criminal behaviour and pathological gambling

15
Criminal Behavior Associated with Pathological Gambling Gerhard Meyer Michael A. Stadler Institute for Psychology and Cognition Research, University of Bremen, Germany The influence of addictive gambling behavior on criminal behavior was examined in this study. A sample of pathological gamblers from in- and outpatient treatment centers and self-help groups (n = 300) and a sample of high and low frequency gamblers from the general population and army (n = 274) completed a comprehensive questionnaire which assessed social attachment, personality, pathological gambling and criminal be- havior variables. The causal analysis of a Lisrel Model leads to the following results: addictive gambling behavior is an important criminogenic factor. This predisposing factor alone cannot sufficiently explain criminal behavior associated with pathological gambling. Personality variables also directly influence the intensity of criminal behavior. Social attachment variables have only an indirect effect. As far as property offenses are concerned, it was found that the direct causal effect of addiction behavior is greater than that of personality. INTRODUCTION The majority of clinical studies on pathological gamblers in in- and outpatient treatment centers and self-help groups indicates that a high percentage of subjects committed offenses in order to finance their gambling. Research has shown the existence of a causal relation- ship between criminal behavior and pathological gambling (Meyer, Althoff and Stadler, 1998; Rosenthal and Lesieur, 1996). Journal of Gambling Studies Vol. 15(1), Spring 1999 © 1999 Human Sciences Press, Inc. 29 This research was supported by a grant from the Federal Health Ministry. Address correspondence to Dr. Gerhard Meyer, Institute for Psychology and Cognition Re- search, University of Bremen, Grazer StraBe 4, 28359 Bremen, Germany.

Upload: irma-novalic

Post on 18-Jul-2016

14 views

Category:

Documents


0 download

DESCRIPTION

sdhfausdfh

TRANSCRIPT

Page 1: Criminal Behaviour and Pathological Gambling

Criminal Behavior Associatedwith Pathological Gambling

Gerhard MeyerMichael A. Stadler

Institute for Psychology and Cognition Research, University of Bremen, Germany

The influence of addictive gambling behavior on criminal behavior was examined inthis study. A sample of pathological gamblers from in- and outpatient treatment centersand self-help groups (n = 300) and a sample of high and low frequency gamblers fromthe general population and army (n = 274) completed a comprehensive questionnairewhich assessed social attachment, personality, pathological gambling and criminal be-havior variables. The causal analysis of a Lisrel Model leads to the following results:addictive gambling behavior is an important criminogenic factor. This predisposingfactor alone cannot sufficiently explain criminal behavior associated with pathologicalgambling. Personality variables also directly influence the intensity of criminal behavior.Social attachment variables have only an indirect effect. As far as property offenses areconcerned, it was found that the direct causal effect of addiction behavior is greaterthan that of personality.

INTRODUCTION

The majority of clinical studies on pathological gamblers in in-and outpatient treatment centers and self-help groups indicates that ahigh percentage of subjects committed offenses in order to financetheir gambling. Research has shown the existence of a causal relation-ship between criminal behavior and pathological gambling (Meyer,Althoff and Stadler, 1998; Rosenthal and Lesieur, 1996).

Journal of Gambling Studies Vol. 15(1), Spring 1999© 1999 Human Sciences Press, Inc. 29

This research was supported by a grant from the Federal Health Ministry.Address correspondence to Dr. Gerhard Meyer, Institute for Psychology and Cognition Re-

search, University of Bremen, Grazer StraBe 4, 28359 Bremen, Germany.

Page 2: Criminal Behaviour and Pathological Gambling

JOURNAL OF GAMBLING STUDIES

Lesieur (1984) describes the development of excessive gamblingincluding gambling-related crimes within the context of the gambler'sgrowing indebtedness. This is a self-intensifying system in which thegambler gets embroiled in a spiral of options and involvement whichfinally destroys him. Initial winning streaks are followed by losses, thechase philosophy takes over as the driving force, the effort to even outlosses by means of ever increasing stakes, until the available resourceshave finally been depleted. The gambler is under continuous pressureto ensure that he can finance his gambling, this being his only alterna-tive for a speedy solution to his problems. Every new option that mayserve to finance his gambling (such as borrowing) implies further in-volvement. However, ever increasing obligations are opposed by everdecreasing opportunities that may help the gambler solve his problems.

Meyer and Bachmann (1993) concentrate more on the addictivequality of pathological gambling, in analogy to the crime-related con-sequences of drug abuse. The intensity of gambling increases and sodoes the gambler's expenditure. As addictive gambling behavior de-velops, the gambler's perception becomes restricted to the task of se-curing money for further gambling. The pressure to act increases assoon as the gambler has used up all personal and legal financialresources. This is because the gambler will constantly try to obtainmoney to satisfy his craving to gamble (to generate feelings of pleasureor avoid uncomfortable feelings). It increases to such an extent thatthe gambler overcomes higher and higher levels of inhibition and fi-nally commits offenses in order to satisfy his financial and psychologi-cal needs.

Clinical (Blaszczynski and McConaghy,1994) and forensic (Kro-ber, 1991; Schulte, 1994; Meyer and Fabian, 1996) studies show thatnot all offenses committed by pathological gamblers are necessarilygambling-related. Some had already committed crimes before their ac-tual "gambling career" had begun. This fact is essentially explained bythe presence of a subgroup of pathological gamblers who exhibit anti-social personality characteristics—also seen as an indication of malad-justment to social norms—which are generally considered to be acriminogenic factor.

In order to take into account the complexity of criminal behaviorand its development, it is necessary to consider personality dimensionswhen developing explanatory models. Personality theories in particu-lar play an important role as far as multi-factor approaches are con-

30

Page 3: Criminal Behaviour and Pathological Gambling

GERHARD MEYER AND MICHAEL A. STADLER

cerned. Their basic assumption—that individuals with certain behav-ioral traits are more prone to commit offenses, and that especially per-sistent crimes are more likely than in the case of other individuals—has been confirmed in recent longitudinal research. The relationshipsbetween personality features and crime are however only moderate,and their line of development is by no means "linear" (Losel, 1993).Only once linked to sociopsychological aspects, by considering socialrelationships, do they become important.

The control theory is a current criminological explanation (Gott-fredson and Hirschi, 1990) for criminal behavior. This theory uses theindividual's degree of attachment to society and its institutions as ameasure of behavioral adjustment. These authors interpret criminalbehavior to be an action that serves the purpose of achieving short-term and quick satisfaction of needs and further, that it is carried outby people with little self-control, who avoid or reject control. Socializa-tion and its main institutions such as school and family, work andtraining play a major role when acquiring self-control. Social learningprocesses explain the development of attachment, its intensity, preser-vation and consolidation. Here, there is an overlap with socializationconcepts.

Due to the obvious shortcomings of mono-causal models for crim-inal behavior in general and especially for pathological gambling, thepresent research project is based on an open concept, an offender-oriented multi-factor approach. Even if this theory means that thehope of explaining criminal behavior by a closed theory will have to beabandoned, it does at least allow an empirical examination of themeaning of single factors in order to highlight risk factors.

PURPOSE OF THE STUDY

The objective of this study is to investigate if criminal behaviorassociated with pathological gambling is the result of addictive gam-bling behavior. To allow for the complexity of this development, fur-ther possible predisposing factors, such as personality and socialattachment features were included in the study, based on an offender-oriented multi-factor approach. The aim was to test the various hypo-thetical variables against each other in a causal model (Lisrel) and to

31

Page 4: Criminal Behaviour and Pathological Gambling

JOURNAL OF GAMBLING STUDIES

determine if, and to what extent, they influence the committing ofcriminal acts.

METHOD

Subjects and Procedure

In order to determine the influence of pathological gambling be-havior on criminal behavior it was necessary to establish two groups ofsubjects. A group of pathological gamblers should be compared with agroup of high and low frequency gamblers in order to investigate thedifference between pathological and non-pathological gambling in-stead of the influence of gambling as such. This might have been thecase if pathological gamblers had been compared with a sample of thegeneral population.

In order to set up a sample of pathological gamblers from in- andoutpatient treatment centers and self-help groups, the following pro-cedure was chosen. By previous arrangement questionnaires were sentto selected in- and outpatient treatment centers with the request tosend them on to their clients/patients, to actively encourage theircompletion and thus to guarantee a return flow. In this way 53% of thegamblers being treated or counselled in outpatient institutions duringthis survey and 82% of the gamblers at inpatient treatment centerswere taken into consideration for this study. Further appointmentswere made with self-help groups (Gamblers Anonymous) in which thequestionnaires were distributed, queries clarified and questionnairessubsequently collected again. Here the response rate was 93%. In thisway the whole sample of pathological gamblers taken into considera-tion for our study comprised 637 questionnaires returned. This corre-sponds to an average response rate of 71%.

The sample of high and low frequency gamblers was taken fromcasinos, gaming arcades and pubs in different German cities. Eighthundred twenty-one people were asked and only 108 were willing tocomplete the questionnaire (13%). As this procedure to win subjectswas very ineffective, we took a further sample of 225 army membersinto the group of high and low frequency gamblers who all returnedthe completed questionnaires. So our second group consisted of 333high and low frequency gamblers with an average response rate of

32

Page 5: Criminal Behaviour and Pathological Gambling

GERHARD MEYER AND MICHAEL A. STADLER

32%. The 637 pathological gamblers in treatment (PGT) and the 333high and low frequency gamblers (HLG) were subjected to furtherselection: factors of exclusion were female gender, denial tendenciesor the last phase of regular gambling occurred more than a year ago.This led to a reduction of the final sample of n = 300 (PGT) andn = 274 (HLG) respectively.

The described difficulties of establishing two samples of gamblersindicate that any parallelization of both groups regarding sociodemo-graphic features was impossible. Yet the average ages of the PGT-group(32.6 years) and HLG-group (25.5 years) differ only slightly. The levelof education of HLG is slightly higher than that of PGT.

When asked about the type of gambling that caused gamblers un-dergoing treatment problems (multiple responses), 91.3% of the gam-blers named German-styled slot machines; 42.7% exclusively namedthem as their gambling problem. Nearly a third of the treated gam-blers reported gambling in casinos—especially slot machines(30.1%)—and illegal card and dice games for money (31.1%) to beproblematic. In comparison, lottery, horse racing and speculation onthe stock market only play a subordinate role as problematic forms ofgambling.

Measures

The standardized (cross-section) survey was conducted by meansof an extensive questionnaire which contains self-constructed items aswell as scales from psychometric tests. The following variables/scaleswere covered:

Sociodemographic data: Age, Gender, Marital Status, Education, Profes-sion.Social attachment: Change in Living Conditions, Relationship toward Fa-ther/Mother, Parental Education Methods, Emotional and Social At-tachment, Involvement in Conventional Activities, Belief in SocialRules (items based upon the "Questionnaire used in the InternationalComparative Pilot Study on Self-Reported Criminal Behavior of Ado-lescents").Addiction features: Severity of Pathological Gambling Behavior ("ShortQuestionnaire of Gambling Behavior," KFS); Craving to Gamble (items

33

Page 6: Criminal Behaviour and Pathological Gambling

based on the "Liibeck Questionnaire of Craving-Risk-Relapse," LCRR);Duration of Regular Gambling, Amount of Debt, Relation of MonthlyLosses/Income, Emotional and Family-Related Problems, Gambling(Beginning, Type, Frequency etc.).Personality: Aggressiveness, Frankness, Extroversion, Emotionality ("Frei-burg Personality Inventory," FPI-R); Impulsivity ("Trier PersonalityQuestionnaire," TPF); Risk-Motivation ("Risk-Motivation Question-naire," RMF); Antisocial Personality (items based on the "StructuredClinical Interview," SKID-II); Global Mental Stress ("Symptom-Check-list, "SCL-90-R).Intensity of Criminal Behavior: Committing of Different Criminal Of-fenses (Frequency, Duration, Police Contact, Conviction).

Load values were determined in order to judge the intensity ofcriminal behavior. The loading process took place as follows: at firstthe various offenses were given values between 1 and 100 based onlegal assessments and expert ratings (according to Kreuzer, Romer-Klees and Schneider, 1991). For instance theft without aggravating cir-cumstances loads with a value of 2, forgery of documents with a valueof 12 and robbery with a value of 85. These values were multipliedwith load values between 1 to 5 related to the frequency of criminaloffenses committed during the last 12 months of regular gambling.Another value was given as a factor showing whether the behavior of asubject was given attention to the police (2 points) or has been con-demned by court (6 points). A third factor was the time span betweenthe first and last criminal act conceded (1 to 5 points, 5 points for along time span).

RESULTS

Intensity of Criminal Behavior of Pathological Gamblers in Treatment andHigh and Low Frequency Gamblers

Of the pathological gamblers, 89.3% admitted having committedat least one crime during their lifetime, in contrast to only slightlymore than half (51.8%) of the high and low frequency gamblers;59.3% of the pathological gamblers and 22.3% of the comparativegroup admitted having committed at least one criminal offense during

34 JOURNAL OF GAMBLING STUDIES

Page 7: Criminal Behaviour and Pathological Gambling

the last 12 months of (regular) gambling. Thirty-five percent of thetreated gamblers confirmed having had to deal with the police on atleast one occasion and 28.3% have been convicted at least once. Onthe contrary, only a small percentage of the high and low frequencygamblers (6.2 and 3.3% respectively) reported such experiences.

When questioned about the relationship between gambling andfrequency of criminal behavior 35.3% of the pathological gamblersstated that they often or very often obtained money for gambling pur-poses, or for payment of gambling debts by means of "more or lessforbidden acts." Furthermore, approximately one quarter occasionallybehaved in a similar fashion. In comparison, only 8.2% often or veryoften and 20.2% occasionally committed illegal acts that were unre-lated to gambling.

When considering the types of offenses committed during the last12 months of (regular) gambling (Fig. 1), it can be seen that fraudstands out in the case of gamblers undergoing treatment. Every thirdgambler (37.7%) confirms at least one such offense during the givenperiod of time. The frequency in percent is therefore higher than thatfor minor acts such as travel without paying, driving under the influ-ence of alcohol and the consumption of soft drugs. Theft from theworkplace (23.3%) and family (21%) as well as embezzlement (21.7%)are further outstanding types of offenses.

Examination of a Hypothetical Causal Model for Criminal BehaviorAssociated with Pathological Gambling

In recent times causal modeling with the Lisrel Model has beenintroduced to research on gambling behavior (Gupta & Derevensky,1998). The analysis using Lisrel modeling requires an explicit assump-tion of the causal relation between independent and dependent vari-ables. The causes and effects are derived from theoretical considera-tion. The Lisrel Model calculates the empirical causal relationships onthe basis of correlations between the hypothetical constructs.

Theoretical models provide a range of potential predisposing fac-tors for criminal behavior associated with pathological gambling, aswell as fundamental statements concerning their interrelationships.Based on this, a hypothetical causal model can be postulated whichreflects the suspected causal dependencies between the factors "SocialAttachment," "Personality," "Addiction" and "Criminal Behavior."

GERHARD MEYER AND MICHAEL A. STADLER 35

Page 8: Criminal Behaviour and Pathological Gambling

The following hypotheses can be derived from the theoreticalmodels:

H1,2: Inadequate social attachment leads to (1) a pronounced ad-dictive behavior and (2) a higher intensity of criminal be-havior (designated by the positive value of the path coeffi-cient).

36 JOURNAL OF GAMBLING STUDIES

Figure 1Distribution of Pathological Gamblers in Treatment (n = 300) andHigh and Low Frequency Gamblers (n = 274) by Type of Offense

Committed in the Last 12 Months of Regular Gambling(Multiple Responses)

Page 9: Criminal Behaviour and Pathological Gambling

GERHARD MEYER AND MICHAEL A. STADLER

H3,4: A more pronounced impulsive, antisocial, risk-motivatedpersonality structure leads to (3) a more pronounced addic-tive behavior and to (4) a higher intensity of criminal behav-ior (designated by the positive value of the path coefficient).

H5: A more pronounced addictive behavior leads to a higherintensity of criminal behavior (designated by the positivevalue of the path coefficient).

A causal analysis using the Lisrel Model was used to examine whetherthe theoretical relationships correspond to the empirically deriveddata set.

The results of the Lisrel analysis including the intensity values(C1-C3) for property offenses (based on 16 iterations) are presentedin Fig. 2.

Evaluation of the Measuring Model. The Lisrel results indicate thatthe latent variables "Addiction" and "Criminal Behavior" are well mea-sured by the indicator variables:

— The three intensity measures of criminal behavior opera-tionalize criminal behavior well. Their scores for this factor liebetween .72 and .93. The indicator C3 has a reliability coeffi-cient of .86 and is thus the most reliable measurement.

— The scales S0 through to S9 are, without exception, good indi-cators of the hypothetical construct "Addiction." The correla-tions between observed variables and latent variable (factorscores) lie between .45 and .97. The strongest correlation is.97, the correlation of the scale "Severity of Pathological Gam-bling" with the scale "Addiction," i.e. the latent variableexplains 94% of the indicator variables' variance, only 6%remains unexplained and can be ascribed to error in mea-surement and other possible variable effects that may havebeen neglected.

The hypotheses concerning the assignment of the indicator vari-ables to the hypothetical constructs "Social Attachment" and "Person-ality" were only partly confirmed:

37

Page 10: Criminal Behaviour and Pathological Gambling

— The assumed factor "Social Attachment" is operationalizedsatisfactorily by five of the seven scales (scores of .42 to .71,however with low reliability of .17 up to .50).

— The scales "Risk-Motivation," "Impulsivity," "Emotionality,""Global Mental Stress" and "Non-conformal Social Orienta-tion" (most reliable measurement at .59) satisfactorily load(.40 to .77) on the factor "Personality" which is a very broadconstruct. The other indicators only have correlations be-tween .04 and .36 with this factor.

Evaluation of the Structure Model. The signs of the path coefficientsthat were empirically estimated using the Lisrel Model, are in accor-dance with the theoretically postulated signs. The assumed causal rela-tionships explain 60% of the variance of the latent variable "CriminalBehavior" (residual variance: £2

= .40). When considering the broadlyapplied theoretical model this percentage can be rated as more thansatisfactory. The explained variance of the construct "Addiction" is50%. The path coefficients between the exogenous variables "Person-

38 JOURNAL OF GAMBLING STUDIES

Figure 2Results of LISREL Analysis Including the Criminal Behavior

Referring to Property Offense (Completely Standardized Solution)

Page 11: Criminal Behaviour and Pathological Gambling

GERHARD MEYER AND MICHAEL A. STADLER

ality" and "Social Attachment" as well as the endogenous variable "Ad-diction" have the values y12 = .52 and y11 = .25 and thus verify thatthe exogenous variables have a direct influence on addictive behavior.Twenty-seven percent and 6.3% of the variance of the variable "Addic-tion" is determined by the variance of the variables "Personality" and"Social Attachment" respectively. There is a correlation of <p21 = .63between the two exogenous variables, i.e. "Social Attachment" and"Personality" covary and thus the assumption that the predictors areindependent is not plausible.

The path coefficients between "Addiction," "Personality" and"Criminal Behavior" have the values B12 = .48 and y22 = .33 and thusindicate direct causal effects. The variance of the variable "Addiction"explains 23% and the variable "Personality" 10.9% of the variance ofthe variable "Criminal Behavior." Furthermore, "Personality" indirectlyaffects criminal behavior due to a mediatory influence on the addic-tive behavior .25 (.52*.48).

The variable "Social Attachment" has no direct influence on"Criminal Behavior" (y11 = .04). The indirect influence on addictivebehavior is valued at .12 (.25*.48), which emphasises the role "Addic-tion" plays as a mediator.

By taking into account measures of criminal behavior for all typesof offenses, somewhat less of the variance of the latent variable "Crimi-nal Behavior" (54%) can be explained by the postulated causal rela-tionship than by the analysis based on property crimes alone. The pathcoefficients between "Addiction," "Personality" and "Criminal Behav-ior" have the values B12 = .38 and y22 = .40 and implies equally largedirect causal effects. In this solution the variance of the variable "Crim-inal Behavior" is determined by 14.4% of the variance of the variable"Addiction" and by 16% of the variance of the variable "Personality"(Meyer, Altoff, and Stadler, 1998).

Evaluation of the Model as a Whole. The statistical criteria for thequality of adjustment of the theoretical model's structure to the empir-ical data suggest that the solution is well adjusted. The "Goodness-of-Fit-Index (GFI)" that measures the relative amount of variance andcovariance, taking the model as a whole into account, has a value ofGFI = 0.97; i.e. the structure of the model explains 97% of the wholeinitial variance.

The "Adjusted-Goodness-of-Fit-Index (AGFI)," a further measure

39

Page 12: Criminal Behaviour and Pathological Gambling

of the explained variance used in the model which additionally con-siders the number of degrees of freedom, has a value of 0.96 which isalso close to 1.

The "Root-Mean-Square-Residual (RMR)," a measure of the aver-age variance and covariance not explained by the model, has a valueof RMR = 0.06. Thus, also according to this criterion—the aspiredvalue being < .05—the model has achieved an adequate adjustment.

It should be mentioned that, when evaluating the quality of adap-tation, the chi-square value is not a suitable test statistic for the presentinvestigation. Firstly, the assumption of normal distribution is notgiven for all variables and, secondly, the approximations used in thecalculations are based on a random sample correlation matrix. Thechi-square value is also very sensitive to the population size: the biggerthe sample, the higher the chances of a model being rejected. Thereason for this being that small deviations are already significant in thecase of larger N.

The exclusion of unreliable indicators may improve the quality ofadaptation for the model as a whole, but would lead to a reduction ofthe model's contents.

DISCUSSION

Empirical data confirm the theoretical causal model for gambling-related criminal behavior which is based on potential predisposing fac-tors such as social attachment, personality and addictive behavior. Theempirical confirmation of the theoretical model does, however, not im-ply causality. A single set of empirical correlations can always be ex-plained by several—often very different—causal models. The validity ofthe assumed causal relationships can ultimately only be verified by logic.

Alternative causal models are plausible. Empirical research on theaetiological involvement of criminal behavior and substance-relatedaddictions shows, for example, complex connections which also ex-hibit a certain independence. Admittedly the dependency on drugssuch as opiates and cocaine influences the variance of criminal behav-ior, however drug consumption has been found to be an amplifierrather than an initiator of criminal behavior. Based on this back-ground, Speckart and Anglin (1985) designed a causal model in which

40 JOURNAL OF GAMBLING STUDIES

Page 13: Criminal Behaviour and Pathological Gambling

GERHARD MEYER AND MICHAEL A. STADLER

four relationship patterns explain the link between drug consumptionand crime. According to this model criminal behavior may be the re-sult of drug consumption, but may also lead to it. Both behavior pat-terns may, however, also result from the factor deviance and therebymutually influence each other. In the fourth variation they also de-velop simultaneously yet independently of each other. According tothese hypotheses an alternative causal model is plausible: the patho-logical or addictive gambling behavior develops from an existing crimi-nal constellation which, for its part, designates a deviant primary devel-opment. Furthermore, based on research and forensic reports it isknown that personality traits and social attachment do not solely influ-ence the addictive behavior—including gambling addiction—but thatthe reverse is also applicable. Addiction leads to a general decrease ofsocial links up to social isolation and destruction of the social person(Meyer and Bachmann, 1993). At the same time obvious personalitychanges (depression, irritability, loss of self-respect etc.) are identifia-ble as a result of the addictive behavior.

The ascertained rest variance explains that the given system ofhypotheses does not include all relevant variables that may influencethe intensity of criminal behavior. The addition of cognitive factorssuch as "Belief in Social Rules" and "Moral Justification" or factorssuch as "Opportunity" and "Occupational Situation" suitably modifiesthe model.

By trying out different causal effects on the present data or byreducing or increasing the structure of the model, the Lisrel approachadditionally allows the model to achieve a better quality of fit. Thebasic principle of this approach is, however, violated when a given the-ory is modified as a result of the Lisrel analysis. Such a step is onlyvalid as a last resort when the "new" model is tested using additionaldata.

Apart from the evaluation of the model on a statistical level, it isalso important that the model should be plausible and can be substan-tiated on the content level (e.g. as seen from a developmental psychol-ogy perspective: clear time sequence). In addition, it fulfills certainrequirements on the model-theory level: the particular (Addiction,Criminal Behavior) can be derived from the general (Personality, So-cial Attachment) and the particular can be modelled more stably thanthe general.

41

Page 14: Criminal Behaviour and Pathological Gambling

JOURNAL OF GAMBLING STUDIES

CONCLUSIONS

The present study does not allow one to draw conclusions concern-ing (pathological) gamblers in general in Germany. However an exclu-sive examination of the subpopulation pathological gamblers in treatmentinstitutions for whom the data presented here can be considered repre-sentative (as comparisons with German samples show), indicates thatthe criminogenic effect of gambling with a higher addiction potentialcan no longer be neglected from the point of view of criminal politics.

Comparisons with high and low frequency gamblers, as well aswith German crime statistics and results of criminological studies indi-cate that gamblers show a high intensity of criminal behavior (seeMeyer, Althoffand Stadler, 1998).

Gamblers who seek treatment due to gambling, such as gamblingon slot machines and roulette, show that as a consequence of the par-ticipation in legal offers, persons at risk develop a pathological gam-bling behavior. The percentage of gamblers who lose control overtheir gambling behavior is not known in Germany. Undoubtedly themajority of gamblers make problem-free use of the offers during theirleisure time, and therefore only suffer light losses.

Rough estimates of 90 000 to 150 000 gamblers in need of counsell-ing and treatment in Germany (0.11 to 0.19% of the population) whichare based on the empirically as yet unconfirmed assumption of a ratesimilar to that of the demand for therapy by alcoholics and gamblingaddicts, and approximately 30,000 registered bannings from Germancasinos (more than half on request of the gamblers themselves), areonly first clues as to the extent of this problem (Meyer, 1998).

In cases where addiction led to property crimes, the treatment ofpathological gambling behavior is an effective strategy to reduce therisk of relapse as far as criminal offenses are concerned. According toBrown's (1987) clinical experience once the gambling behavior ischanged, no further convictions ensue. However empirical data and adifferentiation between subgroups of pathological gamblers are re-quired in order to make any fundamental statements.

REFERENCES

Blaszczynski, A.P. & McConaghy, N. (1994). Antisocial personality disorder and pathological gam-bling. Journal of Gambling Studies, 2, 129-145.

42

Page 15: Criminal Behaviour and Pathological Gambling

GERHARD MEYER AND MICHAEL A. STADLER

Brown, R.I.F. (1987). Pathological gambling and associated patterns of crime: Comparisons withalcohol and other drug addictions. Journal of Gambling Behavior, 2, 98-114.

Gottfredson, M.R. & Hirschi, T. (1990). A general theory of crime. Stanford: Stanford UniversityPress.

Gupta, R. & Derevensky, J.L. (1998). An empirical examination of Jacob's general theory of addic-tions: Do adolescent gamblers fit the theory. Journal of Gambling Studies, 1, 17-47.

Kreuzer, A., Romer-Klees, R. & Schneider, H. (1991). Beschaffungskriminalilat Drogenabhangiger.Wiesbaden: Bundeskriminalamt.

Krober, H.-L. (1991). Automatenspieler und Roulettespieler—psychiatrische und kriminologischeDifferenzen. Nervenarzl, 62, 670-675.

Lesieur, H.R. (1984). The chase: Career of the compulsive gambler. Rochester, VT: SchenkmanBooks.

Losel, F. (1993). Taterpersonlichkeit. In G. Kaiser, H.-J. Kerner, F. Sack & H. Schellhoss (eds.)Kleines kriminologisches Worterbuch (3. ed.) (pp. 529-540). Heidelberg: C.F. Muller.

Meyer, G. (1998). Glucksspiel—Zahlen und Fakten. In Deutsche Hauptstelle gegen die Sucht-gefahren (ed.) Jahrbuch Sucht 99 (pp. 89-103). Geesthacht: Neuland.

Meyer, G. & Bachmann, M. (1993). Glucksspiel—Wenn der Traum vom Gluck zum Alptraum wird.Heidelberg: Springer.

Meyer, G. & Fabian, T. (1996). Pathological gambling and criminal culpability: An analysis of foren-sic evaluations presented to german penal courts. Journal of Gambling Studies, 1, 33-47.

Meyer, G., Althoff, M. & Stadler, M.A. (1998). Glucksspiel und Delinquenz. Frankfurt/M.: PeterLang Verlag.

Rosenthal, R.J. & Lesieur, H.R. (1996). Pathological gambling and criminal behavior. In L.B. Schle-singer (ed.) Exploration in criminal psychopathology (pp. 149-169). Springfield, II: Thomas.

Schulte, R.-M. (1994). Pathologisches Spielen—Glucksspielsucht als neuer Abhangigkeitstypus?Mitteilungen der LVA Wurttemberg, 9, 328-333.

Speckart, G. & Anglin, M.D. (1985). Narcotics and crime: An analysis of existing evidence for acausal relationship. Behavioral Sciences and the Law, 3, 259-282.

Received September 30, 1998; final revision April 13, 1999; accepted May 28, 1999.

43