gambling and problem gambling in mississippi · problem gambling uses this term to indicate all of...

Gambling and Problem Gamblingin Mississippi

A Report to theMississippi Council on Compulsive Gambling

Rachel A. VolbergGemini Research

Social Science Research CenterJanuary 1997

Social Research Report Series 97-1

Arthur G. Cosby, Director

Melvin C. Ray, Interim Vice President

Donald W. Zacharias, President

Arthur G. CosbyMississippi State University

While legalized gambling has become a growth area in both the American and Mississippieconomies, there has been a remarkable lack of research designed to monitor, evaluate and assessthe extent and consequences of gambling behaviors. Dr. Rachel Volberg’s Gambling and ProblemGambling in Mississippi is of special significance since it is the first comprehensive assessment ofgambling behavior in the Mississippi population.

This important benchmark study can trace its origins to Governor Kirk Fordice’s 1995Mississippi Public Policy Think Tank. This Think Tank, with the support of the Mississippi CasinoAssociation, brought together leaders from local and state governments, the gaming industry, thetreatment community, universities and other groups interested in gambling. Out of theirdeliberations came a call for the establishment of a council on compulsive gambling in Mississippi.Subsequently, the Mississippi Council on Compulsive Gambling was established in 1996 under theleadership of Executive Director Dick Lankford, with funding from both the Mississippi legislatureand the Mississippi Gaming Association. An objective assessment of gambling and problemgambling among Mississippians became one of the Council’s first objectives. During the fall of 1996the Council, in cooperation with Mississippi State University’s Social Science Research Center,commissioned a survey entitled The 1996 Mississippi Survey of Gambling and Problem Gambling.This survey provided the data for the current study.

Dr. Volberg is uniquely qualified to address issues of gambling prevalence and problemgambling. She has earned a reputation as the leading scholar on these topics. Dr. Volberg has ledor participated in major gambling prevalence studies in Washington State, Georgia, Iowa, Louisiana,New York, North Dakota and New Zealand as well as in a number of Canadian jurisdictions. Thisreport benefits from the expertise and experience gleaned from her distinguished career ofresearching gambling behaviors.

Dr. Volberg’s report describes the characteristics of gamblers and non-gamblers inMississippi and details their preferences for types of games and locations. Extensive, comparativeinformation that contrasts the nature and extent of gambling in Mississippi with that in other states,especially those of our southern neighbors of Louisiana and Georgia, greatly expands the utility ofthe report. Dr. Volberg should be commended for her fine contribution to our understanding of thisimportant aspect of Mississippi life.

F O R E W O R D

Table of Contents

ACKNOWLEDGEMENTS

INTRODUCTION .................................................................................................................. 1Background ...................................................................................................................... 2Defining Problem and Pathological Gambling ............................................................... 2

METHODS ............................................................................................................................. 3Questionnaire .................................................................................................................. 4Sample Design ................................................................................................................. 4Data Analysis and Reporting .......................................................................................... 5

GAMBLING IN MISSISSIPPI .............................................................................................. 6Gambling in the General Population .............................................................................. 6Patterns of Gambling Participation ................................................................................ 7Expenditures on Gambling ............................................................................................. 9Summary ........................................................................................................................ 11

PROBLEM AND PATHOLOGICAL GAMBLING IN MISSISSIPPI ................................. 12Lifetime Prevalence ....................................................................................................... 12Current Prevalence ....................................................................................................... 13Natural Recovery ........................................................................................................... 14Comparing Problem Gambling Across States .............................................................. 15Summary ........................................................................................................................ 16

COMPARING NON-PROBLEM AND PROBLEM GAMBLERS IN MISSISSIPPI .......... 17Demographics ................................................................................................................ 17Weekly Gambling .......................................................................................................... 17Expenditures.................................................................................................................. 19Legal versus Illegal Gambling ...................................................................................... 21Prevalence by Type of Gambling................................................................................... 21Other Significant Differences........................................................................................ 22Summary ........................................................................................................................ 24

COMPARING MISSISSIPPI, LOUISIANA AND GEORGIA ............................................ 25Comparing the Questionnaires ..................................................................................... 26Comparing the Samples ................................................................................................ 27Comparing Gambling Involvement............................................................................... 28Comparing Problem Gambling Prevalence .................................................................. 30Summary ........................................................................................................................ 32

SUMMARY AND CONCLUSION ....................................................................................... 33Summary ........................................................................................................................ 33Directions for the Future............................................................................................... 35

REFERENCES ..................................................................................................................... 36

APPENDIX A: Methods to Assess Problem Gambling in the General Population .......... 39

List of Tables and Figures

Table 1: Comparing the Actual Sample, Weighted Sample and the General Population ....... 5

Table 2: Demographics of Gamblers in Mississippi .................................................................. 8

Table 3: Monthly Expenditures by Different Groups in Mississippi ...................................... 11

Table 4: Reported Monthly Expenditures on Gambling in Mississippi ................................. 11

Table 5: Comparing Lifetime Problem Gamblers with Non-Problem Respondents .............. 13

Table 6: Comparing Current Problem Gamblers with Non-Problem Respondents............... 14

Table 7: Demographics of Non-Problem and Problem Gamblers ........................................... 18

Table 8: Weekly Gambling of Non-Problem and Problem Gamblers ..................................... 18

Table 9: Average Monthly Expenditures of Non-Problem and Problem Gamblers ............... 20

Table 10: Legal and Illegal Gambling by Non-Problem and Problem Gamblers ..................... 21

Table 11: Mean Starting Age of Non-Problem and Problem Gamblers ................................... 23

Table 12: Other Significant Differences Between Non-Problem and Problem Gamblers ....... 24

Table 13: Types of Gambling Included in Prevalence Surveys ................................................ 26

Table 14: Demographic Characteristics of Respondents .......................................................... 27

Table 15: Demographics of Gamblers ........................................................................................ 28

Table 16: Lifetime Gambling Participation Rates .................................................................... 29

Table 17: Lifetime and Current Prevalence Rates .................................................................... 30

Table 18: Comparing Lifetime Problem Gamblers ................................................................... 31

Table 19: Comparing Gambling by Problem Gamblers ............................................................ 32

Figure 1: Lifetime and Past Year Gambling Participation Rates in Mississippi, 1996 ............. 7

Figure 2: Comparing Lifetime Prevalence Rates in the United States ................................... 15

Figure 3: Comparing Current Prevalence Rates in the United States .................................... 16

Figure 4: Prevalence by Types of Gambling ............................................................................. 22

Acknowledgements

We would like to thank the residents of Mississippi who were interviewed for thissurvey. Their contribution has been vital in adding to our knowledge of changes ingambling and gambling-related problems in the United States. We would also liketo thank the Mississippi Council on Compulsive Gambling for commissioning andfunding this study as well as the staff of the Social Science Research Center ofMississippi State University who conducted the interviews for this study.

Gambling and Problem Gamblingin Mississippi

A Report to theMississippi Council on Compulsive Gambling

Rachel A. Volberg*

Introduction

Until recently, the legalization of gambling hasproceeded apace with little consideration of thepotentially negative impacts that gambling canhave on individuals, families and communities.This study, initiated and funded by the Mis-sissippi Council on Compulsive Gambling, ex-amines the extent of gambling and problemgambling in Mississippi in 1996 and comparesthe findings to similar studies conducted else-where in the United States.

The main purpose of this study is to establisha baseline measure of the prevalence of gam-bling-related problems among the adult popu-lation in Mississippi. Another purpose is toidentify the types of gambling causing thegreatest difficulties for the citizens of Missis-sippi. The results of this study will be usefulin documenting the impact of legal gamblingon the prevalence of gambling problems in thegeneral population in Mississippi. The results

*Dr. Rachel A. Volberg is the President of Gemini Research in Roaring Spring, Pennslyvania.

will also be valuable in the development of pre-vention and treatment services for problemgamblers in Mississippi.

This report is organized into several sectionsfor clarity of presentation. The Introductionincludes a definition of the terms used in thereport while the Methods section addresses thedetails of conducting the survey. The follow-ing three sections detail findings from the sur-vey, with a focus on gambling in general, onthe prevalence of problem gambling in Missis-sippi and on differences between non-problemand problem gamblers in the state. The finalsection compares the results of the study inMississippi with the results of similar surveysin Georgia and Louisiana. The report concludeswith recommendations for the future develop-ment of services for problem gamblers in Mis-sissippi.

2

issue. The Think Tank ended its deliberationswith a long-term strategy to address problemgambling in Mississippi, including the forma-tion of the Mississippi Council on CompulsiveGambling.

The Mississippi Council on Compulsive Gam-bling was initially funded with a grant of$100,000 from the Mississippi Gaming Asso-ciation. Matching funds were provided througha legislative appropriation in 1996. The Coun-cil is presently seeking additional funds fromother private sector sources. The Council is anon-profit organization whose activities focuson education, training, referrals to qualifiedcounselors and healthcare providers, a 24-hourstate-wide helpline and prevalence research(Mississippi Council on Compulsive Gambling1996).

❐❐❐❐❐ Defining Problem and PathologicalGambling

Since the 1970s, legal gambling has become apopular recreational pastime throughout NorthAmerica. In 1974, the first, and only, nationalsurvey of gambling in the United States foundthat 68% of the adult respondents had at sometime wagered on one or more types of legal orillegal gambling (Kallick-Kaufmann 1979). Inthe 1980s and 1990s, studies in different stateshave found lifetime gambling participationrates that range from a low of 74% in Georgiato a high of 92% in New Jersey (Volberg 1994c,1995a). The majority of people who partici-pate in legal gambling are social gamblerswho gamble responsibly, for entertainment andto socialize with friends and family.

The term problem gambling has been usedin different ways. The term is sometimes usedto refer to individuals who fall short of the di-agnostic criteria for pathological gambling butare assumed to be in a preliminary stage ofthis progressive disorder (Lesieur andRosenthal 1991). The term has also been usedto refer to individuals who lose excessiveamounts of money through gambling, relativeto their income, although without reference to

❐❐❐❐❐ Background

Until the 1990s, the only form of legal gam-bling in Mississippi was bingo for charitablepurposes. In 1990, dockside casino gamblingon the Mississippi River and on the Gulf Coastwas approved in a special session of the Mis-sissippi Legislature. Dockside casino gamblinghas expanded rapidly from its origins in Biloxiin 1992. Dockside casinos are now establishedin several areas of the state, including the Tu-nica area, Vicksburg, Natchez, Greenville,Coahoma County, Bay St. Louis, Biloxi andGulfport. There is also a casino owned by theMississippi Band of Choctaw Indians nearPhiladelphia. In addition, there are 129 bingohalls currently operating throughout the state.Both dockside casinos and bingo halls are regu-lated by the Mississippi Gaming Commission.

There are presently 31 casinos in Mississippioffering table games, slot machines, videopoker and keno. In both square footage and ingross gaming revenues, Mississippi hasequaled or surpassed the Atlantic City casinomarket. The casino industry in Mississippi (notincluding the Choctaw casino) generated grossrevenues of $1.72 billion in 1995. Tax revenuesto the state from these operations are estimatedat $122 million while tax revenues to local gov-ernments are estimated at $65 million. Thecasinos employ nearly 30,000 individuals andhave probably generated another 20,000 jobsin construction and ancillary services. In con-trast, charitable gambling or bingo in Missis-sippi generated approximately $620,000 in taxrevenues for the state and provided approxi-mately $9 million to charities for operationsand programs in 1995 (International Gamingand Wagering Business 1996; McKinley 1996;Sheffield 1996).

Initially, there was little attention paid in Mis-sissippi to the issue of problem gambling. In1995, however, the Mississippi Public PolicyThink Tank, organized by the Office of theGovernor and the Mississippi Casino Associa-tion, brought together major stakeholders inthe public and private sectors to develop a con-sensus about how to address this important

3

specific difficulties that they may experience(Rosecrance 1988). The National Council onProblem Gambling uses this term to indicateall of the patterns of gambling behaviorthat compromise, disrupt or damage per-sonal, family or vocational pursuits (Na-tional Council on Problem Gambling 1994).

Pathological gambling lies at one end of aspectrum of problem gambling and was firstrecognized as a psychiatric disorder in 1980(American Psychiatric Association 1980). Re-cent changes have been made to the psychiat-ric criteria for pathological gambling to incor-porate empirical research that links pathologi-cal gambling to other addictive disorders likealcohol and drug dependence. The essentialfeatures of pathological gambling are acontinuous or periodic loss of control overgambling; a progression in gambling fre-quency and amounts wagered, in the pre-occupation with gambling and in obtain-ing monies with which to gamble; and acontinuation of gambling involvementdespite adverse consequences (AmericanPsychiatric Association 1994).

In prevalence surveys, individuals are catego-rized as problem gamblers or probablepathological gamblers on the basis of theirresponses to the questions included in theSouth Oaks Gambling Screen (see Appendix A

for a discussion of the methods used to assessproblem and pathological gambling in the gen-eral population). The term probable distin-guishes the results of prevalence surveys,where classification is based on responses toquestions in a telephone interview, from a clini-cal diagnosis. Respondents scoring three orfour out of a possible 20 points on the SouthOaks Gambling Screen items are classified as“problem gamblers” while those scoring five ormore points are classified as “probable patho-logical gamblers.”

In prevalence surveys conducted since 1990, adistinction is also made between “lifetime” and“current” problem and probable pathologicalgamblers. Lifetime problem and probablepathological gamblers are individuals whohave, at some time in their lives, met the SouthOaks Gambling Screen criteria for problem orpathological gambling. Current problem andprobable pathological gamblers are individu-als who have met these criteria in the past year.Not all lifetime problem and probable patho-logical gamblers meet sufficient criteria to beclassified as current problem and probablepathological gamblers. For example, a middle-aged individual who experienced significantgambling-related difficulties in youth but nolonger has such difficulties would be referredto as a lifetime problem gambler.

Methods

Nearly all of the surveys of gambling and prob-lem gambling completed to date have beenbaseline surveys, assessing these behaviorsin a jurisdiction for the first time. Baselineprevalence surveys provide estimates of thenumber of individuals in the general popula-tion who have experienced or are experiencingdifficulties controlling their gambling as wellas information about the demographic charac-

teristics and gambling activities of these indi-viduals.

Replication surveys permit more precise de-terminations of the impact of new gaming op-portunities on the prevalence of gambling-re-lated problems in a jurisdiction. This infor-mation is useful in planning the expansion ofgaming opportunities as well as in targeting

4

services for problem gamblers and their fami-lies. Replication surveys have been conductedin only a few jurisdictions, including Iowa,Minnesota, New York, South Dakota and Texasin the United States (Emerson and Laundergan1996; Volberg 1995b, 1996a; Volberg andStuefen 1994; Wallisch 1996).

The baseline survey in Mississippi was com-pleted in three stages. In the first stage of theproject, Gemini Research consulted with stafffrom the Mississippi Council on CompulsiveGambling as well as from the Social ScienceResearch Center, the organization responsiblefor data collection, regarding the final designof the questionnaire and the stratification ofthe sample. In the second stage of the project,staff from the Social Science Research Centercompleted telephone interviews with a sampleof 1,014 residents of Mississippi aged 18 yearsand older.1 All interviews were completed be-tween September 25 and October 6, 1996 andthe average length of these interviews was 10minutes. The Social Science Research Centerthen provided Gemini Research with the datafor the third stage of the project which includedanalysis of the data and preparation of thisreport.

❐❐❐❐❐ Questionnaire

The questionnaire for the survey in Mississippiwas composed of three major sections (a copyof the questionnaire can be obtained from theSocial Science Research Center at MississippiState University). The first section includedquestions about 13 different types of gamblingavailable to residents of the state. For eachtype of gambling, respondents were askedwhether they had ever tried this type of gam-bling, whether they had tried it in the past year,how often each month they participated andwhether they participated once a week or more

in this type of gambling. Respondents werealso asked to estimate their monthly expendi-tures on the types of gambling that they hadtried in the past year. The second section ofthe questionnaire was composed of the lifetimeand current South Oaks Gambling Screenitems. The third section of the questionnaireincluded questions about the demographiccharacteristics of each respondent.

❐❐❐❐❐ Sample Design

Information about how survey samples are de-veloped is important in assessing the validityand reliability of the results of the survey.While a fully random design is the most desir-able approach in developing a representativesample of the population, this method oftenresults in under-sampling demographic groupswith low rates of telephone ownership, suchas young adults, minorities and individualswith low education and income. To obtain arepresentative sample for the Mississippi sur-vey, random selection of households and ran-dom selection of respondents within householdswas used.

— Response Rate —

To determine whether a representative samplewas obtained, it is helpful to calculate the re-sponse rate for the sample as a whole. Theresponse or completion rate for the Mississippisurvey was calculated by taking the numberof completed interviews and dividing it by thenumber of completes plus refusals plus par-tial interviews (including terminations by re-spondents during an interview as well as indi-viduals identified as language-impaired orhearing-impaired by the interviewer). Usingthis method, the response or completion rateamong valid respondents for the Mississippisurvey was 70%, which compares well with re-sponse rates for similar surveys in recent years.

— Weighting the Sample —

To determine whether the sample was repre-sentative of the population, the demographics

1Three interviews were completed with respon-dents aged 17 years old. These respondents weredropped from the analysis in order to maintain com-parability with similar studies in other jurisdictions.

5

of the Mississippi respondents were comparedwith information from the United States Bu-reau of the Census. After comparing thesample to the known characteristics of theMississippi population, weights were appliedto the sample to ensure that the sample wouldbe representative of the distribution of thepopulation of the state in terms of gender andethnicity. Table 1 shows key demographiccharacteristics of the sample before and afterweighting and compares these characteristicsto information from the 1990 census.

Table 1 shows that the actual sample substan-tially under-represented males and blacks inthe population. The actual sample also slightlyunder-represented individuals under the ageof 25 in the population. The weighted sampleaccurately represents the population in termsof gender and ethnicity. No effort was made toweight the sample for age since the differencebetween the actual sample and the populationin terms of age was small and since the resultsof complex weighting formulae can be difficultto predict.

❐❐❐❐❐ Data Analysis and Reporting

For easier comparisons of data from the Mis-sissippi survey with other jurisdictions, de-tailed demographic data on age, ethnicity, edu-

cation, income and marital status were col-lapsed to have fewer values. Age was collapsedinto four groups (“18 to 20,” “21 to 29,” “30 to54” and “55 and over”) for purposes of analy-sis. Ethnicity was collapsed from five groups(“Caucasian/White,” “Native American,” “Afri-can-American/Black,” “Asian” and “Other”) intothree groups (“White,” “Black” and “Other”).Marital Status was collapsed from five groups(“Married,” “Separated,” “Divorced,” “Wid-owed,” “Never Married”) into four groups(“Married,” “Widowed,” “Separated/Divorced”and “Never Married”). Education was col-lapsed from five groups into two groups (“Lessthan High School” and “High School Gradu-ate”). Income was collapsed from six groupsinto two groups (“Less than $25,000” and“$25,000 Plus”) for purposes of analysis andcomparison.

Chi-square analysis and analyses of variancewere used to test for statistical significance.In order to adjust for the large number of sta-tistical tests conducted, p-values smaller than.01 are considered highly significant whilep-values at the more conventional .05 level areconsidered significant. In reading the tablespresented in this report, asterisks in the right-hand column of each table indicate that one ofthe figures in the row or in a demographicgroup is significantly different from other fig-ures in the same row or demographic group.

Table 1: Comparing the Actual Sample, Weighted Sample and the General Population

Actual Weighted 1990Sample % Sample % Census %

Gender Male 36.9 46.1 46.4Female 63.2 53.9 53.6

Age 18 - 20 4.3 4.9 7.721 - 24 6.3 6.8 8.025 - 54 59.1 58.8 55.055 and over 30.3 29.5 29.3

Ethnicity White 71.1 67.4 67.5Black 27.1 31.7 31.6Other 1.9 1.0 0.9

6

All survey results are subject to margins oferror. For data based on the total number ofcompleted interviews in this survey (N=1,014),the margin of error is ±3.1% assuming a 95%confidence interval and assuming that the to-tal proportion of the sample responding in oneway or another to the question is relativelylarge. For example, if 50% of all the respon-

dents surveyed answered a question in a par-ticular way, then we can be sure, nineteentimes out of twenty, that if the entire popula-tion of Mississippi had been interviewed, theproportion of the population answering in thesame way would be between 46.9% and 53.1%based on the responses of individuals in thesample.

Gambling in Mississippi

In 1996, legal gambling in Mississippi includeddockside casino gambling, bingo and othercharitable games. Many residents of Missis-sippi also have relatively easy access to salesby the state lotteries in Florida and Louisiana,to racetracks in Arkansas, Alabama and Loui-siana and to riverboat casinos and video pokerestablishments in Louisiana. To assess the fullrange of gambling activities available to Mis-sissippi residents, the questionnaire for thesurvey collected information about 13 differ-ent wagering activities, both legal and illegal.Respondents were asked if they had ever betor spent money on the following activities:

• lottery

• casinos

• bingo

• charitable games

• card games for money not at a casino

• horses, dogs or other animals

• slot machines, poker machines or othergambling machines not at a casino

• bowling, pool, golf or other games of skill

• dice games not at a casino

• stock or commodities markets

• sports events

• the numbers

• any other type of gambling

❐❐❐❐❐ Gambling in the General Population

In every recent survey of gambling and prob-lem gambling, the majority of respondents ac-knowledge participating in one or more of thegambling activities included in the question-naire. In the United States, the proportion ofrespondents who have ever gambled rangesfrom 74% in Georgia to 92% in New Jersey(Volberg 1994c, 1995a). In 1996, 64% of the re-spondents in Mississippi acknowledged partici-pating in one or more of 13 gambling activi-ties. This is the lowest rate of lifetime gam-bling participation identified in any jurisdic-tion where similar surveys have been com-pleted.

Figure 1 on the following page shows lifetimeand past-year participation rates for the typesof gambling included in the 1996 survey. Life-time and past-year participation is highest forcasinos, charitable wagering and lottery prod-ucts. Lifetime and past-year participation arelower but still substantial for card games,sports, bingo, parimutuel events such as horseor dog races and games of skill. Past-year par-

7

ticipation on sports is higher than past-yearparticipation for other types of gambling withsimilar lifetime participation rates.

Respondents who had participated in any typeof gambling in the past year were asked howoften they did this type of gambling in a typi-cal month. Among the most popular types ofgambling, those who wagered on card gamesin the past year did so on average nearly twicea month while those who spent money on chari-table gambling in the past year did so on aver-age once every two months. Those whogambled at a casino in the past year did so onaverage just under once a month while thosewho purchased lottery tickets in the past yeardid so on average just over once a month.

❐❐❐❐❐ Patterns of Gambling Participation

To understand patterns of gambling participa-tion, it is helpful to examine the demographicsof respondents who wager with increasing fre-quency. To analyze levels of gambling partici-pation, we divide respondents into four groups:

• non-gamblers who have never partici-

pated in any type of gambling (36% of thetotal sample);

• infrequent gamblers who have partici-pated in one or more types of gambling butnot in the past year (15% of the totalsample);

• past year gamblers who have participatedin one or more types of gambling in the pastyear but not on a weekly basis (42% of thetotal sample); and

• weekly gamblers who participate in oneor more types of gambling on a weekly ba-sis (7% of the total sample).

Table 2 on the following page shows differ-ences in the demographic characteristics ofnon-gamblers, infrequent gamblers, past yeargamblers and weekly gamblers in Mississippias well as differences in the mean number ofgambling activities these groups have evertried.

Table 2 shows that non-gamblers in Missis-sippi are significantly more likely than gam-blers to be older black women with relativelylow levels of education as well as annual house-

Figure 1: Lifetime and Past Year Gambling Participation Rates in Mississippi, 1996

0%

10%

20%

30%

40%

50%

Cas

ino

Cha

ritab

le

Lotte

ry

Car

dG

ames

Spo

rts

Bin

go

Par

imut

uel

Gam

es o

fS

kill

Gam

blin

gM

achi

nes

Sto

cks

Dic

eG

ames

Num

bers

Oth

er

Lifetime Past Year

8

hold income. The non-gambling group also hasthe highest proportion of widowed respondentsand individuals keeping house. Among thegamblers, infrequent gamblers are most simi-lar to the non-gambling group. Infrequentgamblers are most likely to be older blackwomen with low levels of education and income.This group includes a high proportion of wid-owed and retired respondents.

Past-year and weekly gamblers are more likelythan non-gamblers or infrequent gamblers tobe young unmarried white men, to have com-pleted high school and to have relatively high

annual household income. These two groupsare also most likely to be working full-time orpart-time. Table 2 also shows that the num-ber of gambling activities that gamblers haveever tried increases significantly with in-creased levels of participation.

In general in Mississippi, men are more likelythan women to have wagered on games of skill,dice games, sports, the numbers and cardgames. The only type of gambling that womenare more likely to have done than men is bingo.Respondents under the age of 30 are morelikely to have wagered on games of skill, sports

Table 2: Demographics of Gamblers in Mississippi

Non- Infrequent Past Year WeeklyGamblers % Gamblers % Gamblers % Gamblers %

(N=365) (N=149) (N=425) (N=76)

Gender Male 33.7 53.0 51.7 60.9 **Female 66.3 47.0 48.3 39.1

Age 18 - 20 5.6 2.7 4.7 6.6 **21 - 29 11.0 12.6 20.5 19.730 - 54 43.4 50.6 55.7 45.255 and over 40.0 34.0 19.1 28.5

Ethnicity White 60.1 67.9 72.0 75.4 *Black 38.9 31.4 27.0 23.3Other 1.0 0.7 1.0 1.3

Marital Status Married 53.6 59.3 57.6 57.5 **Widowed 17.1 12.8 4.7 3.7Divorced/Separated 12.0 11.2 14.9 16.5Never Married 17.3 16.7 22.7 22.3

Education Less than HS 24.7 17.0 10.8 4.8 **HS and Over 75.3 83.0 89.2 95.2

Income HH Income < $25,000 58.2 44.1 32.1 34.5 **HH Income > $25,000 41.8 55.9 67.9 65.5

Employment Working 53.0 56..6 73.7 80.2 **Other 43.9 41.5 25.3 19.8Unemployed 3.1 1.9 1.0 —

Gambling Activities — 2.35 3.59 5.09 **

* Significant** Highly significant

9

and card games while older respondents aremore likely to have wagered on stocks and com-modities, parimutuel events such as horse ordog races and the numbers. White respondentsare more likely than black respondents to havewagered on most types of gambling, particu-larly stocks and commodities, card games,horse or dog races and games of skill. The onlytype of gambling that black respondents aremore likely than white respondents to havedone is play the numbers.

Respondents who have never married or areseparated or divorced are significantly morelikely than married or widowed respondentsto have wagered on dice games, games of skill,sports and card games. Respondents with ahigh school education and those with annualhousehold incomes over $25,000 are more likelythan respondents without a high school di-ploma or with lower incomes to have wageredon lottery games, at casinos, on parimutuelevents, on games of skill and on sports as wellas on stocks and commodities.

— Gambling Preferences and Location —

Favorite Type: Respondents who had evergambled were asked to identify their favoritetype of gambling. Among respondents who hada favorite type of gambling (N=444), 54% pre-ferred casino gambling. Preferences for othertypes of gambling were much lower. For ex-ample, only 8% of these respondents indicateda preference for lottery games; 6% preferredwagering on sports and 6% on card games notat a casino; 5% identified sports betting or bingoor charitable gambling as their preferred ac-tivity.

Casino Preferences: Respondents who ac-knowledged going to a casino in the past yearwere asked about their preferences for particu-lar casino products and locations. Among re-spondents who had been to a casino in the pastyear (N=332), there was a slight preference forthe dockside casinos located on the MississippiRiver along the western border of the state.Two-fifths of the respondents who had been toa casino in the past year (43%) indicated that

they preferred to go to a casino in Tunica,Vicksburg, Natchez, Greenville or Coahoma.Just over one-third of respondents who hadbeen to a casino in the past year (35%) pre-ferred to go to a casino in Bay St. Louis, Biloxior Gulfport, and 17% of these respondents in-dicated a preference for the casino in Philadel-phia.

The preferences that respondents who havegambled at a casino in the past year expressfor particular gambling places differ signifi-cantly by place of residence. For example, 98%of respondents who reside in Gulf Coast coun-ties and have gambled in casinos in the pastyear prefer to do so at Gulf Coast casinos while85% of respondents who reside in counties onthe western side of the state and have gambledin casinos in the past year prefer to do so atcasinos on the Mississippi River. Preferencesfor particular casinos are evenly divided amongrespondents who do not live in or near a countywith dockside casinos. Among these respon-dents (N=168), 37% prefer the MississippiRiver casinos, 29% prefer the Gulf Coast casi-nos and 34% prefer to go to some other casinoincluding the casino in Philadelphia.

In contrast to the slight preference for the Mis-sissippi River casinos, there was a strong pref-erence among respondents who had been to acasino in the past year for slot machines. Overthree-quarters (77%) of the respondents whohad been to a casino in the past year indicatedthat they usually play slot machines when theygo to a casino. Only 18% of these respondentsusually play card games and 3% prefer to playdice games when they go to a casino.

❐❐❐❐❐ Expenditures on Gambling

Reported estimates of expenditures ob-tained in this and similar surveys arebased on recollection and self-report.These estimates do not include amountsspent on gambling within a jurisdictionby non-residents and tourists. Data onreported expenditures are best suited foranalyzing the relative importance of dif-

1 0

respondents had participated in this activityin the past year.

— Variations in Expenditures —

Using the approach detailed above, we calcu-late that respondents in Mississippi (N=1,014)spend an average of $41 per month or $490 peryear on gambling activities. It is worth reiter-ating that reported expenditures on gamblingare based on recollection and self-report andshould not be regarded as reflections of actualspending on different types of gambling in ajurisdiction. As in other jurisdictions, thereare statistically significant differences inmonthly expenditures on gambling across de-mographic groups. Table 3 on the followingpage shows differences in the mean reportedexpenditures on gambling by specific demo-graphic groups.

Table 3 shows that men in Mississippi esti-mate that they spend nearly three times asmuch on gambling as women and that respon-dents under the age of 30 estimate that theyspend significantly more than older respon-dents. In contrast to most other jurisdictions,there is no significant difference in expendi-tures on gambling between white and blackrespondents. Not surprisingly, respondentswith lower education and household income re-port spending significantly less on gamblingthan respondents with a high school educationor those with annual household incomes over$25,000.

Table 4 on the following page shows total re-ported monthly expenditures for the sample ondifferent types of gambling in Mississippi aswell as the proportion that each type of expen-diture represents of total adjusted monthly ex-penditures on gambling. Only those types ofgambling for which total monthly expendituresexceeded 1% of the total monthly expenditureare shown.

Table 4 shows that reported expenditures oncasino gambling make up more than half of allreported expenditures on gambling among Mis-sissippi respondents. Reported expenditures

ferent types of gambling among ajurisdiction’s residents rather than for as-certaining absolute spending levels on dif-ferent types of wagering.

To determine expenditures on gambling in thegeneral population, the total monthly expen-diture for each gambling activity is calculatedby summing the amount of money reportedspent by each respondent on each gamblingactivity. The total amount spent in a typicalmonth by all respondents on all gambling ac-tivities is then calculated. The proportion ofthe total monthly expenditure spent on eachgambling activity is calculated by dividing theamount spent on each activity by the totalmonthly expenditure. The total monthly ex-penditure on all gambling activities is dividedby the total number of respondents in the sur-vey to obtain an average amount spent per re-spondent.

— Adjustments to Expenditures —

While stocks and speculative investments arenot universally regarded as a gambling activ-ity, there are people who experience difficul-ties due to their involvement in these activi-ties. For this reason, stocks and speculativeinvestments are routinely included in the ques-tionnaire for gambling surveys. However, incalculating the total monthly expenditure ongambling, expenditures on stocks and specu-lative investments are typically excluded. Thisis done in order to clearly explicate the rela-tive gambling expenditures of the majority ofrespondents. This adjustment is also made toallow comparisons of expenditure data fromMississippi with data from other United Statesjurisdictions.

In every jurisdiction where similar surveyshave been completed, amounts spent on stocksand speculative investments reflect largeamounts of money spent by a relatively smallnumber of respondents. Amounts spent onstocks and speculative investments in Missis-sippi constituted 45% of the unadjusted totalmonthly expenditure although only 4% of the

1 1

on every other type of gambling are muchlower, with none greater than 10% of total re-ported expenditures. Since 77% of the respon-dents who report gambling in casinos in thepast year prefer slot machines, we estimatethat 41% of total reported expenditures on gam-bling in Mississippi are for slot machines atdockside casinos.

As in other jurisdictions, the majority of re-spondents in Mississippi report spendingrather small amounts on gambling per month.Approximately two-thirds of respondents inMississippi (68%) report spending less than $10on gambling in a typical month. Another 23%of the respondents report spending between$10 and $99 on gambling in a typical month,and 8% of the respondents report spending$100 or more on gambling in a typical month.However, this highest-spending group of re-spondents accounts for 80% of reportedmonthly expenditures on gambling in Missis-sippi.

Like weekly gamblers, respondents in the high-est spending group in Mississippi are signifi-cantly more likely to be male, under the age of30 and unmarried than respondents in lowerspending groups. These higher spending re-spondents are also significantly more likely tohave graduated from high school, to be em-ployed and to have annual household incomesover $25,000 than respondents who spend lesson gambling.

❐❐❐❐❐ Summary

In this section, we examined patterns of gam-bling participation in the sample as a whole.In 1996, 64% of the respondents in Mississippiacknowledge participating in one or more gam-bling activities, 49% acknowledge over partici-pating in one or more gambling activities inthe past year and only 7% acknowledge gam-bling on a weekly basis. Lifetime participa-tion is highest for casino gambling, charitablewagering and lottery games. As in other juris-dictions, young unmarried white men withrelatively high levels of education and income

Table 3: Monthly Expendituresby Different Groups in Mississippi

Average MonthlyExpenditure $

Male 62.85 **Female 21.94

18 - 20 79.79 *21 - 29 68.2330 - 54 44.4355 and over 15.19

White 44.23Black 32.05Other 89.42

Married 39.43Widowed 3.47Divorced/Separated 64.75Never Married 48.12

Less than HS 9.19 *HS or higher 47.38

HH Income < $25,000 28.21 *HH Income > $25,000 62.31

Working 45.38Unemployed 1.08Other 34.80


Table 4: Reported Monthly Expenditureson Gambling in Mississippi

Monthly %Expenditure $ of Total

Casino 22,299 53.9Dice Games 3,812 9.2Sports 2,633 6.4Bingo 2,619 6.3Lottery 2,254 5.4Card Games 2,079 5.0Charitable 1,697 4.1Games of Skill 1,667 4.0Parimutuels 929 2.2Other 696 1.7

Total 41,366 100.0

1 2

are the respondents most likely to have evergambled in Mississippi. By a large margin,Mississippi respondents who do gamble preferto do so on slot machines at casinos near theirplace of residence.Casino gambling accounts for over half of thereported expenditures on gambling reported inMississippi. As with gambling participation,young unmarried men with relatively higheducation and income are most likely to report

spending the largest amounts of money on gam-bling. Apart from the much lower participa-tion rates in gambling in Mississippi, the pat-terns of gambling participation identified inMississippi are similar to patterns identifiedin other jurisdictions. In the next section, weturn our attention to the prevalence of prob-lem and probable pathological gambling in thesample.

Problem and PathologicalGambling in Mississippi

Following established criteria for discriminat-ing between respondents without gambling-re-lated difficulties and those with moderate tosevere problems (Abbott and Volberg 1996;Lesieur and Blume 1987), Mississippi respon-dents’ scores on the lifetime and current (past-year) South Oaks Gambling Screen items weretallied. In accordance with these criteria,prevalence rates were calculated as follows:

• lifetime problem gamblers are those re-spondents who score 3 or 4 points on thelifetime SOGS items. In Mississippi, 3.7%(±1.16%) of the respondents scored as life-time problem gamblers.

• lifetime probable pathological gam-blers are those respondents who score 5 ormore points on the lifetime SOGS items.In Mississippi, 3.1% (±1.07%) of the respon-dents scored as lifetime probable pathologi-cal gamblers.

• current problem gamblers are those re-spondents who score 3 or 4 points on thepast year SOGS items. In Mississippi, 2.8%(±1.01%) of the respondents scored as cur-rent problem gamblers.

• current probable pathological gam-blers are those respondents who score 5 ormore points on the past year SOGS items.In Mississippi, 2.1% (±0.88%) of the respon-dents scored as current probable pathologi-cal gamblers.

In the tables that follow in this and the nextsection, lifetime and current problem and prob-able pathological gamblers are grouped to-gether. This approach is based on discrimi-nant analysis that has established a strong andsignificant separation between non-problemgamblers and those who score as problem andprobable pathological gamblers (Abbott andVolberg 1996; Volberg and Abbott 1994).

❐❐❐❐❐ Lifetime Prevalence

According to the 1990 census, the populationaged 18 and over in Mississippi is 1,825,845individuals. Based on these figures, we esti-mate that between 46,400 (2.54%) and 88,700(4.86%) of Mississippi residents aged 18 andover can be classified as lifetime problem gam-blers. In addition, we estimate that between37,100 (2.03%) and 76,100 (4.17%) of Missis-

1 3

sippi residents aged 18 and over can be classi-fied as lifetime probable pathological gamblers.Table 5 on the following page shows that life-time problem and probable pathological gam-blers in Mississippi are significantly morelikely than other respondents in the sample tobe male, under the age of 30 and never mar-ried. This table also shows that lifetime prob-lem and probable pathological gamblers in Mis-sissippi are more likely than other respondentsin the sample to be black, employed and to haveannual household incomes over $25,000 al-though these differences do not attain statisti-cal significance.

❐❐❐❐❐ Current Prevalence

Based on current prevalence and 1990 censusinformation, we estimate that between 32,700(1.79%) and 69,600 (3.81%) of Mississippi resi-dents aged 18 and over can be classified as cur-rent problem gamblers. In addition, we esti-mate that between 22,300 (1.22%) and 54,400(2.98%) of Mississippi residents aged 18 andover can be classified as current probablepathological gamblers.

Comparison of Table 5 and Table 6 on thefollowing page shows that most of the differ-ences between respondents who score as life-

Table 5: Comparing Lifetime Problem Gamblers with Non-Problem Respondents

Non-Problem Problem & PathologicalRespondents % Respondents %

(N=945) (N=69)

Gender Male 45.3 56.8 *Female 54.7 43.2

Age 18 - 20 4.6 9.0 **21 - 29 15.1 25.730 - 54 49.8 48.755 and over 30.5 16.7

Ethnicity White 68.2 56.5Black 30.9 42.0Other 0.9 1.5

Marital Status Married 57.2 45.0 **Widowed 10.9 1.2Divorced/Separated 13.1 18.9Never Married 18.8 34.9

Education Less than HS 16.2 15.9HS and Over 83.8 84.1

Income Annual Income <$25,000 43.4 33.9Annual Income > $25,000 56.6 66.1

Employment Working 63.3 76.5Other 34.8 22.3Unemployed 1.9 1.2


1 4

time problem or probable pathological gam-blers and the remainder of the sample in Mis-sissippi hold true for current problem and prob-able pathological gamblers. The greatest dif-ference is that current problem and probablepathological gamblers are even more likelythan lifetime problem and probable pathologi-cal gamblers to be under the age of 30, divorcedor separated and employed.

❐❐❐❐❐ Natural Recovery

Gambling surveys conducted since 1990 havecollected information on current as well as life-time prevalence rates of problem and probable

pathological gambling. The difference betweenlifetime and current prevalence rates repre-sents individuals who have experienced a gam-bling problem at some time in their lives butdo not score as having a gambling problem cur-rently. Since there are few available treatmentservices for problem and pathological gamblers,these individuals can be regarded as problemand pathological gamblers in natural recov-ery.

The proportion of problem and pathologicalgamblers in natural recovery in the generalpopulation ranges from 29% in New Brunswickto 57% in British Columbia (Baseline MarketResearch 1992; Angus Reid Group and Gemini

Table 6: Comparing Current Problem Gamblers with Non-Problem Respondents

Non-Problem Problem & PathologicalRespondents % Respondents %

(N=964) (N=50)

Gender Male 45.6 56.4Female 54.4 43.6

Age 18 - 20 4.7 7.9 **21 - 29 15.0 32.830 - 54 50.0 45.255 and over 30.3 14.1

Ethnicity White 68.0 55.8Black 31.1 42.1Other 0.9 2.0




Employment Working 63.3 82.2 *Other 34.8 17.8Unemployed 1.9 0.0


1 5

gambling in all of the United States jurisdic-tions where surveys based on the South OaksGambling Screen have been completed. Instates where replication surveys have beencompleted (Iowa, Minnesota, New York, SouthDakota and Texas), the most recent prevalencerates are shown. Figure 2 shows that the life-time prevalence rate of problem and probablepathological gambling in Mississippi is higherthan every other state where similar surveyshave been conducted except Louisiana and NewYork.

Figure 2 shows that, in general, lifetime preva-lence rates are lower in Central and Midwest-ern states than in the Northeast, South andWest. Central and Midwestern states are ju-risdictions where gambling has only recentlybeen legalized. States in the Northeast andWest tend to be ethnically more diverse thanstates in the Midwest and to have had accessto legal gambling for longer periods of time.Like the Northeast and West, states in theSouth tend to be ethnically diverse. However,legal gambling is a recent introduction in allof the Southern states where surveys of gam-bling and problem gambling have been com-pleted.

Figure 3 on the following page shows the cur-rent prevalence rates of problem and probable

Research 1994). Another explanation of thesenumbers is that one in three lifetime problemgamblers in New Brunswick is currently expe-riencing difficulties while one in two lifetimeproblem gamblers in British Columbia is cur-rently experiencing problems.

As in other jurisdictions, a proportion of theMississippi respondents who score as lifetimeproblem or probable pathological gamblers donot score as having a current problem or pa-thology. In Mississippi, 32% of lifetime prob-lem and probable pathological gamblers do notscore as having a current problem or pathol-ogy. Another explanation of this number is thattwo out of every three individuals who haveever experienced gambling problems in Mis-sissippi are currently experiencing such diffi-culties.

❐❐❐❐❐ Comparing Problem GamblingAcross States

The jurisdictions where problem gambling sur-veys have been done in the United States dif-fer substantially in the types of gambling avail-able, in levels of gambling participation and inthe demographic characteristics of the generalpopulation. Figure 2 shows prevalence ratesof lifetime problem and probable pathological

Figure 2: Comparing Lifetime Prevalence Rates in the United States

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

WA CA TX LA MS GA MT MN ND SD IA MD NJ CT MA NY

1 6

other states where similar surveys have beencompleted.

In Mississippi, lifetime problem and probablepathological gamblers are significantly morelikely than other respondents to be male, un-der the age of 30 and never married. Currentproblem and probable pathological gamblersare even more likely than lifetime problem andprobable pathological gamblers to be under theage of 30, divorced or separated and employed.Two out of every three individuals who haveever experienced gambling problems in Mis-sissippi are experiencing those difficulties now.In this section, we have examined the preva-lence of problem and probable pathologicalgambling among respondents in the Missis-sippi survey. Here, and in the first section ofthe report, our focus has been on the entiresample of 1,014 respondents. In the next sec-tion, we turn our attention to differences be-tween non-problem and problem gamblers inthe Mississippi survey. Only those respondentswho acknowledged having tried one or moretypes of gambling (N=649) are included inanalyses of the differences between non-prob-lem and problem gamblers in the following sec-tion.

pathological gambling in United States juris-dictions where both lifetime and current preva-lence data have been collected. Figure 3 showsthat the current prevalence rate of problem andprobable pathological gambling in Mississippiis equaled only by the current prevalence ratein Louisiana. As with lifetime prevalence rates,current prevalence rates tend to be higher injurisdictions where casino gambling has re-cently been introduced. In the Midwest, Iowaand Minnesota have the highest current preva-lence rates of problem and probable pathologi-cal gambling. In the South, the current preva-lence rates in Louisiana and Mississippi alsoreflect the impact of the introduction of casinogambling.


In Mississippi, 3.7% (±1.16%) of the respon-dents scored as lifetime problem gamblers andan additional 3.1% (±1.07%) of the respondentsscored as lifetime probable pathological gam-blers. In Mississippi, 2.8% (±1.01%) of the re-spondents scored as current problem gamblerswhile 2.1% (±0.88%) of the respondents scoredas current probable pathological gamblers.Both lifetime and current prevalence of prob-lem and probable pathological gambling inMississippi in 1996 are higher than in most

Figure 3: Comparing Current Prevalence Rates in the United States

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

WA TX LA MS GA MT MN ND SD IA NY

1 7

Comparing Non-Problem andProblem Gamblers in Mississippi

In considering the development of policies andprograms for problem gamblers, it is impor-tant to direct these efforts in an effective andefficient way. The most effective efforts at pre-vention, outreach and treatment are targetedat individuals who are at greatest risk of expe-riencing gambling-related difficulties. Sincethe purpose of this section is to examine indi-viduals at risk, our focus will be on differencesbetween individuals who gamble, with andwithout problems, rather than on the entiresample.

In addition to looking only at respondents whogamble, our analysis in this section is limitedto differences between non-problem gamblersand lifetime problem and probable pathologi-cal gamblers. For reasons explained in Appen-dix A, individuals who score as lifetime prob-lem and probable pathological gamblers aremost likely to reflect fully the characteristicsof individuals in the community who are cur-rently experiencing gambling difficulties. Fur-ther, since problem and probable pathologicalgamblers are statistically associated and arestrongly separated from non-problem gamblers(Abbott and Volberg 1996; Volberg and Abbott1994), these respondents are treated as a singlegroup and are referred to as problem gam-blers in this section.

❐❐❐❐❐ Demographics

Table 7 on the following page shows that, incontrast to other jurisdictions, there are fewdemographic differences between non-problemand problem gamblers in Mississippi. Whileproblem gamblers in Mississippi are signifi-cantly more likely than non-problem gamblersto be black and never married, differences inage, education and income do not attain sta-tistical significance.

While information about the demographic char-acteristics of problem gamblers is useful, it isalso important to understand differences in thegambling behavior of non-problem and prob-lem gamblers. Information about the behav-ioral correlates of problem gambling can helpagencies develop effective public education andprevention materials, effectively identify at-risk individuals and provide appropriate treat-ment services.

❐❐❐❐❐ Weekly Gambling

Behavioral correlates of problem gambling in-clude regular involvement with continuousforms of gambling (Dickerson 1993; Ladouceur,Gaboury, Dumont and Rochette 1988; Walker1992). Regular gambling is defined as weeklyor more frequent involvement with one or moretypes of gambling. Continuous forms of gam-bling are characterized by rapid cycles of playas well as the opportunity for players to im-mediately reinvest their winnings. Legal formsof continuous gambling in Mississippi includecasino table games and slot machines as wellas bingo. Illegal forms of continuous gamblinginclude wagering on sports, cards, dice andgames of skill.

Table 8 on the following page shows differ-ences in weekly involvement in different typesof wagering by non-problem and problem gam-blers. As in other jurisdictions, problem gam-blers in Mississippi are significantly morelikely than non-problem gamblers to partici-pate weekly in one or more gambling activi-ties. Four times as many problem gamblersas non-problem gamblers in Mississippi wa-ger weekly or more often. In striking contrastto other jurisdictions, where problem gamblersparticipate weekly in many more types of gam-bling than non-problem gamblers, problem

1 8

gamblers in Mississippi are significantly morelikely to participate weekly in only a few typesof gambling. These include casino gambling,card games and sports wagering. These ac-tivities are all considered continuous formsof gambling.

In light of the low levels of weekly gamblingparticipation in Mississippi, we examined dif-ferences between non-problem and problemgamblers in monthly gambling participation.As with weekly gambling, there are significantdifferences between non-problem and problemgamblers in mean monthly participation ratesonly for casino gambling, wagering on cardgames and wagering on sports. While non-problem gamblers who had been to a casino inthe past year did so on average less than once

Table 7: Demographics of Non-Problem and Problem Gamblers

Non-Problem ProblemGamblers % Gamblers %

(N=580) (N=69)

Gender Male 52.6 56.8Female 47.4 43.2

Age 18 - 20 3.9 9.021 - 29 17.8 25.730 - 54 53.8 48.755 and over 24.5 16.7

Ethnicity White 73.2 56.5 **Black 25.9 42.0Other 0.9 1.5





Table 8: Weekly Gambling ofNon-Problem and Problem Gamblers

Games Non-Problem ProblemPlayed Gamblers % Gamblers %Weekly (N=580) (N=69)

Sports 2.1 13.0 **Casino 0.7 10.1 **Card Games 0.0 8.7 **

Weekly Gambling(1+ activities) 8.6 37.7 **


1 9

a month, problem gamblers who had been to acasino in the past year did so on average oneand a half times per month. Non-problem gam-blers who wagered on card games in the pastyear did so just over once a month while prob-lem gamblers who wagered on card games inthe past year did so nearly three times a month.Non-problem gamblers who wagered on sportsin the past year did so just over once a monthwhile problem gamblers who wagered on sportsin the past year did so more than twice a month.

In considering differences between non-prob-lem and problem gamblers, it is helpful to com-pare respondents’ preferences for specific typesof gambling. While 43% of non-problem gam-blers indicate that their favorite type of gam-bling is casino gambling, 60% of problem gam-blers in Mississippi identify casino gamblingas their preferred activity. While only 3% ofnon-problem gamblers say that sports is theirfavorite type of gambling, 11% of problem gam-blers identify sports wagering as their pre-ferred activity.

There are also differences among problemgamblers in their preferences for specific typesof gambling. While 75% of female problemgamblers identify casino gambling as their fa-vorite type of gambling, only 47% of male prob-lem gamblers say that casino gambling is theirpreferred activity. While 74% of white prob-lem gamblers prefer casino gambling, only 41%of black problem gamblers identify casinos astheir favorite type of gambling. While 50% ofproblem gamblers under 30 identify casinogambling as their favorite type, 66% of prob-lem gamblers over 30 identify casinos as theirpreferred activity. These findings suggest thatthere may be two distinct groups of problemgamblers in Mississippi: older white men andwomen who prefer casino gambling andyounger black men who prefer wagering on il-legal types of gambling including sports andcard games not at casinos.

In most jurisdictions, problem and pathologi-cal gamblers are significantly more likely thanother gamblers or non-gamblers to be young,minority, blue-collar men embedded in a cul-

ture where gambling is broadly accepted, if notalways legal. Combined with the stresses thatare part of the life of young minority and blue-collar men, gambling on sports, dice or locallyavailable gaming machines as well as at casi-nos presents a challenging opportunity to getsome “action”, demonstrate control of theirlives, beat the system and gain prestige amongtheir friends (Volberg, Reitzes and Boles 1996).

The most significant change in gambling inNorth America since the 1970s has been thegrowing involvement of the middle class. Whilegambling has long been condoned among theupper classes and broadly tolerated among thelower classes, the same activities were frownedupon by the middle class (Rosecrance 1988).With little gambling experience, new middleclass gamblers have no repertoire of techniquesfor dealing with the periodic losses that are anintegral part of gambling. Until these gamblersdevelop the skills and strategies to gambleregularly without incurring disastrous losses,they are increasingly likely to experience diffi-culties (Rosecrance 1985). In Mississippi, wemay be seeing the emergence of the middleclass problem gamblers predicted byRosecrance.

❐❐❐❐❐ Expenditures

In addition to gambling regularly on continu-ous types of wagering, an important behavioralcorrelate of problem gambling is heavy gam-bling losses (Dickerson 1993). Although gam-bling losses should be considered relative toincome, comparisons of reported gambling ex-penditures provide insight into the far greaterfinancial impact of gambling on problem gam-blers and their families.

Table 9 on the following page shows differ-ences in the average monthly expenditures ongambling for non-problem and problem gam-blers in Mississippi. Only those types of gam-bling for which expenditures among problemgamblers exceeded those of non-problem gam-blers by $5 or more per month are shown.

2 0

Table 9 shows that the greatest differencesbetween non-problem and problem gamblersin Mississippi in average gambling expendi-tures are for casinos, bingo, card games andsports. The difference between average expen-ditures by non-problem and problem gamblerson casino games does not attain statistical sig-nificance because of the very large standarddeviation around the mean for both groups.2

Table 9 also shows that average total monthlyexpenditures on gambling are three and a halftimes higher for problem gamblers than fornon-problem gamblers in Mississippi.

In our discussion of gambling expenditures inthe total sample, we identified a small groupof respondents (N=81) who reported spending$100 or more on gambling in a typical month(see the discussion of Variations in Expen-ditures). This small group of respondents ac-

counted for 80% of reported monthly expendi-tures on gambling in Mississippi. In consider-ing risk factors associated with problem gam-bling, it is worth noting that 47% of the prob-lem gamblers in Mississippi (N=32) fall intothis heavy-spending group.

In addition to significant differences in gam-bling expenditures between non-problem andproblem gamblers, there are significant differ-ences among problem gamblers in terms of ex-penditures. For example, male problem gam-blers report spending an average of $200 permonth while female problem gamblers reportspending an average of $151 per month. Dif-ferences in expenditures between male andfemale problem gamblers are greatest for bingo(female problem gamblers spend nine timesmore on bingo than male problem gamblers),card games and sports.

Similarly, there are substantial differences be-tween younger and older problem gamblerswith regard to gambling expenditures. Whileproblem gamblers aged 18 to 20 report spend-ing an average of $90 per month, those aged21 to 29 report spending $311 per month andthose aged 30 to 54 report spending $152 permonth. Problem gamblers aged 55 and olderreport spending an average of $102 per month.Problem gamblers under the age of 30 reportspending the most on bingo, casino games, cardgames, games of skill and dice. Problem gam-blers aged 30 to 54 report spending the moston casino games and sports wagering whileproblem gamblers aged 55 and older reportspending the most on casino games, card gamesand the numbers.

Among white problem gamblers, the averageamount spent per month on gambling is $203compared to $136 among black problem gam-blers. Only one problem gambler is from an-other ethnic group and this individual reportsspending an average of $505 per month on gam-bling. White problem gamblers report spend-ing the most on casino games, bingo, sports andcard games. Black problem gamblers reportspending the most on card games, casinogames, sports and bingo.

2A large standard deviation around the meanis caused by very large differences in the responsesby individuals in each group. The standard devia-tion for casino spending among non-problem gam-blers is $220 while the standard deviation amongproblem gamblers is $122.

Table 9: Average Monthly Expenditures ofNon-Problem and Problem Gamblers

Type of Non-Problem ProblemGambling Gamblers $ Gamblers $

(N=580) (N=69)

Casino 31.13 61.37Bingo 1.16 28.18 **Card Games .96 22.03 **Sports 2.45 17.56 **Games of Skill 1.70 9.91 **Lottery 2.75 9.57 *Gambling Machines .15 5.62 **

Total Expenditures 49.99 179.09 **


2 1

❐❐❐❐❐ Legal versus Illegal Gambling

As we have seen, problem gamblers are sig-nificantly more likely than non-problem gam-blers to be involved in several types of gam-bling and to spend more on their gambling. Dif-ferences in the wagering of non-problem andproblem gamblers become even clearer whenwe examine involvement in legal versus illegalgambling among Mississippi respondents.Legal gambling includes lottery purchases,casino gambling, bingo, charitable gamblingand parimutuel wagering. Illegal gamblingincludes card games, dice games, sports andgames of skill as well as the numbers and ille-gal gambling machines.

Table 10 shows that while lifetime involve-ment in legal types of gambling is not signifi-cantly different for non-problem and problemgamblers, there are significant differences inlifetime involvement in illegal gambling as wellas in past year and weekly involvement in bothlegal and illegal types of gambling.

While nearly all non-problem and problemgamblers have participated at some time in alegal form of gambling, only half of the non-

problem gamblers compared to three-quartersof the problem gamblers have gambled illegallyat some time. Two out of three non-problemgamblers compared to nearly nine out of tenproblem gamblers have participated in the pastyear in legal types of gambling. In contrast,only one out of four non-problem gamblers andhalf of the problem gamblers have participatedin the past year in some type of illegal gam-bling. These differences hold for weekly gam-bling as well.

❐❐❐❐❐ Prevalence by Type of Gambling

The question most often asked about the rela-tionship between gambling and problem gam-bling is: What type of gambling is most likelyto add to the number of problem and patho-logical gamblers in the general population?

We have examined the relationship betweenweekly involvement, gambling expendituresand problem gambling among respondents inthis survey to help answer this question forMississippi. Our analysis shows that for life-time problem and pathological gamblers, con-tinuous forms of gambling including casinogambling, sports betting and wagering on cardgames not at a casino present the greatest risk.

Another approach is to examine the prevalenceof gambling problems among individuals whohave participated in specific types of gambling.Figure 4 on the following page illustrates theprevalence of lifetime problem and pathologi-cal gambling for the total sample from Missis-sippi, for respondents who have ever gambledand for respondents who have ever participatedin different types of gambling.

Figure 4 shows that lifetime prevalence ratesare substantially higher among individualswho have participated in specific types of wa-gering than among the sample as a whole oramong gamblers in general. In Mississippi,prevalence rates are highest among individu-als who have ever wagered on illegal types ofgambling such as dice games, illegal gamblingmachines and on card games not at casinos.

Table 10: Legal and Illegal Gambling byNon-Problem and Problem Gamblers

Type of Non-Problem ProblemGambling Gamblers % Gamblers %

(N=580) (N=69)

Lifetime Legal 94.2 97.7 Illegal 54.3 73.5 **

Past Year Legal 67.3 85.9 ** Illegal 25.9 56.9 **

Weekly Legal 4.1 16.7 ** Illegal 3.4 23.5 **


2 2

These types of gambling are most attractive toyoung men in the general population. Preva-lence rates are also high among individualswho have ever wagered on parimutuel eventsand bingo which are legal forms of gamblingattractive to older gamblers in the generalpopulation.

Prevalence rates are lower for most of the le-gal types of gambling available to Mississippiresidents. Mississippi does not have a state lot-tery so problem gamblers must travel out-of-state to purchase lottery tickets. Dockside ca-sinos have only been operational in Mississippisince 1992. Since gambling problems can takeseveral years to develop fully (Volberg 1988),it is likely that the prevalence of problem andpathological gambling among casino gamblersin Mississippi will increase to a level closer tothe prevalence rates for sports, bingo andparimutuel events in the future.

❐❐❐❐❐ Other Significant Differences

In addition to their demographics and gam-bling involvement, there are other significantdifferences between non-problem and problem

gamblers in Mississippi. These include differ-ences in respondents’ perceptions of their gam-bling involvement, the amount of time theyusually gamble and the largest amount theyreport losing in a single day.

— Starting to Gamble —

One important difference between non-prob-lem and problem gamblers is the age at whichthey start gambling. While the mean age atwhich non-problem gamblers in Mississippistarted gambling is 34 years old, the mean ageat which problem and pathological gamblersin Mississippi started gambling is significantlyyounger at 27 years old. Table 11 on the fol-lowing page shows that there are significantdifferences in the age at which respondentsstarted gambling by gender, age and maritalstatus for both non-problem and problem gam-blers. In contrast to other jurisdictions, differ-ences in mean starting age among ethnicgroups in Mississippi are not significantly dif-ferent.

Table 11 shows that men, whether non-prob-lem or problem gamblers, start gambling at asignificantly earlier age than women in Mis-

Figure 4: Prevalence by Types of Gambling

0%

5%

10%

15%

20%

25%

Tot

alS

ampl

e

Gam

bler

s

Cha

ritab

le

Cas

inos

Lotte

ry

Bin

go

Par

imut

uel

Spo

rts

Ski

ll

Car

dG

ames

Mac

hine

s

Dic

e

Lifetime Problem Lifetime Pathological

2 3

sissippi. Younger respondents, whether non-problem or problem gamblers, also recall start-ing to gamble at significantly earlier ages thanolder respondents. Finally, problem gamblerswho have never married recall starting togamble at a significantly earlier age than non-problem or problem gamblers who are married,divorced or separated, as well as non-problemgamblers who have never married. Problemgamblers who are divorced or separated arethe only group that recall starting to gamblelater in life than their non-problem counter-parts. This suggests that separation and di-vorce may be a risk factor in the developmentof gambling problems in older adults.

— Gambling Experiences and Resources —

Table 12 on the following page shows thatproblem gamblers are significantly more likelythan non-problem gamblers in Mississippi tohave felt nervous about their gambling and tohave felt that one or both parents had a gam-bling problem. Table 12 also shows that thereare significant differences between non-prob-

lem and problem gamblers in Mississippi interms of the time and resources that they de-vote to gambling. Problem gamblers are sig-nificantly more likely than non-problem gam-blers to spend six or more hours gambling persession, to have lost $1,000 or more in a singleday, and to travel 60 or more miles in order togamble.

— Help-Seeking —

As in other jurisdictions, very few problemgamblers in Mississippi acknowledge desiringor seeking help for a gambling problem. Only6% (N=4) of problem gamblers in Mississippihave desired help for a gambling problem andonly 4% (N=3) have sought such help. One in-dividual sought help from a problem gamblingtreatment program in Mississippi and anothersought help from a problem gambling treat-ment program outside of Mississippi. The othertwo respondents declined to indicate wherethey had sought help for their gambling prob-lems.

Table 11: Mean Starting Age of Non-Problem and Problem Gamblers


(N=580) (N=69)

Gender Male 29 23 **Female 38 32

Age 18 - 20 16 16 **21 - 29 23 1930 - 54 32 3055 or over 45 36

Ethnicity White 34 27Black 32 27Other 39 39

Marital Status Married 35 27 **Widowed 51 —Divorced/Separated 30 36Never Married 26 19


2 4

— Location —

In planning the development of services forproblem gamblers in Mississippi, it is helpfulto know where these individuals reside andwhere they prefer to gamble. In Mississippi,30% of the problem gamblers reside in coun-ties on the western side of the state, 16% ofthe problem gamblers reside in the Gulf Coastcounties and 54% of the problem gamblers iden-tified in the survey reside in counties wheredockside casino gambling is not readily avail-able.

In light of their residential distribution, it isinteresting to note that problem gamblers liv-ing outside of counties in Mississippi with

dockside casinos are significantly more likelythan problem gamblers living within suchcounties to indicate that they usually travel60 miles or more to gamble. Over half of prob-lem gamblers living in non-dockside casinocounties (52%) indicate that they usually travel60 miles or more to gamble. In contrast, 35%of the problem gamblers living in western Mis-sissippi and 68% of the problem gamblers liv-ing in the Gulf counties indicate that they usu-ally travel less than 15 miles to gamble.


Our focus in this section has been on the riskfactors associated with problem gambling in

Table 12: Other Significant Differences Between Non-Problem and Problem Gamblers


(N=580) (N=69)

Ever Felt Nervous About Your Gambling 10.8 57.6 **Parent Ever Have Gambling Problem 2.4 16.2 **

Usually Gamble With Alone 17.0 17.0 Spouse/Partner 28.5 25.9 Other Family 14.5 12.2 Friends 31.0 38.0 Other 5.3 6.9

Usual Time Spent Gambling ** < 1 to 2 hours 75.8 36.9 3 to 5 hours 21.2 52.3 6 or more hours 3.0 10.8

Largest Amount Lost in One Day ** < $1 to $9 25.5 3.9 $10 to $99 55.5 28.8 $100 to $999 16.7 50.5 $1,000 or more 2.3 16.8

Usual Distance to Gamble ** 0 to 15 miles 53.8 27.0 15 to 60 miles 29.0 38.3 60 or more miles 17.3 34.7


2 5

the general population. To identify these riskfactors, we compared problem and non-prob-lem gamblers in Mississippi as well as in otherjurisdictions where similar surveys have beencompleted. As predicted by the research lit-erature, regular gambling involvement, in par-ticular with continuous forms of gambling andheavy gambling losses, are the factors associ-ated with gambling-related difficulties in Mis-sissippi.

Although there are few demographic differ-ences between non-problem and problem gam-blers in Mississippi, problem gamblers are sig-nificantly more likely than non-problem gam-blers to be black and never married. Informa-tion on the age at which non-problem and prob-lem gamblers started gambling suggests thatseparation or divorce may be a risk factor inthe development of gambling problems in olderadults.

Problem gamblers are significantly more likelythan non-problem gamblers to wager regularlyalthough they do so on only a few activities,including casino gambling, card games andsports wagering. While the majority of prob-lem gamblers in Mississippi identify casinogambling as their preferred activity, this pref-erence is most pronounced among older whitefemale problem gamblers. Despite this ex-pressed preference, problem gambling preva-lence rates are highest among individuals who

have wagered on illegal types of gambling suchas dice games, non-casino gambling machinesand card games and among individuals whohave ever wagered on parimutuel events andbingo.

These findings suggest that there may be twodistinct groups of problem gamblers in Missis-sippi: older white men and women who oncepreferred parimutuels and bingo but may nowprefer casino gambling and younger black menwho prefer sports wagering as well as gamblingon card games, dice games and games of skill.Given these findings, prevention, outreach andtreatment services should probably be aimedprimarily at young black men engaged in ille-gal types of gambling and at middle-class gam-blers who gamble regularly at dockside casi-nos. It will be especially important to directsome of these efforts at women problem gam-blers who comprise nearly half of the lifetimeand current problem gamblers.

In this section, we have identified several ma-jor risk factors associated with gambling-re-lated difficulties among respondents in Mis-sissippi. Our focus has been on respondentswho acknowledge gambling, whether or notthey experience difficulties related to this in-volvement. In the next section, we will focuson comparisons between the survey in Missis-sippi and two similar surveys completed re-cently in Georgia and Louisiana.

Comparing Mississippi,Louisiana and Georgia

This section focuses on comparisons betweenthe survey in Mississippi and two recent sur-veys completed in Georgia and Louisiana.These two jurisdictions were chosen for com-parison with Mississippi because (1) these arethe other two Southern states where surveysof gambling and problem gambling have been

completed, and (2) Louisiana is one of the fewstates besides Mississippi where riverboat ca-sinos have been legalized while Georgia hadno legal gambling except charitable bingo anda one-year old state lottery at the time the sur-vey was completed.

2 6

This section examines similarities and differ-ences in the questionnaires used in the threesurveys. Differences and similarities betweenthe samples from the three states are also ex-amined. We then look at differences and simi-larities in gambling involvement and expendi-tures, in the prevalence of problem gamblingand in the characteristics of problem gamblersin these three jurisdictions.

❐❐❐❐❐ Comparing the Questionnaires

In the Methods section, we noted that the ques-tionnaire for the 1996 Mississippi survey in-cluded three major sections: gambling partici-pation, the lifetime and current South OaksGambling Screen and questions about demo-

graphic characteristics. Care was taken indesigning the questionnaire for the Mississippisurvey to ensure that respondents’ involvementin different types of gambling could be com-pared with similar surveys in nearby jurisdic-tions. Table 13 shows how the different typesof gambling included in the Mississippi sur-vey were matched to the types of gambling in-cluded in surveys in Louisiana and Georgia:

Table 13 shows that the surveys in Georgiaand Louisiana obtained greater detail aboutsome types of gambling than the survey in Mis-sissippi. For example, respondents in Louisi-ana were asked about their casino play in Loui-siana on a riverboat, in Louisiana at an Indiancasino and outside of Louisiana. Respondentsin Georgia were asked about their lottery gam-

Table 13: Types of Gambling Included in Prevalence Surveys

Mississippi Louisiana Georgia

• Lottery • Louisiana lottery game • Instant or scratch-off lottery games• Daily lottery games• Lotto-type lottery games

• Been to a casino • Louisiana riverboat casino game • Gaming devices or slot machines at• Louisiana Indian Reservation an out-of-state casino casino game • Card or dice games at an out-of-state• Out-of-state casino game casino

• Bingo • — • Bingo• Card games • Private card games • Card games for money• Horses, dogs or other animals • Louisiana horse race at the track • Horse or dog races (includes

• Louisiana horse race at OTB parlor on-track, off-track or with a bookie)• Cockfights or dogfights

• Stocks or commodities • Speculative stock or commodity • Speculative investments or stock investment market

• Gambling machines • Louisiana video poker game not at • — Indian casino or riverboat casino

• Games of skill • Private game of skill such as • Games of skill billiards, bowling or golf

• Dice games • Private game of chance such as dice • Craps or other dice games• Sports • Outcome of a public sporting event • Bet on sports with friends,

acquaintances, co-workers• Sports pools• Sports with a bookie

• Charitable • Louisiana charitable game • Raffles, casino nights or other small (raffle, bingo, keno) stakes charitable gaming

• Numbers • — • Numbers games (not the daily lottery game)

• Other • Other • Other

2 7

bling in greater detail than respondents in Mis-sissippi or Louisiana since the state lottery inGeorgia is one of the few legal types of gam-bling in that state. Respondents in Georgiawere not asked about wagering on gamblingmachines not at casinos while respondents inLouisiana were not asked about wagering onthe illegal numbers game.

❐❐❐❐❐ Comparing the Samples

In contrast to the sample from Mississippi, thesamples from Louisiana and Georgia were notweighted. Consequently, the samples fromthese states are not fully representative of thepopulation in the states. In Louisiana, blacksare under-represented while in Georgia, malesand blacks are under-represented. Table 14

shows that the three samples are significantlydifferent in terms of gender and ethnicity.Respondents from Georgia are the most likelyto be female while those from Louisiana arethe most likely to be male. Respondents fromMississippi are the most likely to be black al-though this may be the result of weights ap-plied to the Mississippi data and not to the datafrom Georgia or Louisiana.

Table 14 also shows that the three samplesare significantly different in terms of genderand ethnicity. Respondents from Mississippiare the most likely to be widowed while respon-dents from Georgia are the least likely to haveever married. Respondents from Mississippiare the least likely to have graduated from highschool while respondents from Louisiana are

Table 14: Demographic Characteristics of Respondents

Mississippi Louisiana Georgia(N=1,014) (N=1,818) (N=1,550)

Gender Male 46.1 48.1 41.6 **Female 53.9 51.9 58.4

Age Under 30 20.8 22.3 21.3Over 30 79.2 79.7 78.7

Ethnicity White 67.4 75.3 71.6 **Black 31.7 20.9 23.6Other 1.0 3.8 4.7

Marital Status Married 56.4 57.7 56.1 *Widowed 10.3 8.9 9.2Divorced/Separated 13.5 12.7 16.5Never Married 19.9 20.6 18.3

Education Less than HS 16.2 14.0 12.3 *HS and Over 83.8 86.0 87.7

Income Annual Income <$25,000 42.7 43.5 35.1 **Annual Income > $25,000 57.3 56.5 64.9

Employment Working 64.2 61.9 66.7 *Other 34.0 35.1 31.3Unemployed 1.8 3.1 2.0


2 8

the least likely to have annual household in-comes over $25,000. Respondents from Loui-siana are the least likely to be working andthe most likely to be unemployed.

❐❐❐❐❐ Comparing Gambling Involvement

Lifetime gambling participation is significantlyhigher in Louisiana and Georgia than in Mis-sissippi. While only 64% of the Mississippi re-spondents have ever tried one or more types ofgambling, 81% of the Louisiana respondentsand 74% of the Georgia respondents have triedone or more types of gambling. Differences inlifetime gambling participation hold true forpast-year and weekly participation. While 75%of Louisiana respondents and 65% of Georgia

respondents have wagered on one or more typesof gambling in the past year, only 49% of Mis-sissippi respondents have done so. While 38%of Louisiana respondents and 28% of Georgiarespondents wager weekly on one or more typesof gambling, only 7% of Mississippi respon-dents wager weekly.

In general, differences between the samplesfrom the three states are reflected in the de-mographics of those respondents who have evergambled. Table 15 shows differences in thedemographic characteristics of respondentswho have ever gambled in Mississippi, Louisi-ana and Georgia. Gamblers in Mississippi andLouisiana are more likely to be male than gam-blers in Georgia. Black respondents in all threestates are slightly less likely to have ever

Table 15: Demographics of Gamblers

Mississippi Louisiana Georgia(N=641) (N=1,469) (N=1,151)

Gender Male 53.1 50.4 45.4 **Female 46.9 49.6 54.6

Age Under 30 23.1 23.4 23.1Over 30 76.9 76.6 76.9

Ethnicity White 71.5 78.6 73.0 **Black 27.6 17.9 22.4Other 0.9 3.5 4.6

Marital Status Married 58.0 59.5 55.8 *Widowed 6.4 6.1 6.3Divorced/Separated 14.3 13.3 18.5Never Married 21.3 21.0 19.4

Education Less than HS 11.5 11.0 9.0HS and Over 88.5 89.0 91.0

Income Annual Income <$25,000 34.9 39.6 31.4 **Annual Income > $25,000 65.1 60.4 68.6

Employment Working 70.6 65.3 72.2 **Other 28.3 31.8 25.9Unemployed 1.1 3.0 2.0


2 9

gambled than white respondents. Widowed re-spondents in all three states are less likely tohave ever gambled than other respondents. Inall three states, respondents with higher lev-els of education and income are more likely tohave ever gambled.

Table 16 shows differences in participation inspecific types of gambling among respondentsin the three states. While the proportion ofrespondents from Louisiana and Georgia whohave ever gambled is significantly higher thanamong respondents from Mississippi, many ofthe differences in participation in specific typesof gambling are due to differences in availabil-ity in the three states.

With no state lottery in place, respondents inMississippi are the least likely to have ever pur-chased lottery products. Since riverboat anddockside casinos are legal in Mississippi andLouisiana but not in Georgia, respondents fromthese two states are significantly more likelyto have gambled at a casino. With no race-tracks in Georgia or Mississippi, respondentsfrom these states are significantly less likelyto have wagered on parimutuel events thanrespondents from Louisiana. With legal video

poker available throughout the state, Louisi-ana respondents are significantly more likelyto acknowledge wagering on gambling ma-chines than respondents in Mississippi.

Differences across the states in illegal gamblingare somewhat more difficult to explain. Re-spondents from Louisiana are significantlymore likely than those from Georgia and Mis-sissippi to have played dice and card gamesfor money. Respondents from Georgia are sig-nificantly less likely than respondents fromMississippi and Louisiana to have wagered ongames of skill while respondents from Missis-sippi are the least likely to have wagered onsports. Reasons for these differences may becultural as well as demographic, such as theinfluence of the Cajun culture on attitudes to-ward gambling in Louisiana and the impor-tance of the Baptist Church in Mississippi.However, this remains a hypothesis to be testedin future research.

Reported expenditures on gambling also differsignificantly across the three states. While theaverage monthly expenditure on gambling is$82 in Georgia and $59 in Louisiana, the aver-age monthly expenditure in Mississippi is $41.

Table 16: Lifetime Gambling Participation Rates

Type of Gambling Mississippi Louisiana Georgia(N=1,014) (N=1,818) (N=1,550)

Lottery 32.2 68.3 61.2 **Casino 45.0 45.7 26.3 **Bingo 17.0 — 17.6Charitable 36.2 43.0 37.6 **Card Games 19.9 25.1 20.1 **Parimutuels 14.8 26.2 17.5 **Gambling Machines 8.8 28.9 — **Games of Skill 13.2 13.9 9.4 **Dice Games 6.0 10.3 5.0 **Stocks or Commodities 7.2 15.0 15.2 **Sports 19.7 26.8 24.9 **Numbers 1.7 — 3.7 **Other 1.4 4.8 1.5 **


3 0

Among Georgia respondents, average expen-ditures are highest for instant lottery games,Lotto and slot machines at casinos (out-of-state). Among Louisiana respondents, aver-age expenditures are highest for Louisianariverboat casinos, sports and out-of-state casi-nos. Among Mississippi respondents, as wehave seen, average expenditures are highestfor dockside casinos.

❐❐❐❐❐ Comparing Problem GamblingPrevalence

Table 17 shows the point estimates and stan-dard deviations (rounded to one decimal point)for lifetime and current problem and probablepathological gambling for Mississippi, Louisi-ana and Georgia as well as the combined life-time and current prevalence rates.

While there are small overlaps in the standarddeviations for both lifetime and current prob-lem and probable pathological gambling, thedifferences in prevalence rates between Mis-sissippi and Louisiana, on the one hand, andGeorgia, on the other, are significant. To-gether, these figures show that there is a sub-stantial and significant difference in the life-time and current prevalence rates of problemand pathological gambling in these threeSouthern states.

In contrast to Georgia, casino gambling is avail-able in Louisiana and Mississippi. However,gambling involvement and expenditures are farhigher in Louisiana than in Mississippi. Howcan we account for the equally high prevalencerates of problem gambling in Mississippi andLouisiana?

One possible explanation of the much higherlevel of gambling participation in Louisiana isthat gambling of all kinds is more broadly ac-cepted in this state, where Creole and Cajuncultural influences are strong, while Missis-sippi remains a Baptist stronghold. The lowrate of gambling involvement and the high rateof gambling problems in Mississippi may bedue to the lower per capita income in Missis-sippi compared with either Louisiana or Geor-gia. Since the prevalence of gambling prob-lems is much higher among lower socio-eco-nomic groups, the low socio-economic status ofa large proportion of the population in Missis-sippi may contribute disproportionately to ahigher rate of problem and pathological gam-bling in the state. A third explanation is thatthe availability of casino gambling has contrib-uted rapidly to an equally high prevalence ratein the two states regardless of levels of involve-ment and expenditure.

Table 17: Lifetime and Current Prevalence Rates

Mississippi Louisiana Georgia(N=1,014) (N=1,818) (N=1,550)

Lifetime Problem 3.7 (±1.2) 4.5 (±.0.9) 2.8 (±0.8) *Probable Pathological 3.1 (±1.1) 2.5 (±0.7) 1.6 (±0.6) *Total 6.8 (±1.5) 7.0 (±1.2) 4.4 (±1.0) **

Current Problem 2.8 (±1.0) 3.4 (±0.8) 1.5 (±0.6) **Probable Pathological 2.1 (±0.9) 1.4 (±0.5) 0.8 (±0.4) **Total 4.9 (±1.3) 4.8 (±1.0) 2.3 (±0.7) **


3 1

— Demographics —

In contrast to differences in the demographiccharacteristics among the overall samples andamong gamblers, Table 18 shows that thereare few differences among problem gamblersfrom Mississippi, Louisiana and Georgia.While the small size of these groups makes itdifficult to establish statistical significance,these results correspond with findings fromnumerous other studies in the United Statesand Canada (Ladouceur 1996; Volberg 1996b).

— Gambling Involvement —

In contrast to substantial differences in gam-bling involvement among respondents from

Mississippi, Louisiana and Georgia, Table 19on the following page shows that there are fewsignificant differences in gambling participa-tion by problem gamblers in these three states.With no lottery in Mississippi, problem gam-blers from this state are significantly less likelyto have purchased lottery products than prob-lem gamblers from Louisiana and Georgia.Given the availability of casinos in Mississippiand Louisiana, problem gamblers in these twostates are significantly more likely to havegambled at a casino than problem gamblersfrom Georgia. With the availability of videopoker in Louisiana, problem gamblers in thatstate are significantly more likely than prob-lem gamblers in Mississippi to have wageredon gambling machines not at a casino.

Table 18: Comparing Lifetime Problem Gamblers

Mississippi Louisiana Georgia(N=69) (N=127) (N=68)

Gender Male 56.8 62.2 63.2Female 43.2 37.8 36.8

Age Under 30 34.7 39.7 41.2Over 30 65.3 60.3 58.8

Ethnicity White 56.5 59.1 52.2Black 42.0 33.9 38.8Other 1.5 7.1 9.0

Marital Status Married 45.0 37.8 36.8Widowed 1.2 3.1 8.8Divorced/Separated 18.9 19.7 20.6Never Married 34.9 39.4 33.8

Education Less than HS 15.9 22.8 19.1HS and Over 84.1 77.2 80.9

Income Annual Income <$25,000 33.9 47.5 32.8Annual Income > $25,000 66.1 52.5 67.2

Employment Working 76.5 66.9 74.2Other 22.3 29.0 21.2Unemployed 1.2 4.0 4.5


3 2

Although there are few significant differencesin lifetime gambling involvement among prob-lem gamblers in Mississippi, Louisiana andGeorgia, problem gamblers in Mississippi aresignificantly less likely to gamble on a weeklybasis than problem gamblers in Louisiana andGeorgia. Over three quarters (77%) of prob-lem gamblers in Louisiana and 63% of prob-lem gamblers in Georgia gamble weekly com-pared to only 38% of problem gamblers in Mis-sissippi. While average monthly expendituresare $660 among problem gamblers in Louisi-ana and $299 among problem gamblers inGeorgia, average monthly expenditures areonly $179 among problem gamblers in Missis-sippi.


Our focus in this section has been on compari-sons between the survey in Mississippi and twosimilar surveys completed recently in Georgiaand Louisiana. In comparing these surveys,we looked at the questionnaires used and atthe characteristics of the samples from eachstate. In considering similarities and differ-

ences between the three studies, we looked atgambling involvement, problem gamblingprevalence rates and at the characteristics ofproblem gamblers in the three states.

Review of the questionnaires from the threesurveys found that while the surveys in Geor-gia and Louisiana obtained greater detail aboutsome types of gambling than the survey inMississippi, all three surveys obtained identi-cal information with which to measure theprevalence of problem and pathological gam-bling as well as identical information about thedemographic characteristics of the respon-dents.

The three samples are significantly differentin terms of gender and ethnicity. Unlike thesample from Mississippi, the samples fromLouisiana and Georgia were not weighted.Consequently, the samples from these statesare not fully representative of the populationin the states. The extent to which significantdifferences between the samples are due toactual differences in the population of thesethree states is difficult to determine althoughthe differences between weighted andunweighted samples in Louisiana and Geor-

Table 19: Comparing Gambling by Problem Gamblers

Type of Gambling Mississippi Louisiana Georgia(N=69) (N=127) (N=68)

Lottery 63.6 89.0 89.7 **Casino 83.8 65.4 54.4 **Bingo 40.2 — 27.9Charitable 56.7 63.0 50.0Card Games 56.9 54.3 52.9Parimutuels 40.0 48.0 32.4Gambling Machines 27.0 58.3 — **Games of Skill 33.0 29.9 20.6Dice Games 20.0 33.1 27.9Stocks or Commodities 11.6 19.7 23.5Sports 44.9 59.8 51.5Numbers 9.2 — 5.9Other 9.0 10.2 7.4


3 3

gia all contribute to conservative results in theprevalence of problem gambling.

Lifetime, past-year and weekly gambling par-ticipation are all significantly higher in Loui-siana and Georgia than in Mississippi. Withregard to specific types of gambling, many ofthe differences in participation are due to dif-ferences in availability in the three states.There is no state lottery in Mississippi and noriverboat or dockside casinos in Georgia. Nei-ther Georgia nor Mississippi have legal horseracing events or legal gambling machines. Par-ticipation in these types of gambling in thethree states differs significantly depending onavailability. Reported expenditures on gam-bling are significantly higher in Georgia andLouisiana than in Mississippi. In Georgia, ex-penditures are highest for lottery games andfor slot machines at casinos out-of-state. InLouisiana, expenditures are highest forriverboat casinos, sports and out-of-state casi-

nos. In Mississippi, average expenditures arehighest for dockside casinos.

There is a substantial and significant differ-ence in the lifetime and current prevalencerates of problem and pathological gambling inthe three states. While the availability of ca-sino gambling has contributed to the highprevalence rates of problem gambling in Loui-siana and Mississippi, there are other factorsthat may contribute to differences in preva-lence rates across states in the same region,including cultural influences and socio-eco-nomic status. Despite significant differencesin the prevalence of problem and pathologicalgambling, there are few differences in the char-acteristics of problem gamblers from the threestates. As in the larger population, problemgamblers in Louisiana and Georgia gamblemore often and spend more resources on gam-bling than problem gamblers in Mississippi.

Summary and Conclusion

An additional purpose of this study was to es-tablish a baseline measure of the prevalenceof gambling-related problems among adults inMississippi. The other main purpose of thestudy was to identify the types of gamblingcausing the greatest difficulties for the citizensof Mississippi. The results of this study showthat substantial numbers of Mississippi resi-dents participate in legal gambling and thatmost residents spend small to moderateamounts on gambling. However, the study alsoshows that there is a significant proportion ofthe adult population of Mississippi experienc-ing difficulties related to their gambling. Weestimate that, at a minimum, there are pres-ently 22,300 adult Mississippi residents expe-riencing severe difficulties related to their in-volvement in gambling.


In this report, we have examined patterns ofgambling participation for the general popula-tion in Mississippi, the prevalence of lifetimeand current problem and pathological gamblingin the state, the risk factors associated withproblem and pathological gambling, and dif-ferences and similarities between Mississippi,Louisiana and Georgia in gambling involve-ment and problem gambling prevalence.

Mississippi has the lowest rates of lifetime,past-year and weekly gambling identified insimilar surveys in the United States. Lifetimeparticipation in Mississippi is highest for ca-sino gambling, charitable wagering and lotterygames. As in other jurisdictions, young un-married White men with relatively high levels

3 4

of education and income are most likely to haveever gambled in Mississippi. By a large mar-gin, Mississippi respondents who do gambleprefer to do so on slot machines at casinos neartheir place of residence. Casino gambling ac-counts for over half of the reported expendi-tures on gambling reported in Mississippi.Apart from much rates of participation, thegambling patterns identified in Mississippi aresimilar to patterns identified in other jurisdic-tions.

Both lifetime and current prevalence of prob-lem and probable pathological gambling in Mis-sissippi in 1996 are higher than in most otherstates where similar surveys have been com-pleted. Lifetime and current problem and prob-able pathological gamblers in Mississippi aresignificantly more likely than other respon-dents to be male, under the age of 30, nevermarried and employed. Two out of every threeindividuals who have ever experienced gam-bling problems in Mississippi are experienc-ing those difficulties now.

As predicted by the research literature, regu-lar gambling involvement, in particular withcontinuous forms of gambling, and heavygambling losses are the factors associated withgambling-related difficulties in Mississippi.Problem gamblers in Mississippi are signifi-cantly more likely than non-problem gamblersto be black and never married. Informationon the age at which non-problem and problemgamblers started gambling suggests that sepa-ration or divorce may be a risk factor in thedevelopment of gambling problems in olderadults in Mississippi. Problem gamblers inMississippi are significantly more likely thannon-problem gamblers to spend six or morehours gambling per session and to have lost$1,000 or more in a single day. While half ofthe problem gamblers in Mississippi live out-side of counties where dockside casino gam-bling is available, a substantial proportion ofthese individuals usually travel 60 miles ormore to gamble. In contrast, problem gamblersliving in counties where dockside casino gam-bling are more likely to travel less than 15miles to gamble.

Comparison of a variety of risk factors for prob-lem gambling among Mississippi respondentsshows that there may be two distinct groupsof problem gamblers in Mississippi: older whitemen and women who once preferredparimutuels and bingo but may now prefercasino gambling and younger black men whoprefer sports wagering as well as gambling oncard games, dice games and games of skill.

In comparing the survey in Mississippi withthose in Louisiana and Georgia, we looked atthe questionnaires used and at the character-istics of the samples from each state as well asat gambling involvement, problem gamblingprevalence and the characteristics of problemgamblers in the three states. The threesamples are significantly different in terms ofgender and ethnicity although the extent towhich these differences are due to actual dif-ferences in the population is difficult to deter-mine.

While lifetime, past-year and weekly gamblingparticipation are all significantly higher inLouisiana and Georgia than in Mississippi, dif-ferences in participation in specific types ofgambling appear to be due to differences inavailability. Reported expenditures on gam-bling are significantly higher in Georgia andLouisiana than in Mississippi. In Georgia,expenditures are highest for lottery games andfor slot machines at casinos out-of-state. InLouisiana, expenditures are highest forriverboat casinos, sports and out-of-state casi-nos. In Mississippi, average expenditures arehighest for dockside casinos.

There is a substantial and significant differ-ence in the lifetime and current prevalencerates of problem and pathological gambling inthe three states. While the availability of ca-sino gambling is one factor contributing to thehigh prevalence rates of problem gambling inLouisiana and Mississippi, there are other fac-tors that may also contribute to differences inprevalence rates across these states, includ-ing cultural influences and socio-economic sta-tus. Despite significant differences in theprevalence of problem and pathological gam-bling, there are few differences in the charac-

3 5

counselor certification and the funding of thisprevalence study. Additional directions for thefuture include:

• development of public education and preven-tion services targeted toward at-risk and under-served groups in the population, includingyoung men as well as women problem gam-blers;

• establishment of treatment services and the de-velopment of innovative treatment alternativesto provide a variety of options for individualsseeking help for gambling problems;

• evaluation of the effectiveness of the servicesthat are established, based on uniform data col-lected from existing providers and the helpline;

• continued monitoring of gambling participa-tion and problem gambling prevalence in thestate to assess the impacts of the introductionof new dockside casinos as well as the possibleintroduction of other types of legal gamblingon the residents of Mississippi and to refineexisting efforts to minimize the negative im-pacts of gambling; and

• research activities including a thorough exami-nation of the prevalence and characteristics ofproblem gamblers among under-served and/orminority groups as well as among adolescentsin Mississippi.

This report represents the first opportunity toassess the prevalence of problem and patho-logical gambling in Mississippi. These dataprovide insights that will be valuable in on-going policy and planning efforts in the state.In the future, it will be important for everyoneinvolved with legal gambling in Mississippi tocontinue to work together to develop ways tohelp the increasing number of individuals inthe state who experience difficulties related totheir gambling and to prevent any further in-creases in the prevalence of problem gamblingin Mississippi.

teristics of problem gamblers from the threestates. As in the larger population, problemgamblers in Louisiana and Georgia gamblemore often and spend more on gambling thanproblem gamblers in Mississippi.

❐❐❐❐❐ Directions for the Future

The costs of problem and pathological gamblingcan be high, not only for individuals but forfamilies and communities. Pathological gam-blers experience physical and psychologicalstress and exhibit substantial rates of depres-sion, alcohol and substance abuse and suicidalideation. The families of pathological gamblersexperience physical and psychological abuse aswell as harassment and threats from bill col-lectors and creditors. Other significant impactsinclude costs to employers, creditors, insurancecompanies, social service agencies and the civiland criminal justice systems.

The State of Mississippi is in the process ofdeveloping services for problem and pathologi-cal gamblers and their families. Interestedstakeholders in the state have come togetherto achieve consensus about how such servicesshould be developed. This survey of gamblingand problem gambling among adults in Mis-sissippi is one of the first steps in the process.Information about the characteristics of indi-viduals with gambling problems in Mississippisuggests that, as expected, the prevalence ofproblem and pathological gambling amongyoung black man in Mississippi is high. How-ever, there also seems to be an emerging groupof problem gamblers who are middle-class andmiddle-aged whites with a strong preferencefor casino gambling. The services developedfor problem gamblers in Mississippi will haveto reflect the diversity of problem gamblers inthe general population.

The Mississippi Council has already under-taken several important activities in the de-velopment of services for problem gamblers andtheir families in Mississippi, including the es-tablishment of a 24-hour helpline, training fortreatment professionals to achieve gambling

3 6

References

Abbott, M. W. and R. A. Volberg. 1991. Gamblingand Problem Gambling in New Zealand:Report on Phase One of the National Sur-vey. Research Series No. 12. Wellington:New Zealand Department of Internal Af-fairs.

Abbott, M. W. and R. A. Volberg. 1992. FrequentGamblers and Problem Gamblers in NewZealand: Report on Phase Two of the Na-tional Survey. Research Series No. 14.Wellington: New Zealand Department ofInternal Affairs.

Abbott, M. W. and R. A. Volberg. 1996. “The NewZealand National Survey of Problem andPathological Gambling.” Journal of Gam-bling Studies, 12 (2): 143-160.

American Psychiatric Association. 1980. Diagnos-tic and Statistical Manual of Mental Disor-ders, Third Edition. Washington, DC:American Psychiatric Association.

American Psychiatric Association. 1994. Diagnos-tic and Statistical Manual of Mental Disor-ders, Fourth Edition. Washington, DC:American Psychiatric Association.

Angus Reid Group and Gemini Research. 1994. So-cial Gaming and Problem Gambling in Brit-ish Columbia. Report to the British Colum-bia Lottery Corporation.

Baseline Market Research. 1992. Prevalence Sur-vey of Problem Gambling in NewBrunswick. Report to the New BrunswickDepartment of Finance.

Christiansen/Cummings Associates. 1992. LegalGambling in Connecticut: Assessment ofCurrent Status and Options for the Future.Report to the Connecticut Division of Spe-cial Revenues.

Criterion Research Corporation. 1993. ProblemGambling Study. Report to the ManitobaLotteries Foundation.

Dickerson, M. G. 1993. “A Preliminary Explora-tion of a Two-Stage Methodology in the As-sessment of the Extent and Degree of Gam-bling-Related Problems in the AustralianPopulation.” In Gambling Behavior andProblem Gambling. W. R. Eadington andJ. A. Cornelius (eds). Reno: University ofNevada Press.

Emerson, M. O. and J. C. Laundergan. 1996. “Gam-bling and Problem Gambling Among AdultMinnesotans: Changes 1990 to 1994,” Jour-nal of Gambling Studies 12 (3): 291-304.

International Gaming and Wagering Business. 1996.North American Gaming Report. Supple-ment to International Gaming and Wager-ing Business.

Kallick-Kaufmann, M. 1979. “The Micro and MacroDimensions of Gambling in the UnitedStates,” Journal of Social Issues 35: 7-26.

Kallick, M., Suits, D. Dielman, T. and J. Hybels. 1979.A Survey of American Gambling Attitudesand Behavior. Research Report Series, Sur-vey Research Center, Institute for SocialResearch. Ann Arbor: University of Michi-gan Press.

Ladouceur, R. 1993. “Prevalence Estimates of Patho-logical Gamblers in Quebec, Canada.” InGambling Behavior and Problem Gambling,W. R. Eadington and J. A. Cornelius (eds).(Pp. 379-384). Reno: University of Nevada,Reno.

Ladouceur, R. 1996. “Prevalence of PathologicalGamblers in Canada and Related Issues,”Journal of Gambling Studies 12 (2): 129-142.

Ladouceur, R., A. Gaboury, M. Dumont and P.Rochette. 1988. “Gambling: RelationshipBetween the Frequency of Wins and Irratio-nal Thinking,” Journal of Psychology 122:409-414.

3 7

Laundergan, J. C. 1992. Measuring PathologicalGambling in a General Population and aChemically Dependent Population UsingSOGS and a Modified SOGS (MOGS). Pa-per presented at the National Conference onGambling Behavior. Cleveland, OH.

Laundergan, J. C., J. M. Schaefer, K. F. Eckhoff andP. L. Pirie. 1990. Adult Survey of Minne-sota Gambling Behavior: A Benchmark,1990. Report to the Minnesota Departmentof Human Services, Mental Health Division.

Lesieur, H. R. 1994. “Epidemiological Surveys ofPathological Gambling: Critique and Sugges-tions for Modification,” Journal of GamblingStudies 10 (4): 385-398.

Lesieur, H. R. and S. B. Blume. 1987. “The SouthOaks Gambling Screen (SOGS): A New In-strument for the Identification of Pathologi-cal Gamblers,” American Journal of Psychia-try 144: 1184-1188.

Lesieur, H. R., S. B. Blume and R. M. Zoppa. 1986.“Alcoholism, Drug Abuse, and Gambling,”Alcoholism: Clinical and Experimental Re-search 10: 33-38.

Lesieur, H. R. and R. Klein. 1985. Prisoners, Gam-bling and Crime. Paper presented at theMeetings of the Academy of Criminal Jus-tice Scientists.

Lesieur, H. R. and R. J. Rosenthal. 1991. “Patho-logical Gambling: A Review of the Literature(prepared for the American Psychiatric As-sociation Task Force on DSM-IV Committeeon Disorders of Impulse Control Not Else-where Classified),” Journal of GamblingStudies 7 (1): 5-40.

McKinley, R. 1996. Personal Communication. (RobMcKinley, Gambling Program Coordinator,Pine Grove Recovery Center).

Mississippi Council on Compulsive Gambling. 1996.Background Information. Material providedby the Mississippi Council on CompulsiveGambling.

National Council on Problem Gambling. 1994. TheNeed for a National Policy on Problem andPathological Gambling in America. Reportprepared by the Research and Public PolicyCommittees of the National Council on Prob-lem Gambling.

Omnifacts Research. 1993. An Examination of thePrevalence of Gambling in Nova Scotia. Re-port to the Nova Scotia Department ofHealth, Drug Dependency Services.

Rosecrance, J. 1985. “The Next Best Thing: A Studyof Problem Gambling,” International Jour-nal of the Addictions 20: 1727-1740.

Rosecrance, J. 1988. Gambling Without Guilt: TheLegitimation of an American Pastime.Belmont, CA: Wadsworth.

Sheffield, C. 1996. “Secondary DevelopmentEmerges in Casino’s Wake,” MississippiBusiness Journal (November 4, 1996): 20-22.

Smith, G. J., R. A. Volberg and H. J. Wynne. 1994.Gambling and Problem Gambling in Alberta.Report to the Alberta Lotteries and GamingCommission.

Sommers, I. 1988. “Pathological Gambling: Esti-mating Prevalence and Group Characteris-tics,” International Journal of the Addictions23: 477-490.

Thompson, W. N., R. Gazel and D. Rickman. 1996.The Social Costs of Gambling in Wisconsin.Report to the Wisconsin Policy Research In-stitute. Volume 9, No. 6.

Volberg, R. A. 1988. Compulsive Gambling Treat-ment Program Evaluation: Final Report.Report to the New York State Office of Men-tal Health.

Volberg, R. A. 1992. Gambling Involvement andProblem Gambling in Montana. Report tothe Montana Department of Corrections andHuman Services.

Volberg, R. A. 1993. Gambling and Problem Gam-bling in Washington State. Report to theWashington State Lottery.

3 8

Volberg, R. A. 1994a. “Assessing Problem and Patho-logical Gambling in the General Population:A Methodological Review.” In Gambling inCanada: The Bottom Line, Colin Campbell(ed). Vancouver: Simon Fraser UniversityPress.

Volberg, R. A. 1994b. Gambling and Problem Gam-bling in Saskatchewan. Report to theMinister’s Advisory Committee on SocialImpacts of Gaming.

Volberg, R. A. 1994c. “The Prevalence and Demo-graphics of Pathological Gamblers: Implica-tions for Public Health,” American Journalof Public Health 84 (2): 237-241.

Volberg, R. A. 1995a. Gambling and Problem Gam-bling in Georgia. Report to the Georgia De-partment of Human Services.

Volberg, R. A. 1995b. Gambling and Problem Gam-bling in Iowa: A Replication Survey. Reportto the Iowa Department of Human Services.

Volberg, R. A. 1995c. Wagering and Problem Wa-gering in Louisiana. Report to the Louisi-ana Economic Development and GamingCorporation.

Volberg, R. A. 1996a. Gambling and Problem Gam-bling in New York: A 10-Year ReplicationSurvey, 1986 to 1996. Report to the NewYork Council on Problem Gambling.

Volberg, R. A. 1996b. “Prevalence Studies of Prob-lem Gambling in the United States,” Jour-nal of Gambling Studies 12 (2): 111-128.

Volberg, R. A. and M. W. Abbott. 1994. “LifetimePrevalence Estimates of Pathological Gam-bling in New Zealand,” International Jour-nal of Epidemiology 23 (5): 976-983.

Volberg, R. A. and M. G. Dickerson (eds.). 1996. Jour-nal of Gambling Studies, Special Issue onInternational Prevalence Studies and Re-lated Treatment Developments. Vol. 12, No.2.

Volberg, R. A., D. C. Reitzes and J. Boles. “Explor-ing the Links Between Gambling, ProblemGambling and Self-Esteem.” Deviant Behav-ior, submitted for review.

Volberg, R. A. and E. Silver. 1993. Gambling andProblem Gambling in North Dakota. Reportto the North Dakota Department of HumanServices, Division of Mental Health.

Volberg, R. A. and H. J. Steadman. 1988. “RefiningPrevalence Estimates of Pathological Gam-bling,” American Journal of Psychiatry 145:502-505.

Volberg, R. A. and R. M. Stuefen. 1991. Gamblingand Problem Gambling in South Dakota. Re-port to the Governor’s Office of South Da-kota.

Volberg, R. A. and R. M. Stuefen. 1994. Gamblingand Problem Gambling in South Dakota: AFollow Up Survey. Report to Citizens Unit-ing for Gambling Reform.

Walker, M. B. 1992. The Psychology of Gambling.Oxford: Pergamon Press.

Wallisch, L. 1993. The 1992 Texas Survey of AdultGambling Behavior. Report to the TexasCommission on Alcohol and Drug Abuse.

Wallisch, L. 1996. Gambling in Texas: 1995 Sur-veys of Adult and Adolescent Gambling Be-havior. Report to the Texas Commission onAlcohol and Drug Abuse.

3 9

Appendix A

Increasingly, surveys of gambling and problemgambling in the general population have be-come an essential component in the establish-ment and monitoring of gaming initiatives inAustralia, Canada, Europe and the UnitedStates (Volberg and Dickerson 1996). Infor-mation from such surveys helps identify andminimize the potentially harmful impacts thatlegalized gambling may produce. This proac-tive approach helps ensure that appropriatemeasures are taken to educate the public aboutproblem gambling and that appropriate levelsand types of services for individuals with gam-bling-related difficulties are funded, developedand maintained.

A variety of methodological questions havebeen raised in recent years about research ongambling and problem gambling in the gen-eral population (Dickerson 1993; Lesieur 1994;walker 1992). Questions about surveys of gam-bling and problem gambling in the generalpopulation raised by Lesieur (1994) and Walker(1992) are issues common to all social scienceand survey research. Every researcher whouses survey methods must be concerned withrespondent denial and with rising refusal ratesin telephone surveys. However, these concernsare best addressed through careful attentionto good survey design, including the use of ap-propriate sampling frames and well-designedquestionnaires, as well as an emphasis on ad-equate interviewer training.

Issues related to the substantive topic of gam-bling and problem gambling include questionsabout the validity and reliability of the SouthOaks Gambling Screen as well as challengesto assumptions about the nature of gamblingand problem gambling built into the originalversion of the South Oaks Gambling Screen(Dickerson 1993; Volberg 1994a). In responseto questions about these assumptions, work toimprove the South Oaks Gambling Screen andto extend our understanding of how well the

South Oaks Gambling Screen operates in gen-eral population surveys was carried out in NewZealand in the early 1990s.

❐❐❐❐❐ Development of theSouth Oaks Gambling Screen

Only one survey of gambling and gambling-related difficulties in the general populationwas conducted in the United States prior to1980 (Kallick, Suits, Dielman and Hybels1979). Between 1980 and 1990, state-widesurveys of gambling and problem gamblingwere carried out in California, Connecticut,Iowa, Maryland, Massachusetts, Minnesota,New Jersey, New York and Ohio (Christiansen/Cummings Associates 1992; Laundergan,Schaefer, Eckhoff and Pirie 1990; Sommers1988; Volberg 1994c; Volberg and Steadman1988) as well as in the Canadian province ofQuebec (Ladouceur 1993).

Since 1990, baseline prevalence surveys ofgambling and problem gambling have beencompleted in Georgia, Louisiana, Montana,North Dakota, South Dakota, Texas, Washing-ton State and Wisconsin (Thompson, Gazel andRickman 1996; Volberg 1992, 1993, 1995a,1995c, 1996b; Volberg and Silver 1993; Volbergand Stuefen 1991; Wallisch 1993) as well as inthe Canadian provinces of Alberta British Co-lumbia, Manitoba, New Brunswick, NovaScotia and Saskatchewan (Angus Reid Groupand Gemini Research 1994; Baseline MarketResearch 1992; Criterion Research 1993;Omnifacts Research 1993; Smith, Volberg andWynne 1994; Volberg 1994b). A national preva-lence survey of gambling and problem gamblingwas carried out in New Zealand (Abbott andVolberg 1991, 1992, 1996).

All but three of the prevalence surveys carriedout since 1980 have based on the South OaksGambling Screen (Lesieur and Blume 1987).

4 0

❐❐❐❐❐ The Accuracy of SOGS-BasedPrevalence Rates

The South Oaks Gambling Screen was origi-nally developed for use as a clinical screen andwas adapted slightly in 1986 for use in generalpopulation surveys (Volberg and Steadman1988). Like all screens to detect physical andpsychological maladies, the South Oaks Gam-bling Screen is expected to make errors in clas-sification although misclassification has verydifferent consequences in clinical settings thanin research in the general population.

Misclassification can occur when an individualwithout the malady in question is misdiag-nosed as having the malady. This type of clas-sification error is called a false positive (seetable below).

The Ohio and Wisconsin surveys were basedon screens for problem gambling that havenever been experimentally validated. The re-searchers in Minnesota made such substantialchanges to the South Oaks Gambling Screenthat they designated their instrument theSOGS-M. These changes made it difficult todirectly compare the results of the survey inMinnesota with other SOGS-based surveys(Laundergan 1992).

The South Oaks Gambling Screen is a 20-itemscale based on the diagnostic criteria for patho-logical gambling (American Psychiatric Asso-ciation 1980). Weighted items on the SouthOaks Gambling Screen include hiding evidenceof Gambling, spending more time or moneygambling than intended, arguing with familymembers over gambling and borrowing moneyto gamble or to pay gambling debts. In devel-oping the South Oaks Gambling Screen, spe-cific items as well as the entire screen weretested for reliability and validity with a vari-ety of groups, including hospital workers, uni-versity students, prison inmates and inpatientsin alcohol and substance abuse treatment pro-grams (Lesieur and Blume 1987; Lesieur,Blume and Zoppa 1986; Lesieur and Klein1985).

Surveys of gambling and problem gamblingdirected by Volberg and her associates since1990 have used a revised version of the SouthOaks Gambling Screen developed in NewZealand (Abbott and Volberg 1991, 1996). Inrevising the South Oaks Gambling Screen, thepreliminary section of the questionnaire wasexpanded to collect more detailed informationabout gambling frequency and expenditures inthe general population. In addition, theweighted items of the screen were expandedto assess both lifetime and current prevalenceof problem and pathological gambling. To de-termine if the changes made to the South OaksGambling Screen had any impact on reportedprevalence rates, the revised South Oaks Gam-bling Screen was tested in Iowa in 1991. Thedifference in the prevalence rates for these twoquestionnaires was 0.1% (Volberg and Stuefen1991).

Determining the size of each type of classifica-tion error and correcting for these errors is thekey to establishing more accurate prevalenceestimates. Research in New Zealand used thepositive predictive value and efficiency ap-proaches in efforts to correct lifetime and cur-rent prevalence rates of pathological gambling(Abbott and Volberg 1992, 1996).

The positive predictive value approach is basedon existing information about the sensitivityand specificity of an instrument.1 While thelifetime South Oaks Gambling Screen is knownto have high sensitivity, the specificity of thescreen has differed across different groups in

1 Sensitivity is a measure of the capacity of an in-strument to accurately detect the presence of a par-ticular condition (true and false positives). Speci-ficity is a measure of the rate at which an instru-ment detects true and false negatives.

Condition

Pathological Non-Pathological

Pathological True Positive False Positive

Non-Pathological False Negative True Negative

Classification

4 1

the population (Lesieur and Blume 1987). Sen-sitivity and specificity have never been deter-mined for the current South Oaks GamblingScreen. While the New Zealand researcherswere able to correct the lifetime prevalence ratefor false positives, it proved difficult to makethe correction for false negatives. The research-ers concluded that until more is known aboutthe rate at which the lifetime South Oaks Gam-bling Screen misclassifies pathological gam-blers as non-pathological, the usefulness of thepositive predictive value approach in revisinglifetime prevalence estimates was limited.

The efficiency approach was possible in NewZealand because a two-phase research designwas used to identify true pathological gam-blers among particular groups of respondents(Abbott and Volberg 1992). In the New Zealandstudy, true pathological gamblers were identi-fied in each of four groups included in the sur-vey: (1) probable pathological gamblers, (2)problem gamblers, (3) continuous gamblers and(4) non-continuous gamblers. No error rate wasdetermined for respondents in the NewZealand study who did not acknowledge gam-bling on a regular basis. The efficiency ap-proach involved calculating the rate of truepathological gamblers in each group and di-viding this number by the total number of re-spondents in the sample. The efficiency ap-proach resulted in a revised current prevalenceestimate in New Zealand that was 0.1% higherthan the uncorrected current prevalence rate.

This revised estimate rested on the conserva-tive assumption that there were no false nega-tives among individuals who do not gambleregularly. While the error rates in the fourgroups have an impact on the overall preva-lence rate, the size of the error rate for eachgroup will have a different impact because ofthe differing sizes of these groups in the popu-lation. Even if the number of true pathologi-cal gamblers in the false negative group oramong respondents who do not gamble regu-larly were extremely small, the relatively largesize of these groups contributes to a noticeablyhigher overall prevalence rate. For example,if the non-gambling group is assumed to in-

clude a very small number of pathological gam-blers (1%), the prevalence estimate increasesby 0.7%.

The New Zealand researchers concluded thatthe lifetime South Oaks Gambling Screen isvery good at detecting pathological gamblingamong those who currently experience the dis-order. However, as expected, the screen iden-tifies at-risk individuals at the expense of gen-erating a substantial number of false positives.The current South Oaks Gambling Screen pro-duces fewer false positives than the lifetimemeasure but more false negatives and thusprovides a weaker screen for identifying patho-logical gamblers in the clinical sense. How-ever, the greater efficiency of the current SouthOaks Gambling Screen makes it a more usefultool for detecting rates of change in the preva-lence of problem and pathological gamblingover time (Abbott and Volberg 1996).

Although there are questions about the valid-ity of applying results from research in NewZealand to studies in the United States, theNew Zealand research does suggest that esti-mates of the lifetime prevalence of problem andprobable pathological gambling over-state theactual prevalence of pathological gambling.However, since the lifetime South Oaks Gam-bling Screen does a good job of identifyingpathological gamblers in the general popula-tion, information about the characteristics ofthese respondents is valuable in planning theimplementation and development of servicesfor pathological gamblers in the community.The New Zealand research further suggeststhat estimates of the current prevalence ofproblem and probable pathological gamblingare quite accurate. In future research on gam-bling and problem gambling in Mississippi, itwill be essential to collect information on cur-rent prevalence so that the magnitude ofchanges in the prevalence of gambling-relateddifficulties in Mississippi can be accuratelyassessed.

T he Social Science Research Center(SSRC) has over 45 years experienceanalyzing economic and social issues in

Mississippi, the southeastern region and the nation.Some 45 research fellows, cooperating faculty andresearch associates, supported by an administrativestaff and graduate and undergraduate researchassistants annually conduct around 20-25 sponsoredand numerous unsponsored research projects. TheSSRC is a university-level, multi-disciplinaryresearch unit, organized with university-wideresponsibilities under supervision of the VicePresident for Research. The SSRC has administra-tive responsibilities, for certain programs, to theDirector of the Mississippi Agricultural andForestry Experiment Station (MAFES). The Centeralso serves as the administrative home of theNational Black Graduate Student Association andhas had administrative responsibility for the conductof the Mississippi Alcohol Safety EducationProgram since its inception as an offshoot of anSSRC research project in 1972. Total expendituresfor research and development activities in theCenter exceed three million dollars a year. TheCenter has a strong tradition of research,development and evaluation projects in the areas ofsocial and economic development, the family andchildren, community development, alcohol safetyand substance abuse, race relations, information-agesocietal monitoring, natural resources, gaming,social services, and crime, delinquency and justice.

Housed in the Mississippi Research and TechnologyPark adjacent to the Mississippi State Universitycampus, SSRC researchers benefit from theextensive research infrastructure of the Centerwhich has seven major research and developmentdivisions: the Monitor Laboratory, the SurveyResearch Unit (SRU), the Social and EconomicDevelopment Research Program, the MississippiAlcohol Safety Education Program (MASEP), theMississippi Crime and Justice Research Unit(MCJRU), the Family and Children Research Unit(FCRU) and the Gaming Research Group. Withinthis structure the Center provides support for socialscience researchers to develop individual researchagendas as well as collaborative, interdisciplinaryresearch endeavors.

Director : Dr. Arthur G. CosbySeries Editor: Dr. J. Gipson Wells

Associate Editor: Tan H. Tsai

Mississippi State University does notdiscriminate on the basis of race, color,

religion, national origin, sex, age,handicap/disability, or veteran status.

SSRC Home Page:http://www.ssrc.msstate.edu/

http://www.ssrc.msstate.edu

Recent Titles in the Social Research Report Series:

93-1 Education and Vicious Circles in Mississippi (MAFES Publication D-8189)

93-2 Families and Households in Mississippi: 1960-1990 (MAFES Publication D-8190)

93-3 Health in the Mississippi Delta: Prevention and Policy (SSRC Publication)

93-4 The Life Activities Inventory as a DUI Countermeasure: An Attempted Replication(MASEP/SSRC Publication)

93-5 Flood Control and Water Management in the Yazoo-Mississippi Delta (SSRC Publica-tion)

93-6 Background, Development and Utilization of Drinker Assessment in Mississippi AlcoholSafety Education Program Classes (MASEP/SSRC Publication)

94-1 Best Management Practices for Manufacturing in the Mississippi Delta (SSRC Publication)

94-2 Patterns of Financial Expenditures among College Seniors in Mississippi Public Univer-sities (MAFES Publication D-8542)

95-1 Social and Economic Characteristics of Mississippi Counties, 1990 (MAFES PublicationD-8775)

95-2 A Summary of Mississippi 5-Year Plan: Family Preservation/Family Support Services—The Mississippi Department of Human Services Planning Document (SSRC Publication)

95-3 Geodemographic Approaches to Identifying U.S. Furniture Markets (SSRC Publication)

96-1 Stability and Change in Mississippians’ Political and Partisan Views: Insights from 14Years of Opinion Polling (SSRC Publication)

96-2 Social and Economic Change in the Mississippi Delta: An Update of Portrait Data(SSRC Publication)

96-3 Characteristics of Convicted Drinking Drivers Who Attended the Mississippi AlcoholSafety Education Program, 1992-1994: A Statistical Summary (MASEP/SSRC Publica-tion)

96-4 Recidivism Among Convicted Drinking Drivers: An Analysis of DUI Offenders Referred tothe Mississippi Alcohol Safety Education Program in 1992 (MASEP/SSRC Publication)

96-5 The Social Cost of Growing-Up in Rural America: Rural Development and SocialChange During the Twentieth Century (MAFES Publication D-8985)

96-6 Mississippi Youth: Problems and Issues (MAFES Publication D-8983)

To order publications, write or call:Series EditorSocial Science Research CenterP.O. Box 5287Mississippi State, MS 39762-5287(601) 325-8055

gambling and problem gambling in mississippi · problem gambling uses this term to indicate all of...

Documents