Psychology of Addictive Behaviors Copyright 2002 by the Educational Publishing Foundation

2002, Vol. 16, No. 2, 143–149 0893-164X/02/$5.00 DOI: 10.1037//0893-164X.16.2.143

Development and Validation of the Gamblers’ Beliefs Questionnaire
Timothy A. Steenbergh, Andrew W. Meyers, Ryan K. May, and James P. Whelan
University of Memphis

The Gamblers’ Beliefs Questionnaire (GBQ) is a self-report measure of gamblers’ cognitive distortions.
GBQ test items were constructed on the basis of theory, empirical evidence, and expert review. Four
hundred three adults completed the initial set of items, and 21 items were selected to make up the final
GBQ. The factor structure of the GBQ consisted of 2 closely related factors: Luck/Perseverance and
Illusion of Control. The full scale showed good internal consistency (␣ ⫽ .92) and adequate test–retest
reliability (r ⫽ .77). Problem and pathological gamblers scored higher than nonproblem gamblers on the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

GBQ and its factors. GBQ scores were moderately correlated with the duration of gambling sessions
This document is copyrighted by the American Psychological Association or one of its allied publishers.

among problem and pathological gamblers, and there was no relationship between GBQ scores and social
desirability.

Reviews by Ladouceur and Walker (1996) and Toneatto (1999) At present, a limited set of cognitive assessment strategies is
support the role of cognitive variables in gambling behavior and available to clinicians and researchers working with gamblers. In
suggest that problem gamblers hold cognitive distortions that may fact, only two methods of belief assessment can be found in the
contribute to their inability to adequately control their gambling. literature. Early investigations identified cognitive distortions
Furthermore, there is growing support for interventions designed based on indirect inferences from gamblers’ behavior (Henslin,
to modify problem gamblers’ erroneous beliefs (Bujold, Ladou- 1967; Langer, 1975). For example, Henslin (1967) observed that
ceur, Sylvain, & Boisvert, 1994; Ladouceur, Boisvert, & Dumont, regular craps players would coach new players to take their time
1994; Sylvain, Ladouceur, & Boisvert, 1997; Toneatto & Sobell, when throwing dice and that players would throw the dice harder
1990). Given the likely role of faulty belief systems in problem for larger numbers and softer for smaller ones. Although indirect
gambling, the utility of a valid and efficient method for assessing assessment methods such as these were useful for the initial
gambling-related cognitive distortions for clinical and research identification of common cognitive distortions among gamblers,
applications is clear. The present investigation was designed to they are impractical for most clinical and research applications.
develop and evaluate a self-report measure of cognitive distortions Ladouceur and his colleagues (Gaboury & Ladouceur, 1988;
related to gambling. Ladouceur, Mayrand, Dussault, Letarte, & Tremblay, 1984) de-
A number of cognitive distortions have been identified among
veloped a more direct method of cognitive assessment of gamblers.
gamblers (Toneatto, 1999). Examples of cognitive distortions in-
Their “thinking aloud” method required gamblers to verbalize their
clude an illusion of control, which is the belief that one can
thoughts while gambling. These verbalizations were recorded and
influence the outcome of a chance-determined event (Langer,
later coded as either rational or irrational. Criteria for these cate-
1975), and the gambler’s fallacy, which is the belief that future
gorizations were based on objective rules of randomness and
outcomes can be predicted on the basis of past outcomes (e.g., if
heads has come up on 10 consecutive tosses of a coin, then tails is causality. Using the thinking-aloud method, Gaboury and Ladou-
more likely to come up on the next toss). These and other ceur (1988) found that several cognitive distortions could be
gambling-related cognitive distortions have been hypothesized to identified, including references to luck; perceptions of personal
cause gamblers to misperceive the economic utility of gambling skill, control, or ability; personification of a slot machine; state-
and lead to gambling persistence despite the objective expectation ments regarding a system for winning; predictions; and references
of monetary loss (Ladouceur & Walker, 1996). to cause-and-effect relationships. Although the thinking-aloud
method was instrumental in the identification of gambling-related
cognitive distortions, its time-intensive nature has limited its use in
Timothy A. Steenbergh, Andrew W. Meyers, Ryan K. May, and James
laboratory and clinical research. Laboratory investigations that
P. Whelan, Institute for Gambling Education and Research, Department of
Psychology, University of Memphis. have used the thinking aloud method have typically had small
Support for this study was provided by a Centers of Excellence grant sample sizes (e.g., Gaboury & Ladouceur, 1988), and studies of
from the State of Tennessee to the Department of Psychology at the cognitive interventions for problem gamblers have not evaluated
University of Memphis. We especially thank Robert Ladouceur, Daniel cognitive distortions at pre- and posttreatment (e.g., Bujold et al.,
Kirschenbaum, and Edgar Nottingham for their helpful comments during 1994; Ladouceur et al., 1994; Sylvain et al., 1997).
the development of the GBQ. The assistance of Matt Taylor, Angie Findings from studies that have used inferences from gamblers’
Sheffield, Kim Floyd, and Sharon Rekeita in data collection also is
behavior and the thinking-aloud method have provided a founda-
appreciated.
Correspondence concerning this article should be addressed to Andrew
tion for exploring the contribution of dysfunctional cognition to
W. Meyers, Institute for Gambling Education and Research, Department of problem gambling. However, neither assessment method meets the
Psychology, University of Memphis, Memphis, Tennessee 38152. E-mail: needs of clinical researchers. Such research requires a valid, time-
ameyers@memphis.edu efficient assessment method.

143

On the basis of these findings.00003) each indicated that the suggested. and women outnumbered men (57. Overview as gambling status does not dictate the presence or absence of gambling- related irrational beliefs. and a pencil and. Extant literature regarding gambling-related irratio. Minorities were well represented among participants (61. assesses illusion of control and overestimation of the likelihood of This document is copyrighted by the American Psychological Association or one of its allied publishers.g. he or she was of gambling-related irrational thinking provided the conceptual framework provided a consent form. irrational belief assessment. MEYERS. lottery. Five referred to as the GBQ–1). First. AND WHELAN Walker (1992) and Ladouceur and Walker (1996) have offered cognitive distortion. A majority (85%) of participants reported gambling at least once during their lifetime. ␹2(300. Whether these two belief sets exist independently has not been explored. (b) explore nity members (n ⫽ 203) and undergraduate students over the age of 21 its temporal stability. and core beliefs in luck and the illusion of additional 47 community participants were recruited by introductory psy- control were identified. examine the factor structure and internal consistency of the re- 1996. factors with eigenvalues greater than 1. MAY.2 years (SD ⫽ 15.2). These participants were approached by a research associate at the particular recruitment site.6% African American. Most community participants (n ⫽ 156) were recruited from local workplaces and mass-transit waiting areas in two large urban Method areas in the southern United States. A sample of 403 participants was drawn from commu- instrument’s internal consistency and factor structure. Walker’s (1992) and Ladouceur and Walker’s (1996) two-factor models and asked to participate. and that nonparticipation in the study statement “My knowledge and skill in gambling contribute to the likeli. For the purposes of this study. we selected a final The purpose of the present project was to develop and validate item pool and determined the scale’s factor structure and internal a self-report measure of gambling-related cognitive distortions that consistency. we so that higher scores were indicative of greater gambling-related generated 1.4% of other or Study 1 unspecified ethnicity). there Overview seems to be a general consensus that the illusion of control can be The purposes of Study 2 were to (a) refine the GBQ–1 on the differentiated from other types of cognitive distortions that lead to basis of factor analytic and internal consistency data and (b) overestimation of one’s chances of winning (Ladouceur & Walker. lowing statements carefully. if for scale construction. Nongamblers were included in the study. If the individual chose to participate. was 33.6% vs. introductory psychology participant pool and completed the questionnaire these beliefs were transformed into statements. 144 STEENBERGH. examination of the scree plot suggested that ment and allowed respondents to rate the extent of their agreement either one or two factors should be retained. 1992). participants had gambled approxi- The purpose of Study 1 was to identify the domain of gambling- mately 10 times during the past 12 months and had spent over 3 hr related cognitive distortions and to generate an initial pool of items gambling on each of those occasions. for a gamblers’ beliefs questionnaire (GBQ). 42. We conducted a principal-axis Results factor analysis to explore the factor structure of the GBQ–1. N ⫽ 403) ⫽ 4. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.0% Asian. matrix. All items were reverse scored 1965. a gambling involvement questionnaire that provided The initial set of 21 items was sent to three experts in the areas of demographic information and self-reported gambling behavior. Zwick & Velicer. Once identified. 25. All participants were asked to hood that I will make money. roulette. asked to complete the questionnaire at that time. 1986). We used parallel with each item. an effort was made to generate items that would ticipants over the age of 21 years were recruited through the university’s assess general gambling-related cognitive distortions.4%).49. consent was provided.” control over the outcome of a gambling event than in fact he or she has. Procedure. the belief that in a classroom setting. Given the possible variations in beliefs that may chology students and agreed to come to the university and complete the exist across specific forms of gambling activities (e.93). Exploratory factor analysis. winning.. informed of the study. however.000 random data sets for 403 participants and 25 . through neutral to strongly disagree. Walker. would have no adverse consequences. slot machines). Anchors for the scale ranged from strongly agree analysis to determine the correct number of factors to retain (Horn. For example.0% Cau- casian. The average age of the sample validity. (c) provide evidence for its convergent years (n ⫽ 200) at an urban state university. The second factor involves a set of beliefs that lead gamblers Study 2 to overestimate their chances of winning. and 10. psychology of gambling. In addition. and (d) provide evidence for its discriminant validity.001. some items to be less obvious. we reworded lett’s test of sphericity. Community and undergraduate participants were one holds specific skills that increase his or her chances of winning was informed that they were under no obligation to participate. Rate to what extent you agree or there is a core set of beliefs that the gambler has a greater level of disagree with each statement by circling a number.163. The directions were “Read each of the fol- similar two-factor models of gamblers’ cognitive distortions. 3. p ⬍ . An nal beliefs was examined. a questionnaire packet. The experts were asked to provide feedback regarding item content and to recommend additional items that might be Results and Discussion included in the GBQ. We designed the four remaining Method empirical studies to: (a) refine the measure and examine the Participants. that they could used to represent the illusion of control and was transformed into the drop out of the study at any time.” complete the GBQ–1. and the cognitive other measures that were also used in Studies 4 and 5. maining items. In the first of five interrelated studies we developed the initial content of the instrument. and three cognitive psychotherapy. On average. The Kaiser–Meyer–Olkin measure of sampling adequacy (. questionnaire at a group-based administration session. Bart- On the basis of feedback from one of the experts. four new items were and the determinant of the matrix (. Item refinement resulted in a pool of 25 items (hereafter correlation matrix was appropriate for such an analysis. Undergraduate par- blackjack.0 were extracted from the Seven-point Likert-type scales were included with each state. However.

Eigenvalues for the first two GBQ–1 factors were 8.46 . it will eventually pay off and I will make money.62).16.21 I should keep the same bet even when it hasn’t come up lately because it is bound to win.19). .99 factors.40 In the long run.51 When I am gambling.12 I am pretty accurate at predicting when a “win” will occur. perseverance.61 ⫺.19 Where I get money to gamble doesn’t matter because I will win and pay it back. 1.11 If I lose money gambling. or both. . . Total the third. The 21 values (1. tapping a certain number of times. it revealed that all items correlated (r ⫽ .61 My choices or actions affect the game on which I am betting.07 I should keep track of previous winning bets so that I can figure out how I should bet in the future. and one item failed to load above . including two from the Luck/Perseverance and 1. respectively) scores for the GBQ were determined by adding responses for all extracted from the random data sets. .73 .52 . etc. GBQ–1 correlation matrix were not retained. We solution on the basis of a scree test and parallel analysis. The item that loaded internal consistency of the 24-item GBQ–1 scale and its two most heavily on this factor—“When I lose at gambling. variance in GBQ–1 scores.35.34 .35 on at least one of the two factors.67 I have a “lucky” technique that I use when I gamble.14 .30 explore the factor structure of the GBQ.56 My knowledge and skill in gambling contribute to the likelihood that I will make money.26.30. factor loadings of the remaining 24 items. were deleted the eigenvalues of the random data sets.70 .4. 1. er’s (1992) delineation of cognitive distortions that lead gamblers Internal consistency. my losses are not as bad if I don’t tell my loved ones.02 Note. Thirteen items loaded most heavily on the to gambling wins.0% of the score for the entire participant sample was 54. I must maintain that strategy or plan because I know it will eventually come through for me. increase the interpretability of these factors. We examined the We used principal-axis factor analysis with oblimin rotation to factor loadings for each item. We chose a two-factor on either factor. I should continue because I don’t want to miss a win. . items. .48 . . conducted a second exploratory factor analysis to examine the factors accounted for 48. . The mean GBQ The two factors that were retained accounted for 43. respectively) failed to exceed those of remaining items constituted the final version of the GBQ. ⫺.15 . crossing my fingers.75.41 When I lose at gambling.51 . and both factors were because their item–total correlations with the entire scale and with retained.30 Gambling is the best way for me to experience excitement.59 (SD ⫽ 22.84 ⫺.02 If I continue to gamble.12 .05 I have more skills and knowledge related to gambling than most people who gamble. Possible scores could range from 21 to 147. because their eigen.43 . Rate to what extent you agree or disagree with each statement by circling a number. We assessed with coefficient alpha the to overestimate their chances of winning.33 . This factor was consistent with Walk- were related to the illusion of control. and fifth factors extracted from the their respective factors were below . Factor loadings are presented in Table 1. Directions to the respondent were “Read each of the following statements carefully. fourth.12 Gambling is more than just luck.65 . The two This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.11 . .31 There are certain things I do when I am betting (for example. Factor structure and internal consistency of the GBQ.57 My gambling wins are evidence that I have skill and knowledge related to gambling.61 If I am gambling and losing.03 . . .03.21 Even though I may be losing with my gambling strategy or plan.33. GAMBLERS’ BELIEFS QUESTIONNAIRE 145 variables. Fifteen items factors identified in this analysis matched the Luck/Perseverance loaded most heavily on the first factor and shared a common theme and Illusion of Control factors identified in exploratory analyses in the belief that luck. Three items.3% of the variance and were highly This document is copyrighted by the American Psychological Association or one of its allied publishers. fourth.” Each item was rated on a Likert-type scale that ranged from 1 (strongly agree) to 7 (strongly disagree). would eventually lead used for item selection. . . respectively. I will win more money than I will lose gambling. . and 1. 1. . The loaded above .) which increase the chances that I will win. We dropped this item from the item pool.13 Those who don’t gamble much don’t understand that gambling success requires dedication and a willingness to invest some money. The eigenvalues of these factors exceeded factor and one from the Illusion of Control factor. which Luck/Perseverance factor.43 . and fifth factors (1. The second factor consisted of 9 items. . my losses Table 1 Gamblers’ Beliefs Questionnaire Items and Factor Pattern Matrix Item Factor 1 Factor 2 I think of gambling as a challenge. . We performed oblimin rotation to with scores ranging from 21 to 123. .10 . The third. “near misses” or times when I almost win remind me that if I keep playing I will win.56 . holding a lucky coin in my hand. .74 ⫺. ⫺. . I should try to win it back.

On the basis of this assumption. That coefficient was higher than the have been due to the modest eight items in this subscale. 1992).71 We examined the internal consistency for the 21-item GBQ. p ⫽ .4 years (SD ⫽ 5. Luck/Perseverance and Illusion of Control factor scores were .02. Unfortunately.25) and. ␹2(1.00. The cognitive perspective of problem gam- Method bling holds that gambling-related irrational beliefs are important in the maintenance of excessive gambling behavior (e. MAY. N ⫽ 199) ⫽ 0.8% male and were bling problems. For example. N ⫽ 198) ⫽ 5. Scores for this factor could range from 13 to 91. a widely used measure of general GBQ and the Luck/Perseverance factor was favorable. respectively.2% female and 39. p ⫽ .90 and vided the first demonstration of the temporal stability of gambling- .0.84. p ⫽ . related cognitive distortions. and .07. In addition. undergraduates. SD ⫽ 5. ␹2(2.97. t(2.20.6%. SD ⫽ 18. logical gamblers scored higher on the GBQ than nonproblem ethnicity. losses. p ⫽ . We examined the correlations among full-scale and factor scores across the two test In Study 4 we examined the convergent validity of the GBQ by administrations. Retest correlations for This document is copyrighted by the American Psychological Association or one of its allied publishers. Results and Discussion bling behavior. 146 STEENBERGH. range: 8 –53). A few studies have Walker. illusion-of-control beliefs may play in the development of cogni. higher levels of completed Study 2 agreed to participate in further research for course irrational beliefs should be found among individuals with gam- credit. Overview The purpose of Study 3 was to determine the temporal stability Study 4 of the GBQ and its factors within an adult undergraduate popula- tion. and those of other ethnic background ⫽ 8. The evidence presented here sup. In this study we on an average of 7.g.198) ⫽ 1. and reported a reliability coefficient of .77. data from other studies of gamblers’ beliefs makes interpretation tion of gambling-related cognitive distortions (Ladouceur & of the coefficient reported here difficult. Those who con- 14. age.77. Whether this difference was possibility that illusion of control is not a unitary dimension. higher levels of cognitive distortions would be expected to was 25. examined (a) whether participants identified as problem or patho- There was no difference in gender.92. For example. The mean age of the sample lems. it may also serve to disconfirm such beliefs.6%).71. Furthermore. exploring its relationship with measures of problem gambling and gambling behavior. This belief that one can influence the outcome of a chance event has been termed the illusion of control and has been linked to gam. however. Coefficient alpha for the GBQ was .37. among those with gambling prob- ethnically diverse (African American ⫽ 22. Caucasian ⫽ 68. participants gambled be correlated with lengthier gambling sessions.2. The test–retest reliability data reported here were gathered from tive processes related to overestimation of winning.8%. t(2. Those who chose to participate in this study the question remains whether this factor structure applies to a gambled more frequently than nonparticipants from the undergrad- population of problem gamblers. due to measurement error or reflected actual differences in the Further study of the factor structure measured by the GBQ is stability of the constructs measured by the GBQ remains to be warranted. The Illusion irrational beliefs. examined the temporal stability of measures of irrational thinking.188) ⫽ 2.77. indirectly linked to the other items.183) ⫽ 0. and alphas for the Luck/ The examination of the test–retest reliability of the GBQ pro- Perseverance factor and the Illusion of Control factor were . In this Procedure.09. Participants completed the GBQ during class and again Overview within 2 weeks of the initial administration. A subsample of undergraduate students (N ⫽ 93) who & Walker. This may 2-week test–retest interval. AND WHELAN are not as bad if I don’t tell my loved ones”—appeared to be or average duration of gambling sessions. Langer (1975) found that individuals placed higher wagers on games that fostered an illusion of control.07. Walker.86. Although frequent gam- bling may expose individuals to more opportunities to solidify Study 3 gambling-related irrational beliefs. over the past year. 1996). It is unclear what effect gambling frequency may have on the stability of irrational beliefs. the lack of empirical The GBQ was designed to assess the two-factor conceptualiza. Internal consistency for the evaluated the Belief Scale. t(2. ports the ability of the GBQ to measure these two related con. We used Pearson correlation coefficients to examine the rela- tionship between GBQ full-scale and factor scores. uate pool did. The mean factor score in this study was 24. Malouff and Schutte (1986) gambling-related cognitive distortions. in content and referred to a skill orientation toward gambling.8) than did nonparticipants (M ⫽ 3. 1996.04.95. or to the GBQ test–retest coefficient of . p ⫽ .38 (SD ⫽ 9. Participants who had completed Study 2 were telephoned study. MEYERS. sented to participate completed the GBQ 2 weeks after the initial admin- The eight items loading on the second factor were homogeneous istration date. range: 13–76). Scores on the Illusion of Control factor could range from 8 to 56. It is noteworthy that these data represent the first empirical constructs of irrational beliefs are similar to gambling-related support for the differentiation between illusion of control and other cognitive distortions.. gamblers and (b) whether higher GBQ scores were correlated with . the mean Luck/Perseverance factor score was 30.6). Of particular interest is the potentially causal role that determined. The retest This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. it is not clear that findings based on more general structs.89 for a of Control factor had a slightly lower coefficient alpha. correlation for full-scale scores was . Participants reported gambling more frequently over the and that therefore there is no need to tell loved ones of temporary past 12 months (M ⫽ 7. Ladouceur Participants. It may be that respondents who between this subsample and the undergraduates who chose not to partici- endorse this view believe that they will eventually win back money pate in this study. These participants were 60.4 hr gambling per occasion. respectively.21 (SD ⫽ and invited to complete an additional set of questions.2 occasions and spent 2.

004. p ⬍ . tically significant differences at p ⬍ . chusetts Gambling Screen DSM–IV (MAGS DSM–IV. p ⬍ . Lesieur & Blume.21).94.003. Lesieur & Blume. 1994). Method SD ⫽ 18.53. American Psychiatric Associa. d ⫽ . (1996).06. 1980) criteria for pathological gambling. including participants’ age. d ⫽ . Shaffer. p ⫽ . The participants (N ⫽ 371) were undergraduate students gamblers (M ⫽53. Likewise. t(368) ⫽ 4.09. Pearson correlation coefficients revealed moderate relation- ships between the average amounts of time spent per gambling Procedure occasion and GBQ full-scale and Luck/Perseverance factor scores (rs ⫽ .34) scored significantly higher than nonpathological Participants. The results of this second contrast Manual for Mental Disorders (DSM–III. 368) ⫽ 11. These analyses suggested trends for Scanlan. p ⫽ .001.. and (b) 5 and higher to classify nonproblem (n ⫽ 317). and pathological gamblers (M ⫽ 70. 368) ⫽ 7. 1997). than nonproblem gamblers on the Luck/Perseverance. There is provided an index of lifetime gambling problems. t(53) ⫽ 1.50. F(1. respectively.05 between these groups . SD ⫽ 20.21. Shaffer et al. Pathological gamblers also scored significantly higher 2. because they did not complete all of the items on lusion of Control factors. An ANOVA also revealed significant dif- retest reliability (rs ⫽ .001. Lesieur & Blume. respectively. 1987).36. course of their lifetime scored higher on the GBQ and its factors than those Consistent with the theory proposed by Ladouceur and Walker without problems.05). p ⬎ . 1994). and Derevensky (1997) reported that. there was no items pertain to gambling problems over the individuals’ lifetime and are significant difference between problem and pathological gamblers. d ⫽ . Follow-up contrasts indicated that This document is copyrighted by the American Psychological Association or one of its allied publishers. t(368) ⫽ order (r ⫽ . GAMBLERS’ BELIEFS QUESTIONNAIRE 147 longer gambling sessions among problem and pathological Results and Discussion gamblers. of SOGS scores with counselors’ assessments of gambling involvement and F(2. should be interpreted with caution because of the limited sample tion.22. (b) 3 or 4.001) and with classification based on DSM–III–R criteria indicated that problem and pathological gamblers scored higher (American Psychiatric Association. and both measures of problem gambling. SD ⫽ 22. However. participants identified by the SOGS as gambling in prob- ticipants’ past-year gambling behavior. based on criteria from the third edition of the Diagnostic and Statistical t(53) ⫽ 1.43 and . p ⫽ . & Cummings.. and commonly used cutoff scores of 3 or 4 and 5 or 4.94.40). Convergent ferences among SOGS-identified groups on Luck/Perseverance validity for the SOGS has been demonstrated by analyzing the relationship and Illusion of Control factor scores.63. The Massa.93. 368) ⫽ 12.82). p ⫽ . participants’ gam. 369) ⫽ 8. p ⬍ . d ⫽ . respectively). F(1. Second.53. In addition.001. p ⬍ . 1994) is a 12-item self-report measure based on differences. Hardoon. SOGS GBQ. pathological gamblers made more irrational identified groups differed on GBQ full-scale or factor scores. respectively (Volberg.48.001. t(368) ⫽ range from 0 to 20.42 (Cohen. d ⫽ . No differences between problem and greater have been used to identify problem and pathological gamblers. Scores on the SOGS 3.84. which resulted in a power coefficient to be internally consistent (␣ ⫽ . relative to nonpatho- We conducted a second set of ANOVAs to determine whether SOGS. SD ⫽ 22. individ- score of 5 (Shaffer et al. First. We used scores of (a) prior empirical support for these findings. scores for the 52 of the 54 problem and pathological gamblers who 1994). A score of 5 or greater has been used to identify pathological provided an estimate of the average duration of a gambling occa- gamblers (Shaffer et al. Follow-up contrasts (r ⫽ –. problem (n ⫽ 22). sion. 369) ⫽ 8.68.61–1. p ⫽ . t(53) –1.50) revealed significant differences among these three groups. That no statis- problem and pathological gamblers. provided basic demo.0. average duration of gambling sessions. Thirty-two (18%) of the initial 403 participants in Study 2 were not than nonpathological gamblers on the Luck/Perseverance and Il- included in this study. Gambling involvement questionnaire. however. marital 60.001. identified nonproblem (M ⫽ 52.53.50. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 1987) for pathological gambling dis.32. p ⫽ . p ⫽ . 1988).001. SD ⫽ status. had higher classified participants on the basis of their scores on the SOGS.97.48. race. used correlational analyses to explore the relationship between length of There was a trend for SOGS-defined pathological gamblers to gambling session and GBQ full and factor scores among SOGS-identified score higher than problem gamblers on the GBQ. logical gamblers. d ⫽ . (n ⫽ 192) and community members (n ⫽ 179) who had completed Study d ⫽ . 369) ⫽ 10. bling frequency. We then used ANOVAs to determine uals who had gambled at clinically significant pathological levels whether groups differed on GBQ full-scale or factor scores. F(1. we used MAGS DSM–IV scores to classify par. or Illusion of Control factors. p ⫽ . d ⫽ .55.48. This measure has shown time spent per gambling occasion and GBQ full-scale and factor adequate internal consistency for research purposes (␣ ⫽ . 1987).69.90.39. We identified 11 pathological lematic ways over the course of their lifetimes scored significantly gamblers and 360 nonpathological gamblers using a recommended cutoff higher than nonproblem gamblers on the GBQ.. which scores on the GBQ than nonpathological gamblers did. Respondents are asked to indicate the presence or absence of We examined the relationship between the average amount of diagnostic symptoms over the past 12 months. and Illusion of Control factors. 1987) is a 20-item self-report measure that was developed as a screening instrument for pathological gambling. SD ⫽ 18. Gupta.001. ance. The SOGS has been reported sizes and medium effect size.97) and possess adequate 1-month test– of . and wagering F(2.61) South Oaks Gambling Screen. Finally.72. A short gambling involvement An ANOVA of the GBQ full-scale scores among SOGS- questionnaire. No significant relation- We conducted two sets of analyses of variance (ANOVAs) to examine ship between average time spent per gambling occasion and Illu- whether individuals with gambling problems in the past year or over the sion of Control factor scores was detected (r ⫽ . as indicated by MAGS DSM–IV scores. gender.09. F(2. LaBrie. problem and pathological gamblers (M ⫽ 66. which was designed for this study.87.19. d ⫽ . The ANOVA of the GBQ full-scale scores revealed that patho- logical gamblers identified with the MAGS DSM–IV (M ⫽ 75.16.002. their relatively small sample sizes and me- DSM–IV (American Psychiatric Association.73.03. Baboushkin. p ⬍ . levels were assessed. 1994) criteria for pathological dium effects limited the power of follow-up analyses. we in the past year. The South Oaks Gambling Screen scored significantly higher than nonproblem gamblers on the (SOGS. and pathological gamblers (n ⫽ 32). gambling. education level. 21. respectively.86. problem (M ⫽ graphic information. p ⬍ .001.90.55. we statements while gambling. pathological gamblers were detected on either the Luck/Persever- Massachusetts Gambling Screen DSM–IV Questionnaire. p ⬍ . and family income. less than 3.

88. it may be that the GBQ does not assess other hypothesized role that cognitive distortions may play in problem- important domains of irrational beliefs that might potentially dif. 1988).05) or Illusion of Control factors (r ⫽ –. AND WHELAN were found should be viewed tentatively because of the limited scores were not correlated with the Luck/Perseverance (r ⫽ –. & Sukhodolsky.g. gambling-related cognitive distortions. the properties of the GBQ. there may truly be no difference between problem and pathological gamblers in their gambling-related irrational be. This finding. 1998). slot machines. 1964). recent efforts to measure illusion of control across Study 5 was designed to explore the discriminant validity of the specific gambling activities have shown limited success (Kassi- GBQ. Similarly. SDS cognitive distortions. data from more representative and were included in this study. age). It may be that beliefs about cognitive distortions than those without gambling problems. is consistent with that of Ladouceur and Walker (1996). 148 STEENBERGH. Alternatively. statistical power available for this analysis. measure of social desirability. p ⬎ . gamblers in both groups. horse racing).05). Full-scale and factor scores exhibited good internal consistency. Method Future studies are necessary to examine the properties of the Participants who were involved in Study 2 were asked to complete a GBQ across various demographic characteristics (e. tween problem and pathological gamblers may not be due to Initial scale construction was based on existing conceptualiza- endorsement of irrational beliefs. Given the growing concern over problem gambling and the liefs. The GBQ needs further examination. lem and pathological gamblers spent gambling. blackjack. horse track betters. GBQ full-scale directly challenged over the course of the gambling session and. 393 (91. . Fur- one’s ability to control the outcome of a gambling event are thermore. By selecting scale items that were GBQ total and factor scores. Another possible limita- ness” of irrational belief items. which would have skewed the results of this analysis (Abbott & Volberg. slot players) should item is presented in the form of a statement.88) and temporal stability (1 This series of studies provides initial support for the use of the month test–retest r ⫽ . Third. partici- Control factor scores did not correlate with the average duration of pants with gambling-related problems reported higher levels of gambling sessions was surprising. 1996). MAY. indicate whether each statement is true or false as it pertains to them. The final 21-item GBQ measured two related con- scores were correlated with the average length of time that prob.1%) race. and respondents are asked to be examined to ensure the accuracy and utility of the measure.. As a treatment planning and outcome mea- sure.03.05).g. we conducted this series of studies to This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Davidson. and Luck/Perseverance factor scores were moderately correlated with mounting evidence against these beliefs. it is possible that we overlooked new and potentially important domains of the construct. Given the concern of one expert rater about the “obvious. The limited sample of p ⬎ . it may aid clinicians in conducting cognitively based treat- Results and Discussion ments for problem gambling as well as provide important infor- A Pearson correlation coefficient revealed no significant rela. too. either. Finally. Each tions (e. theoretically based. None of the items address beliefs specific to certain Overview types of gambling (e. The SDS has good internal reliability (K-R 20 ⫽ . which is a 33-item self-report measure of an including its predictive utility. MEYERS... That Illusion of and Walker’s (1996) cognitive perspective on gambling. GBQ for clinical assessment. some participants. tions of gambling-related cognitive distortions and input from This study provides additional convergent validity for the GBQ expert raters. GBQ in both research and clinical situations. mation on the impact of such treatments on gambling-related tionship between the SDS and the GBQ (r ⫽ –. true differences be. roulette players. related cognitive distortions could prove quite useful to clinicians and researchers.03. nove. However.02. individuals may quit with the average length of their gambling sessions. coupled with a moderate effect size. who Evidence of convergent validity was provided by two separate hypothesized that cognitive distortions among individuals with analyses that compared GBQ full-scale and factor scores across gambling problems may be responsible for the maintenance of nonproblem and problem gamblers. However. clinical samples of gamblers will also aid in the validation of the These participants completed the Social Desirability Scale (SDS. it is clear that a measure of gambling- ferentiate between problem and pathological gamblers. Crowne & Marlowe. Therefore. among those with gambling problems.g. p ⬎ . divergent validity revealed that GBQ scores were not related to social desirability biases. (Cohen. This finding clearly warrants further exploration. gender. These findings suggest that full-scale and factor scores were not resulted in a 58% chance that a significant difference between affected by participants’ tendencies to respond in a socially desir- groups would have been missed. Given the largely nonclinical nature of the sample used completed every item on both the GBQ and the social desirability measure in the present series of studies. One study of gambling earlier. we deemed it important to examine tion of the GBQ involved the construct-driven approach that was the relationship between social desirability response biases and used to develop the instrument. Consistent with Ladouceur gambling behavior once it has been initiated. if these findings are accurate and the GBQ does not differentiate between problem and pathological gamblers. First. Of those 403 participants. Crowne & Marlowe. We used factor and internal consistency analyses to by showing that GBQ full-scale and Luck/Perseverance factor select items. 1964). if in fact such a difference existed able way. One possible limitation of the GBQ may be the general nature of Study 5 the items. SOGS measurement error may have led to misclassification of develop and evaluate a self-report assessment tool for evaluating This document is copyrighted by the American Psychological Association or one of its allied publishers. structs: belief in luck/perseverance and the illusion of control. atic gambling behavior. then there are at least four possible explanations for such General Discussion findings. within specific gambling popula- individual’s tendency to endorse items in a socially desirable manner.

. J. (1998. Quebec. J. (1987). Washington. June). Cognitive psychopathology of problem gambling.. (1987). I. New York: Wiley. 34. (1996). 47–52.. A. & Dumont. Development and validation of a Underlying cognitions in gambling behavior amongst university stu. 142–163). ual of mental disorders (3rd ed. Journal of Psychology.. San Francisco. J. sented at the 106th Annual Convention of the American Psychological Association. R. R. 32. Zwick. S. 179 –185. Addictive Behaviors. Mayrand. Scanlan. J.. J. (1986). Langer. blers and gambling behavior (pp. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. & Tremblay. Illusion of control: Effect of subject’s participation and impli- Abbott. (1980). Eadington (Ed. GAMBLERS’ BELIEFS QUESTIONNAIRE 149 References Ladouceur. 2001 18. R. Comparison of five rules for Kassinove. Social Henslin. England: Horn. N. Montreal. International Conference on Gambling and Risk Taking: Vol. Revision received June 22. Irrational thinking and gambling. Salkoskvis (Ed. 727–732. M. report to the Mississippi Council on Compulsive Gambling. (1975). LaBrie. R. The New Zealand National cation. H. S. M. 1184 –1188. Psychological Bulletin.. 143–160. R. 275–282. Sylvain. Ladouceur. A.. The South Oaks Gambling Screen American Psychiatric Association.. Washington. Davidson. R. (pp. Journal of Personality and American Psychiatric Association. analysis. & Cummings. (1992). 1593–1604. Cognitive– behavioral Received February 16.. Statistical power analysis for the behavioral sciences behavioral treatment of pathological gambling: A controlled study.. A. C. Science Research Center. Mississippi State University. T. Journal of Behavior Screen (MAGS). Baboushkin. 432– 442. (1994). H. 2001 䡲 . Treatment of Pathological gambling among adolescents: Massachusetts Gambling This document is copyrighted by the American Psychological Association or one of its allied publishers. American Journal of Psychiatry. evaluative dependence.. D. Jour- (2nd Ed. & Velicer. Journal of Gambling Ladouceur. M. Diagnostic and statistical man. ogy. Sylvain. Lesieur. Oxford. Survey of Problem and Pathological Gambling. (1965). pathological gamblers: An experimental study.. NJ: Erlbaum.. R... Gupta. (SOGS): A new instrument for the identification of pathological gam- ual of mental disorders (4th ed. A cognitive perspective on gambling. K. Ladouceur. G. The approval motive: Studies in Toneatto. New York: Wiley. Craps and magic. nal of Consulting and Clinical Psychology. 316 –330. & Sukhodolsky. D. Diagnostic and statistical man. Boisvert. Accepted July 2. R. Ladouceur. (1994).. & Ladouceur.. Crowne. Bujold. DC: Author. 12. 10. (1996). 3.). 339 –362. 2001 treatment for adolescent pathological gamblers.. (1964). C. Poster pre.. L. M. 99. K. Gambling research: Proceedings of The Seventh cognitive behavior therapy: A case report. Social Psychology. A.. Reno: University of Nevada Volberg. American Journal of Sociology.. N.). 144. Pathological gambling treated with W. Gambling and problem gambling in Mississippi: A Press. Behavior Modification. (1997. Letarte. W. Hillsdale. J. J. F. Dussault. J.). J. H. rev. Illusions of control and self-reported gambling experiences. J. (1986). August). Studies. (1997). (1999). Trends in cognitive and behavioural therapies American Psychiatric Association. 860 – 862. R. M. R. Shaffer. & Boisvert. Malouff. 73. Therapy and Experimental Psychiatry. (1997). & Blume. (1994). B. 54. & Walker. J. & Marlowe. R. R. M. T. 117. & Schutte. (1990).). blers. determining the number of components to retain. Hardoon. M. 230 –242. Canada.).. J. T. C. 65. M. E. measure of irrational belief. Paper presented at the Tenth International Conference on Gam.. P. B.. 89 –120).. bling and Risk Taking. (1988). J. DC: Author. A. Walker. L. (1994). 311–328. (1984). DC: Author. 15. In P.). Diagnostic and statistical man. The psychology of gambling. D. Journal of Gambling Studies. ual of mental disorders (3rd ed. A rationale and test for the number of factors in factor Pergamon Press. Volberg. The illusion of control. (1988). & Boisvert. (1967).. R. Psychometrika. Washington. M. In Toneatto.. 25. Cognitive and Cohen. R.. Gam. I. W. Journal of Consulting and Clinical Psychol- dents. W. & Sobell. Gaboury. Substance Use & Misuse. 497–501. 30. & Derevensky. &.