justice, 1998).

The Los Angeles-based Women Against Gun Violence (WAGV), founded in 1994 with support from the American Jewish Congress, also began mobilizing women negatively impacted by gun violence. Its initial activities included requests to regional associations and city councils to support inexpensive handgun bans and the Victims Remembrance Project, which publicized gun-related deaths in Los Angeles County (Women Against Gun Violence, 1998). The WAGV noticeably changed its strategies when the prospects for statewide and national action appeared to improve in the late 1990s. The WAGV then began publicizing gun violence statistics by state legislative district and interacting more with state and congressional leaders.

Each interest group filled a distinct advocacy role in the local government arena: the LCAV provided legal services, the WAGV represented gun victims and testified at public hearings, local law enforcement provided expertise and internal advocacy, and grass-roots activists lobbied their own local governments. Handgun Control, Inc. maintained its focus on gun violence education and statewide lobbying while providing expertise to the new interest groups. This evolution of interest groups is an intriguing example of policy niche partitioning and interest group ecology (Gray & Lowery, 1996), with interest group activities adjusting strategies over time.

The Acceptance of a New Policy Image

According to Baumgartner and Jones (1993, pp. 25-38), policy change occurs when there are changes in both institutional venue and policy image. Shifts in policymaking to different jurisdictions and the promotion of a new understanding of a problem interact together, creating momentum for change. We found that local government policymaking was facilitated by the acceptance of a new image of gun violence as a public health problem, consistent with Baumgartner and Jones' conceptualization.

Gun control supporters traditionally argued that restrictions were needed to reduce crime and stop the inappropriate use of firearms. Medical researchers in the 1980s began promoting a new image of gun violence-framing it as a medical epidemic. The Centers for Disease Control promoted gun violence reduction measures, and medical journals increasingly published studies on firearm injuries and deaths (Bijlefeld, 1997, pp. 131-161). Many of these studies went beyond purely medical research (Kates, Lattimer, & Boen, 1997). For example, the physician-led Violence Research Prevention Program at the University of California, Davis published an influential study on gun manufacturers in southern California and the proliferation of cheap handguns (Wintemute, 1994).

The public health perspective manifested itself in California by the California Wellness Foundation's heavy involvement in the promotion of local gun control measures (California Wellness Foundation, 1992, 1993, 1994, 1995, 1996, 1997). The billion-dollar foundation was founded in 1991 by Health Net, a health maintenance organization. Its Violence Prevention Initiative has provided funding primarily to nonprofit organizations within California in the areas of youth violence policy, community leadership, community action, and research. The Wellness Foundation publicly designated $35 million for grants from 1993 through 1998; our own tally of projects funded from 1992 through 1997 added up to almost $50 million.6 Over $17 million was allocated for projects related to gun violence reduction, including over $7 million in start-up funds for the Pacific Center for Violence Prevention and several million dollars for a public information campaign. The Wellness Foundation has provided funding to virtually every guncontrol interest group and gun-related research project in California, including grass-roots organizations.

One of the Wellness Foundation's most ambitious efforts was commissioning public opinion polls in critical urban areas. Although the polls included leading questions,7 they appeared to signal a groundswell of public support for inexpensive handgun bans and other restrictions to reduce gun violence (EDK Associates, Inc., 1996). The results were touted through an extensive public relations campaign, under a new organization, Prevent Handgun Violence Against Kids and Legal Community Against Violence (1996, 1997); written materials used the title Communities On the Move to encourage local activism.

The Foundation's 1996 Annual Report makes the connection between local government policymaking, a new policy image of gun violence, and its activities explicit:

Today communities no longer accept the conventional view that violence is simply a crime. In the past four years, Californians have rallied around a preventive health strategy against violence that goes after the main agent of this disease-handguns-and the hosts- and hopeless young people-in ways proven over decades to successfully stem epidemics. The Foundation did not begin this effort; it rose out of communities that have long struggled, often in isolation, to defeat society's most virulent virus. But this initiative did help inform the public debate and embolden communities to take action (California Wellness Foundation, 1996, p. 17).

In contrast, gun rights groups relied upon traditional arguments in opposing local government actions: the Second Amendment right to bear arms, the recreational benefits of guns for hunting, and the deterrent effect of guns as a means of self-defense (Lott, 1998; Spitzer, 1995, p. 8). Representatives from the California Rifle and Pistol Association (CRPA) primarily attended local governing board meetings around southern California, while National Rifle Association (NRA) members attended meetings in the Bay Area. The CRPA's letter to the City of Pomona outlined typical arguments: the inexpensive handgun ban did not address criminal use of weapons, might violate state law, would be costly to implement, and was flawed (California Rifle and Pistol Association, 1996).

Following Cobb and Ross' (1997) typology of interest group strategies, the arguments,used by pro-gun groups included both medium-cost attack strategies (disputing the desirability of the proposed solution) and the high-cost strategy of threatening legal action. Pro-gun groups did not follow more basic strategies such as denying that a problem existed or discrediting the new anti-gun interest groups, which often appeared to have more grass-roots support than either the NRA or the CRPA. Arguments for state primacy, directed most often at state legislatures in other states by pro-gun groups, were not well received by local governing bodies in California interested in preserving autonomy. Although the threat of legal action was somewhat mitigated by the LCAV's involvement, it was effective in some communities. For example, Sierra Madre, with no gun shops, voted down an ordinance after hearing arguments from the CRPA.

In spite of the pervasiveness of the public health image and the inherent weaknesses of the gun-rights arguments at the local government level, gun control supporters achieved success in a relatively small number of communities. No cities in urban Orange County considered or adopted new gun control ordinances during the period studied. San Diego County and the City of San Diego belatedly adopted trigger lock requirements in 1998 after previously voting down inexpensive handgun bans. Even in the most successful region of the state, the Bay Area, only 29 of the 98 cities adopted ordinances. The record for grass-roots grant recipients was mixed: the Pasadena coalition, already working in conjunction with its city council, successfully developed an ordinance, but the Orange County Citizens for the Prevention of Violence was not as successful.

The success of gun control supporters came instead from the types of cities adopting ordinances. The adopting cities encompassed 41% of California's incorporated population and majorities of the populations within the major urban regions, giving the appearance of widespread support and framing the gun control issue for political elites, the media, and the general public. In our next two sections we directly discuss the diffusion process and characteristics associated with local governments adopting gun ordinances.

Regional Associations and Communication Links

The adoption of local gun control ordinances closely resembles gateway innovation, which sets the stage for more large-scale change. It is obvious that adoption of a ban on a specific weapon type by a single jurisdiction will have little, if any, impact on gun accessibility or crime, but "after a certain threshold of accumulated adoptions of gateway innovations has been reached, a sudden takeoff may occur" (Zaltman, Duncan, & Holbek, 1973, p. 45). Therefore, we expected and found that interest groups and first adopters of particular gun control policies encouraged additional jurisdictions to adopt policies.

A policy's perceived benefits and its compatibility with organizational values impact its diffusion (Rogers, 1995). The process of diffusion requires contact between jurisdictions, a transmission process, and acceptance of the policy by additional jurisdictions (Lave & March, 1993). The adoption rate and number of adopters can be maximized through professional associations, which speed up the dissemination of knowledge about a proposed innovation, reduce the risks of failure by collective action, and create bandwagon effects. The presence of regional associations and their relative support of gun control measures help to explain differences in adoption patterns.

In southern California, the cities adopting gun control measures generally were located adjacent to or near the City of Los Angeles. Los Angeles Council member Mike Feuer, with the support of the WAGV, established an informal organization in 1995 specifically to promote local government policymaking on gun control issues. Renamed in 1997 as the Municipal Gun Violence Working Group, participants include interest groups, elected officials, and staff from surrounding communities. The Working Group has held regular strategy meetings and periodically hosted conferences and seminars (Legal Community Against Violence, 1999).8

In addition, West Hollywood officials, members of the Working Group, directly promoted their inexpensive handgun ban ordinance to other Los Angeles County jurisdictions through a December 1995 letter distributed even before final adoption. Members of WAGV also approached local governments and the 26-- member San Gabriel Valley Council of Governments about the ban. The Council passed a resolution in favor of inexpensive handgun bans in October 1996, but only three member cities adopted it. Not even the two members that previously had adopted ammunition registration requirements adopted the ban.

Bay Area associations were more organized and activist. The East Bay Public Safety Corridor Partnership, consisting of law enforcement, city, county, and school district representatives along the Interstate 80 Freeway Corridor, quickly endorsed the inexpensive handgun ban when it was presented by LCAV; the Partnership actively promoted adoption to its membership and solicited commitments to a legal defense fund. Fourteen of 16 city members adopted cheap handgun bans, and 15 adopted at least one gun control ordinance.

The larger Association of Bay Area Governments (ABAG) subsequently held a summit on gun violence and passed a resolution in support of local gun ordinances; in a letter to its members ABAG focused on cheap handgun bans, but also encouraged the adoption of trigger lock requirements and new requirements for gun dealers. We found that the additional Bay Area cities that adopted ordinances tended to be located near the area covered by East Bay Public Safety Corridor Partnership or adjacent to the City of San Francisco, long a center of gun control support and home to the LCAV.

As previously noted, trigger lock requirements were adopted by both the City and County of San Diego in early 1998. There was a mild amount of policy diffusion with three nearby cities adopting similar ordinances shortly after the San Diego Union-Tribune (1998), the regional newspaper, called upon local jurisdictions to adopt trigger lock ordinances.

A summary of city gun control measures by region is included in Table 2. While this table shows the apparent diffusion of gun control ordinances to receptive jurisdictions, it must be emphasized that diffusion was not automatic. The core policymaking areas- the Bay Area and the Los Angeles region-were home to interest groups that worked with local governments to develop ordinances and sought out receptive local governments. The active involvement of informal and formal regional associations in these areas also obviously impacted the number of adoptions and diffusion patterns. The lower adoption rates were in the area covered by the San Gabriel Valley Council of Governments, which supported gun control measures only through a resolution, and in the San Diego/Imperial County region, which had a lower level of interest group activity.

Characteristics of Adopting Cities

To identify the characteristics of adopting jurisdictions more precisely, we created a data set with all California cities, except for 14 cities that incorporated after 1990 and 33 cities that did not record individual crime rates.9 Only one of the missing data cities adopted a gun control ordinance. We used logistic regression (logic) analyses to predict which city governments would adopt gun control measures.

In our first analysis we included variables designed to identify the larger, more densely populated and established jurisdictions that were the center of interest group activity. We included regional variables to identify the cities from East Bay Public Safety Corridor Partnership, San Gabriel Valley Council of Governments, and San Diego/Imperial Counties, along with cities adjacent to the East Bay, San Francisco, and Los Angeles. We also hypothesized that cities with certain demographic characteristics would be more supportive of gun control: higher Democratic party registration, similar to partisan differences on gun control issues at the state and national government levels; more highly educated and higherincome populations, in accordance with previous research on local government innovation and public entrepreneurship (Nice, 1994; Rogers, 1995; Schneider & Teske, 1992; Schneider & Teske with Mintrom, 1995); and higher proportions of racial/ethnic minorities, as indicated by previous innovation research and research on state legislative support for gun control (Barrett, 1995; Flanagan, Cohen, & Brennan, 1993). Finally, we hypothesized that contextual variables such as whether a city was home to a gun manufacturer, a city's murder rate, and whether a city had its own police department would be relevant.10

The results are presented in Table 3. Since the data set comprises the universe of cities, rather than a sample, tests of statistical significance are not as relevant. The odds ratio, or Exponent (B), provides an interpretable measure of the relative importance of the variables' in logistic regression models. Therefore, the following discussion emphasizes interpretation of the odds ratios.

As expected, the regional variables were the strongest predictors; the relative odds of being an adopting city was highest for the cities that were members of the East Bay Public Safety Corridor Partnership, but the odds also increased for cities within the other identified regions. The most populous cities were more likely to adopt gun control ordinances, but the odds ratio was close to the base of 1.0, indicating that this factor was relatively less important compared to region. Population density, Democratic party registration, and percentage of college-educated populations also were associated with higher odds ratios. Although median household income was significantly correlated with education level (Pearson's R = 0.767),11 surprisingly it is not a significant predictor in the logistic regression model nor does its inclusion impact the odds of being an adopter. The racial/ethnic composition variables were relatively unimportant in this model. Finally, newly incorporated cities (negative relationship between later incorporation year and adoption) were less likely to adopt gun control ordinances.

The results for the contextual variables were mixed. Three cities with gun manufacturers actually adopted gun control measures, contrary to expectation. The murder rate was in the predicted direction, but higher murder rates did not increase the odds of being an adopter. We ran additional analyses using murder rates from other years, multiyear averages, and substituting violent crime rates, with the same result. Interestingly, this finding is similar to that of a recent interstate study, which also found insignificant relationships between crime rates and gun control support (Bruce & Wilcox, 1998b). Cities with their own police departments actually were less likely to adopt gun control measures, but this result should be interpreted cautiously, given the large standard error and the effect of the regional variables.

Overall, the model correctly predicted the outcomes in 93% of cities. Twenty-nine (81%) of the 36 predicted cities actually adopted ordinances. The model underpredicted the number of adopting jurisdictions, capturing 29 (58%) of 50 adopters, with one adopting jurisdiction excluded due to missing data. The chisquare value of 164 (with related -2 Log Likelihood value of 143), pseudo R^sup 2^ values, Lambda PRE (proportional reduction of error), and Goodman-Kruskal Tau C values indicate that the statistical model has fairly high predictive power. Given the skewed distribution of the dependent variable, particular attention should be paid to the Goodman & Kruskal Tau C score, which is the most appropriate goodness of fit test and has a value of 0.89.

We performed a second analysis without the regional variables, presented in Table 4. The predictive power of the model was reduced; 20 (77%) of 26 predicted cities adopted ordinances, and the model further underpredicted actual adoptions, capturing only 20 (40%) of 50 adopters. The values of the predictive power statistics also were lower. For example, the proportional reduction of error was only 28%, as compared to 44% of the model with regional variables.

However, this statistical analysis yields clearer findings about the impact of racial and ethnic composition. Without the regional variables, it is apparent that the odds of being an adopter increase for cities with higher proportions of African-American and Asian-American populations and decrease slightly as the percentage of population that is Hispanic increases.

The adopting cities tended to have interesting combinations of Democratic and educated populations with higher percentages of Asian-American populations. Several Bay Area cities with high Asian-American populations adopted ordinances, as well as Monterey Park in Los Angeles County. In addition, several Los Angeles County cities with concentrations of African-American residents adopted ordinances. Finally, there are high proportions of Hispanics in the Central Valley region and in several southern California communities that did not adopt gun control ordinances.

We were precluded from using event history analysis and other time series methodologies that can compare early and late adoptions and use proximity to previous adopters as a variable. Most adoptions of the most heavily promoted measure, the ban on inexpensive handguns, took place within a very short time frame, from early 1996 through early 1997, and differences in adoption order often were due simply to differences in procedures for scheduling action on city council agendas. We also found that the smaller number of adopting cities for the other types of gun control measures resulted in skewed distribution problems when attempting to model adoption patterns. Through a close examination of our data, we did find that early adopters of Saturday Night Special ordinances had somewhat higher murder rates. As discussed earlier, areas outside the main focus of interest group activity, such as San Diego, considered gun control measures later and adopted few types of ordinances. We found no other differences in our substantive findings.

Conclusion

The implementation of state term limits in the late 1990s resulted in the election of several new state legislators who had served on city councils that had adopted gun control measures. These legislators, along with gun control groups, played leading roles in promoting gun control in the state legislature during the 1997-98 session. Gun control became a major 1998 campaign issue after gubernatorial vetoes of an inexpensive handgun ban and trigger lock requirements, both based on local government ordinances. With the election of a supportive governor and increased anti-gun legislative majorities, the state government adopted these proposals in 1999, along with a revised assault weapons ban and a one gun per month sales limit. The new attorney general also promoted a more cooperative relationship with local governments on policy implementation of gun control measures.

We have found that the exploitation of a new image of gun violence as a public health problem and the strong financial support of the California Wellness Foundation strengthened the influence of gun control groups. Yet, it would be a mistake to assume that gun politics in California have been shaped solely by financial and organizational resources. The number of local governments adopting gun ordinances was sufficient to build pressure for statewide change, but total adoptions fell well below a majority of cities, even in the most targeted regions. The strategies used by interest groups, focusing events, policy framing, and communication links all had impacts upon which local governments adopted particular policy proposals. The activism of regional associations further affected adoption rates and diffusion patterns.

The ability of interest groups and policy entrepreneurs to engage successfully in policymaking at the local government level depends upon the existence of a critical mass of receptive local governments. Our analysis identified characteristics such as population density, education, year of city incorporation, political culture, and ethnic makeup, which were associated with cities adopting gun control ordinances. The relative unimportance of murder rates suggests that demographic and political variables can serve as a starting point for analysis of other types of policy innovations. There may be key regions or types of local governments that regularly serve as leaders in the initiation of new policies in a variety of areas.

The following factors play important roles in the development and diffusion of local policy innovations that later are adopted by state governments:

1. Circumstances or focusing events that resonate among local government leaders and policy entrepreneurs.

2. The establishment of new interest groups that provide a fresh face to policymakers.

3. The promotion of a new policy image that provides: (a) a new perspective on the problem, (b) new supporters, and (c) new financial and organizational resources.

4. The presence and strength of regional associations that can actively promote policy adoption and diffusion within particular geographical areas.

5. The ability of interest groups and entrepreneurs to target activities correctly toward receptive local governments. Resources should be targeted toward an influential number of governments-universal adoption is not necessary to create pressure for statewide change.

Our results support Baumgartner and Jones' view that the creation of new policy images and involvement of new institutional venues are linked to policy change. We suspect that there will be less obvious observations for policy areas that have not had an extensive history of controversy at the state level, thereby attracting the attention of outside interest groups and financial resources. Nevertheless, we believe that there are a number of policy proposals, especially those in the morality and social regulatory policies arenas, that are adopted at the local level before being considered by state and federal governments, meriting additional research on this subject.