NATIONAL GAMBLING IMPACT STUDY
COMMISSION

RESEARCH AGENDA

(As Unanimously Adopted by the Commission on 10/31/97)

Introduction

The Subcommittee prepared a comprehensive list of 42 specific policy questions for research at its August 14 meeting; these questions rose out of the 6 broad areas specified in the NGISC Act (PL 104-169) Section 4(a) paragraph 2. These questions were approved by the full Commission on August 20, 1997. The Subcommittee then met with research staff (Peter Reuter, Principal Research Consultant; Tim Kelly, Research Director) on October 2. Drs. Reuter and Kelly drafted an agenda, which was reviewed by the Subcommittee at its October 18 meeting in Denver. This further revised document reflects the Subcommittee's recommendations in light of funds availability and the statutory time limits.

The agenda focuses on developing a strong factual base, both for the Commission in making its recommendations and for state and local decision makers later. It accepts the inherent limitations of research in answering broad cause and effect questions, particularly for a phenomenon as complex and emergent as gambling. It emphasizes those issues which seem likely to be the most important for the next decade and gives less attention to some other issues which though also important do not seem currently as urgent. Casino gaming, both commercial and Tribal, gets more attention than other forms of gambling but no legal form is neglected.

Major Research Tasks

The analytic agenda, will be built around four major streams of research, representing the different levels of data collection: individual and community; players and industry. These represent natural foci for empirical research.

(1) National Survey of Gambling Behavior ¹

A national survey of gambling behavior would be a major contribution by the Commission. The last one was conducted in 1976 by the Commission on the Review of the National Policy toward Gambling and it is seriously dated. Many questions concerning the extent of gambling; the distribution of gambling by income class, race and gender; the relationship of gambling to other behaviors (such as family stability and education) can only be answered with such a survey. The 1976 survey also provided an estimate of the prevalence of pathological gambling; the new survey should update that and permit some assessment of differences in this rate among jurisdictions with varying levels of legal gambling available, though the precision of the resulting estimates will necessarily be determined by the sample size.

(i) National surveys of high quality are extremely expensive, because of the difficulty of obtaining a high response rate from a truly representative sample, and the preliminary research required to develop a survey questionnaire that captures all the behaviors of interest in a parsimonious fashion. The Subcommittee recommends that the Commission consider conducting a survey with in-person rather than telephone interviews if this is necessary to provide a strong set of estimates and analyses. ²

The Subcommittee recommends that the staff should begin immediate inquiry into the options for the conduct of such a survey, as well as discussions with other federal agencies that might be willing to provide additional funds. Information is needed as to costs, time for completion and the likely quality of the data under each option. The staff will report back to the Commission Chair and the Subcommittee by the end of November on these options and the Chair in consultation with the Subcommittee will then authorize the development of a Request for Proposals.

Though the Subcommittee enthusiastically supports the proposal for a national survey, it does note the limits of such research. For example, it will not provide data on the association of crime with gambling; though other household surveys do ask about criminal activities, such as drug taking, such questions are time-consuming, reduce response rates and, for predatory and violent crimes, yield low quality data. Moreover the precision of the estimates of problem and pathological gambling is limited by the possibility that problem gamblers are harder to reach in households, less willing to respond to questions and underreport their involvement in gambling. ³

Targeted Surveys. For many reasons it would be desirable to supplement a national survey with additional surveys of populations in communities with locally important forms of gambling and of frequent players, including problem or pathological (p/p) gamblers.

(a) Surveys of communities with casinos or other locally significant forms of gambling (e.g. parimutuel facilities or card rooms). The principal goals of such surveys are to:

(1) estimate the extent of problem/pathological gambling in these communities as compared to the nation generally. This would be done with exactly the same questionnaire and procedures as in the national survey; the cost is finding interviewers and a sampling frame for the specific communities targeted and:

(2) acquire additional data on how individuals and families in these communities are affected by the availability of, and participation in, gambling.

Unfortunately the cost of any survey is largely driven by its sample size. It would be necessary to undertake a substantial number of additional interviews to answer the critical questions concerning these communities. Thus this activity is likely to add very substantially to total survey costs.

The Subcommittee recommends obtaining casino industry data relating to betting patterns (including heavy betting patterns),⁵ the demographics of casino customers in general, and advertising studies and techniques. The subcommittee recognizes that certain privacy concerns of individuals must be respected. A targeted survey based on approaching players as they exit a casino, betting parlor or lottery outlet, will not yield samples which can be described systematically and thus generalized to a larger population. However, such data could be useful as descriptive research intended to provide a snapshot of percentage of p/p gamblers at various locations, as well as helping development of estimates of percentage revenue provided by p/p gamblers versus non-p/p gamblers. The Commission should weigh the cost of such research as well as time limitations and then determine whether or not a review of the current literature on p/p will suffice. Note that this topic is also addressed in the p/p gambling section.

(2) Data Base on Communities.⁶

The Subcommittee recommends that data be collected on a large sample of communities, including most that have introduced casinos in the last decade. The data base will cover the period of approximately 1980 to 1996 and thus include at least five years (and generally 10 years) before the introduction of casinos. For each community, data will be assembled on: (I) Social Problem Indicators including: per capita crime levels (by crime type), suicide rates, divorce rates, child abuse and domestic violence cases brought to the attention of agencies and welfare case loads, (II) Economic Indicators, including: employment levels, unemployment rates, average earnings, government revenues and bankruptcy rates; government expenditures on criminal justice and related services.

The sample will include a matched group of communities that have not introduced casinos. Attention will be paid to the availability of other forms of gambling, such as parimutuel (track and offtrack) and lottery. Thus the sample will consist of perhaps as many as 100 communities (depending on cost) divided into at least four groups: no legal gambling; lottery only; casino only; various combinations including other forms of gambling. It will also include communities that vary substantially by size, rural and urban location; and region of the nation. Some will be selected because they are close to communities that have introduced casinos or parimutuel betting ("feeder communities") They will include communities with different kinds of casinos; large tourist-oriented facilities and others that explicitly cater to predominantly local markets. Communities with Native American casinos will also be included.

Though the vast majority of these data are available from official agencies, the creation of this data base is a major research activity because so much local data is not available from a central source. For example, even crime report and arrest data, in theory available for all jurisdictions through the FBI's Uniform Crime Reports system, sometimes turns out not to be obtainable except by time-consuming contact with the local police department, if then. Mississippi, for instance, does not file local data with the UCR and in Florida and Illinois there were some years in which no local data were provided. For some important measures, like the number of women in shelters because of spousal abuse, no standard data exist. Other measures are not available at the right level for analysis: e.g. bankruptcy data are available only at a county level (from US courts) and that may be a much larger jurisdiction than the town which established casinos.

Case Studies. The data base will be complemented by a smaller number of case studies in a few of the communities in that sample. These case studies will gather much finer data, of the kind that can only be obtained through interviews and visits to the sites. For example, welfare case workers may be interviewed about the extent to which they see gambling as a major factor in their case loads and the basis of their perceptions. Similarly enforcement officials may be interviewed about their deployment of officers around casinos and their perception of the extent to which offenders in their community are heavy gamblers. There may also be focus groups for community members to discuss perceptions of how legal gambling affects the community as whole. It may be possible to obtain data on the extent of enforcement of rules against under-age wagering and the problems that have been encountered in controlling this problem. Additional economic data, for example on the changing characteristics of job opportunities related to the casino industry, may also be gathered. This set of case studies should be published as a separate report but also will inform the analysis of the database; the number of case studies will be determined by cost factors.

Descriptive Analysis. The data base and case studies will initially allow the Commission to describe what has happened in communities with various forms of gambling. This will involve comparisons of how crime rates, law enforcement expenditures, welfare case loads etc. changed in casino towns as compared to changes in towns that serve as feeder communities and others that are quite distant from such facilities. Without claiming to be able to establish cause and effect, it will be possible to determine whether (or what types of) communities with casinos have seen large increases in crime and employment as compared to similarly situated towns and cities. The data base will also provide an important resource for individual communities considering the establishment of casinos by permitting them to examine the experiences of similarly situated towns and cities.

More sophisticated analyses will attempt to determine whether there is any causal relationship between the availability and extent of gambling on the one hand and major economic or social variables on the other. Given how complex it is to explain variations in these outcomes across cities, it cannot be predicted that these analyses will turn up relationships strong enough to allow causal conclusions.

Economic Impact Modeling. A large number of consultant studies of the economic impact of casinos already exist. They have reached widely varying conclusions with respect to specific casino projects. The studies are usually done in the context of advocacy and mostly on behalf of potential developers. A number of them may be carefully executed, well documented and may present a lot of the relevant data and assumptions for estimating the immediate economic impact of casino development and operations. A synthesis, identifying the strengths and weaknesses of different approaches is worth undertaking. However it is not clear whether the Commission should undertake an independent modeling activity, which might well be very expensive.

The problems here may be primarily conceptual rather than empirical. The most important source of difference among the studies is varying assumptions about the extent to which gambling expenditures (player losses) substitute for other recreational expenditures or add to the productivity of labor and capital in such a way as to increase total product (regionally or nationally). Thus the studies differ substantially in their assessment of the multiplier effects on the local and regional economies i.e. the extent to which a dollar of gambling expenditures increases total economic activity. The Commission will seek a review of these issues by a leading scholar in the field of regional economic development who has not previously worked on this topic for any private or advocacy group. One product of that review, to be conducted very promptly, will be a determination of the need for a major additional empirical study as a secondary research topic.

(3) The Industry

The third stream of research is based on industry data. A great deal can be obtained from public agency filings by non-Tribal casinos with the SEC or state regulatory commissions; much information is available for the state lotteries, though from different sources. Little is currently available for Tribal casinos, but the Commission will pursue such data and attempt to ensure that its inquiry is as comprehensive for these casinos as for the non-tribal.

Industry data are the primary source for obtaining data on advertising and marketing practices,⁷ estimating the share of revenues generated by casinos in their home and feeder communities, and detailing the characteristics of jobs in the industry. Industry data will be important for key information such as estimation of total revenues generated by heavy gamblers,⁸ and the quality of specific industry jobs as opposed to non-industry jobs. The Commission will therefore instruct its research contractors to identify all issues for which industry data is required, and will pursue such data with all the means at its disposal.

Industry data will also be helpful in addressing issues raised by the NIGA concerning the differential regulation and management of alcohol, prostitution, and problem gambling in Tribal casinos as opposed to commercial casinos.

(4) Problem and Pathological Gambling

The primary work on this topic will be carried out by a panel created by the National Research Council. That panel, consisting of leading scholars from a wide variety of fields relevant to the study of problem and pathological gambling, will review all the existing literature and data sets, meet with researchers working on this specific topic and commission papers as needed. It will develop, as best as the existing materials permit, estimates of the prevalence, sources and consequences of problem and pathological gambling. Since the NRC proposal uses such a large share of the research budget and the research plan is further developed for this topic than for any other, more detail is provided here about the scope of the NRC work and what the Commission needs to undertake to supplement this area.

The Research Subcommittee developed nine sets of questions related to Pathological and Problem Gambling (see attachment A) before soliciting the NRC proposal. Identifying those issues not dealt with by the NRC proposal provides the basis for decisions about additional research the Commission might undertake in this area, including original data collection.

The National Research Council (NRC) Proposal. The NGISC Act specified that the Commission would contract with the NRC for assistance with research on pathological or problem gambling (sec.7(b)). Accordingly, the NRC has developed a proposed plan of action (see attachment B) which addresses this area based on the Act as well as the Research Subcommittee's nine question sets. The resultant NRC plan lists six possible issues to be addressed, which correspond closely (but not completely) with the nine question sets as follows:

NGISC Research Subcommittee Question Sets	Corresponding NRC Proposal
1. extent of p/p gambling	Issue 1: prevalence data
2. characteristics of p/p gamblers	Issue 1: prevalence data
3. factors affecting p/p gambling	Issue 2: etiology data Issue 5: technology data
4. causes of p/p gambling	Issue 2: etiology data
5. p/p gambling & other ills	Issue 3: comorbidity data Issue 4: outcome data (need to add suicide)
6. percentage of revenue from p/p gamblers	[not addressed]
7. treatment efficacy	Issue 6. treatment impact data (need to define as individual clinical outcome)
8. efforts made to address p/p gambling	[not addressed]
9. a) costs attributable to p/p gambling; b) comparative cost information	a) Issue 4: outcome data b) [comparative data not addressed]

As can be seen, six of the nine question sets on pathological or problem gambling developed by the Research Subcommittee are addressed by the NRC proposal. Of the three areas not addressed, number six (percentage of revenue from p/p gambling) will be researched separately since such data are not readily available to the NRC, number eight (efforts to address p/p gambling) will be added to the NRC proposal, and number nine b) (comparative data) will be dropped since the Subcommittee has determined that the cost of gathering such data outweighs the benefits.

The work of the NRC will be augmented by the national survey of gambling behavior described above. This should substantially strengthen estimates of the prevalence of this problem behavior. In addition, a targeted survey of gamblers exiting gambling locations may be considered as a secondary research initiative in order to provide a snapshot of percentage of p/p gamblers at various locations and help development of estimates of revenue they generated. Also, invited testimony and/or focus group data from p/p gamblers in treatment may be considered as a secondary research initiative to help the Commission understand the personal cost and experience of p/p gambling.

Other Topics

The Commission must undertake certain descriptive tasks, in particular (i) providing a current listing of the forms of gambling available in every major jurisdiction (ii) a data base of statutes and regulations governing these gambling forms,⁹ and (iii) the revenues generated through taxation.¹⁰ Given the unique federal relationship to Tribal casinos, particular attention will be given to laws and regulations related to Indian gaming. These descriptive tasks will be discussed here only briefly because they are straightforward; we propose that they be executed through a contract with the Advisory Commission on Intergovernmental Relations. The full scope of that work will be determined after estimating the costs of high priority primary research areas.

The timeline at the end of this document calls for the immediate commencement of primary research initiatives. However, it is difficult to anticipate all topics that may become important as the work of the Commission progresses. For instance, it is not yet known whether topics such as percentage of revenue from p/p gamblers, or unbiased economic impact modeling, are or are not adequately addressed in the current literature. Furthermore, it is not yet clear how much funding may or may not be available for possible initiatives in such areas. For these reasons, it is proposed that the Commission move ahead immediately with a primary research agenda with the understanding that secondary projects may well be called for as work progresses.

Internet gambling will likely be a very important area in the future, particularly for federal regulation, and it is prominently noted in PL 104-169. However precisely because it is still so new and changing so rapidly, it may be better dealt with primarily through hearings and the activities of other federal agencies and only secondarily through research. The Subcommittee recommends that a review by an established expert be commissioned before the end of the year, with a delivery date of 3 months. After that review the Commission can make a more informed decision as to whether there may be further research possibilities that should be explored.

The Subcommittee identified a number of topics that it may wish to explore on a secondary basis but did not believe could be examined in much depth with the existing research funds. It thus recommends that the Commission consider funding literature reviews and/or limited original research by the leading scholars in the field on an as needed basis. A reserve of at least $100,000 could be set aside for commissioning of reviews and/or original research projects on specific issues that may be needed as the Commission deliberates. The staff would seek a knowledgeable scholar to carry out such work within, say, three to six months. Reviews could cost between $5,000 and $40,000, and original research could cost significantly more depending on the requisite methodology.

Possible topics include:

Gambling Promotion

The extent to which the promotional activities of lotteries and casinos affect the volume and distribution of wagering across demographic groups can only be studied by collecting detailed data on advertising by media, placement of billboards and the content of messages. Unless other funding can be found, this will be addressed only through a synthesis paper, which should be able to draw on the extensive literature that has developed on cigarette and alcohol promotional activities.

Organized Crime

There is no promising research approach to studying this topic, which is more investigative in nature. The Commission, in addition to hearings, may wish to commission a survey of what is known concerning the last twenty years; documentation for earlier eras is widely available. Invited testimony from law enforcement officials might also be considered.

The blurring boundary among gambling industries

It has been noted that state lotteries are offering games with rapid feedback, such as Keno, which are very similar to games offered in casinos. Race tracks in a number of states have become sites for slot machines. The Commission might seek a factual review of the extent of these overlaps and a presentation of the policy issues raised by the blurring of these lines.

Other Issues

The Subcommittee recommends that Commission staff should be directed by the chair to approach other federal agencies to identify potential co-funders of Commission projects, particularly for a major national survey of gambling behavior. Among agencies which might have interest are: the National Institute on Mental Health, the National Institute of Justice and the Center for Mental Health Services.

The Commission's contract with consultants or organizations will require that the Commission be given first right to publish results. However the Subcommittee recommends that if the Commission chooses not to publish, then the authors may publish as they wish, though the Commission can require that they not mention the funding source in the publication.

Next Steps

If this agenda is accepted by the full Commission, the staff will begin by (i) developing cost estimates for research initiatives, (ii) developing Requests For Proposals and (iii) identifying a number of suitable contractors who could respond promptly to the RFP. It is likely that the creation of the community data base and the conduct of case studies should be done under the same contract to ensure consistency in the data collection. It may be appropriate to split off the analysis tasks from data collection. A proposed initial timeline for the process is provided below:

Draft primary research agenda (& NRC Contract) developed and approved	Oct 97
Final primary research agenda completed	Nov 97
Primary research contracts developed and let	Nov 97 - Dec 97
Primary research performed, with regular reports to Commission	Nov 97 - Dec 98
Secondary research contracts developed and let (as needed)	Jan 98 - Jun 98
Secondary research performed, with regular reports to Commission	Jan 98 - Dec 98
Final primary and secondary research reports due to Commission	Dec 98
Commission reviews final research reports	Jan 99 - Mar 99
Commission drafts report and recommendations	Apr 99 - Jun 99

Attachment A
(C) Pathological and Problem Gambling

1. What is the extent of problem and pathological gambling in the U.S. today? How many cases of pathological or problem gambling have diagnosed by medical and psychological professionals using methods approved by the appropriate medical, psychological or scientific authorities?



2. What are the demographic characteristics of problem and pathological gamblers? What percentages of gambling customers and of the general population are problem or pathological gamblers?



3. What are the causes of changes in the levels of problem and pathological gambling over time? To what extent does the availability and/or accessibility of legalized gambling affect the levels of problem and pathological gambling? Is there any difference in the incidence of problem or pathological gambling in the general population versus those areas with a major gambling presence?



4. What are the clinically proven causes of pathological or problem gambling, as defined by the appropriate medical psychological, or scientific authorities? What clinically proven methods exist for medical and psychological professionals to accurately diagnose pathological or problem gamblers?



5. What are the symptoms of problem and pathological gambling, and what are the impacts of those systems on the individual, their families, and their communities? Specifically, what relationship, if any exists between problem or pathological gambling and crime? Between problem or pathological gambling and suicide? Between problem or pathological gambling and alcoholism? Between problem or pathological gambling and consumer debt? Between problem or pathological gambling and bankruptcy? What effect, if any do problem and pathological gambling have on employers and the work environment?



6. For each form of gambling, what percentage of revenues is attributable to pathological and problem gamblers? Do the various forms of gambling have different effects on problem or pathological gamblers? Do some forms of gambling as well as particular games encourage or discourage problem or pathological gambling? If so, in what ways



7. What clinically proven treatments exist for problem and pathological gamblers? What is the cost and efficacy of various forms of problem and pathological gambling treatment?



8. What efforts are the gambling industry, various levels of government, and non-profit and community based organizations making to address problem and pathological gambling? What are the most effective measures in reducing problem and pathological gambling?



9. On average, what monetary and measurable costs are directly attributable to pathological and problem gamblers? How do these costs compare with average, measurable costs directly attributable to people with other compulsive behavioral problems whose similarities with pathological or problem gambling are clinically proven and generally accepted by the appropriate medical or psychological authorities?

Attachment B

1. To identify and review relevant data sets and research studies that address the prevalence of pathological gambling. The strengths and weaknesses of this knowledge base will be documented, especially the ability of existing data to support impact studies of pathological or problem gambling on individuals, families, businesses, social institutions, and economy. Where possible, national, regional, and state level data on prevalence will be examined. Major gaps in available data and information will be identified, and date and analysis requirements for designing new studies of prevalence and impact will also be described.



2. To review the research on the etiology of pathological gambling behavior including: early developmental correlates and risk factors; the role of reasoning and decision making processes, and how and when behavioral control mechanisms cease to function in gambling or other behavior; the influence of the gambling environment on the development of problem gambling behavior; the development of cognitive distortions or perceptions that gambling outcomes can be controlled; and the role of personal values and their relationship to the extent of pathological involvement in gambling.



3. To examine co-morbidity with other disorders such as alcoholism, drug abuse, and depression, and to identify data and analysis requirements that could support comparative studies of the individual and cumulative impacts of pathological gambling and these other disorders.



4. To examine existing data on the social and economic characteristics and costs of pathological gambling in relation to a range of negative outcomes including welfare dependency, delinquency, crime and violence, divorce rates, non-support of families, personal indebtedness (including bankruptcies), lost productivity, and compromised family mental and physical health.



5. To assess the impact of gambling technologies on the development and progression of pathological gambling behavior, particularly video, computer, and other electronic games.



6. To assess research on the impacts of available treatment programs, and to identify research requirements and principles for the design of prevention strategies, focusing on the best use of resources to minimize the onset of problem or pathological gambling and to reduce relapse rates after treatment.







Footnotes


¹ Implementing part of item D "An assessment of the impacts of gambling on individuals [and] families".

² Clearly lower cost surveys can be carried out using telephone interviews only and less stringent sampling methods. Such surveys however would: (1) Do less well at obtaining detailed information; telephone interviews rarely exceed 25 minutes, while in-person interviews can be one hour in length. The longer questionnaire may be required to ensure the inclusion of questions that permit assessment of pathological gambling issues. (2) Miss some important populations, since 5 percent of households do not have phones and these are disproportionately poor; moreover they are likely to be systematically different from poor households that do have telephones.

³ In addition, since this behavior is very unlikely to involve even as many as ten percent of the population, it should be noted that the precision of any resulting estimates will be directly linked to sample size.

⁴ This can be achieved by oversampling specific population groups that are believed, on the basis of other information, to have high rates of regular participation of gambling. Thus the sample might include larger numbers from particular states or in particular age or ethnic groups. In subsequent analyses, respondents in these groups have lower weights.

⁵ This will include many who are pathological/problem gamblers

⁶ Implementing elements of Items B and D. B: "An assessment of the relationship between gambling and levels of crime..". D: "An assessment of the impacts of gambling on social institutions...and the impact of gambling on depressed economic areas."

⁷ Implementing part of item D: "An assessment of the impacts of gambling. ...including the role of advertising in promoting gambling".

⁸ The industry data sets will not allow estimation of the share of total expenditures generated by p/p gamblers, because 1) they have neither the incentive nor the capacity to make such diagnosis, and 2) the data are collected by individual casino companies and so do not cover the individual's betting at other casinos.

⁹ Some or all of this area of research may be performed through contracting with the Advisory Commission on Intergovernmental Relations.

¹⁰ Implementing parts of Items A and E among those specified in 4(a) paragraph 2. A: "A review of existing Federal State, local and Native American tribal government policies and practices with respect to the legalization or prohibition of gambling.." E: "An assessment of the extent to which gambling provides revenues to State, local and Native American

---