The Structure of Gambling Behavior in Adulthood

1
The Structure of Gambling Behavior in Adulthood

• Bethany Cara Bray
• The Methodology Center, The Prevention Research
  Center
• Department of Human Development and Family
  Studies
• The Pennsylvania State University
• Society for Prevention Research
• Annual Meeting
• Thursday, June 1, 2006

2
Acknowledgements

• Co-author Dr. Linda M. Collins
• Funding
• Center for Prevention and Treatment Methodology
  (NIDA) P50-DA-10075
• Prevention and Methodology Training (PAMT)
  Program (NIDA) T32-DA-017629
• Special thanks S. T. Lanza, M. M.
  Maldonado-Molina, T. L. Root, K. J. Auerbach, J.
  L. Schafer

3
Outline

• The Idea
• My Motivation
• My Methods
• Current Results
• Discussion

4
Outline

• The Idea
• My Motivation
• My Methods
• Current Results
• Discussion

5
The Idea

• Are there different types of gamblers for whom
  different targeted prevention programs should be
  designed?
• Important to move beyond the typical
  classification of gamblers and gambling behavior
  
• Non-problem gamblers
• Problem or At-risk gamblers
• Pathological gamblers

6
Research Questions

• Are there identifiable types of gamblers?
• If so, are these types different for men and
  women?
• Is latent class analysis (LCA) a more useful
  method than traditional approaches when
  classifying gambling behavior?
• What can LCA tell us about the performance of the
  diagnostic criteria when they are used to
  identify individuals with non-clinical levels of
  problem gambling for targeted prevention programs?

7
Outline

• The Idea
• My Motivation
• My Methods
• Current Results
• Discussion

8
Issues in the Conceptualization of PG

• Continuous vs. Categorical
• CATEGORICAL
• Manifest vs. Latent
• LATENT

9
Operationalizations of PG

• DSM-III-R
• South Oaks Gambling Screen
• DSM-IV
• Lie/Bet Screen
• General DSM-IV Screens
• Canadian Problem Gambling Index

10
Operationalization of PG

• the unidimensional additive scoring of
  screening instruments is inadequate to represent
  a multidimensional latent state. The method of
  summing endorsed characteristics assumes that all
  dimensions exist on the same additive continuum
  and that all dimensions equally predict gambling
  disorders This equivalence is highly unlikely
  and misleading.
• (Shaffer et al., 2004)

11
The Prevention of PG

• Universal Programs
• Targeted/Indicated Programs
• Screening typically uses diagnostic criteria

12
Outline

• The Idea
• My Motivation
• My Methods
• Current Results
• Discussion

13
The NESARC

• 2001-2002 National Epidemiologic Survey on
  Alcohol and Related Conditions
• Sponsored by the National Institute on Alcohol
  Abuse and Alcoholism (NIAAA)
• A source for information and data on the U.S.
  population for
• Alcohol and drug use
• Alcohol and drug abuse and dependence
• Associated psychiatric and other medical
  comorbidities
• Representative sample of the U.S. population,
  aged 18 years and older
• N 43,093

14
Participants

• Screening question
• Includes all participants who had ever gambled
  five or more times in any one year
• N 11,153
• Males n 6,000
• Females n 5,153

15
Measures

• Questionnaire
• 15 questions operationalize 10 diagnostic
  criteria for lifetime pathological gambling
• Lifetime pathological gambling indicators
• 1 indicator created for each of the 10 diagnostic
  criteria

16
Measures

• DSM-IV Diagnostic Criteria
• Preoccupation
• Tolerance
• Loss of control
• Withdrawal
• Escape
• Chasing
• Lying
• Illegal acts
• Risking significant relationship
• Bailout

17
Measures

• DSM-IV Diagnostic Criteria
• Preoccupation with gambling
• Needing to gamble with increasing amounts of
  money
• Being unsuccessful at controlling/stopping
  gambling
• Being restless/irritable when controlling
  gambling
• Gambling to escape problems or a dysphoric mood
• Returning another day to get even (chasing)
• Lying to conceal extent of gambling involvement
• Committing illegal acts to finance gambling
• Risking significant relationship/job/opportunity
• Relying on others to relieve a financial situation

18
Traditional Analyses

• Classify participants based on the number of
  diagnostic criteria endorsed
• 0 2 criteria Non-problem gamblers
• 3 4 criteria Problem gamblers
• 5 criteria Pathological gamblers
• Examine proportion of participants endorsing each
  individual criterion

19
Latent Class Analysis

• Statistical method
• Identifies exclusive groups of individuals
• Groups characterized by similar patterns of
  behavior
• Models underlying group structure of a single,
  static, categorical latent (unobserved) variable
• Uses categorical indicators of behavior

20
LCA Parameters

• ? Gamma marginal probability of latent class
  membership
• Probability of membership in the non-problem
  gambler latent class
• ? Rho measurement parameter describes how
  individuals within a latent class response to
  indicators
• Probability of endorsing the preoccupation
  indicator, conditional on latent class membership

21
Models

• Different numbers of latent classes
• 2, 3, 4, 5 class models
• Bayesian Information Criterion (BIC) used to
  select the most well-fitting model
• Examined three groups of participants
• All participants
• Males
• Females

22
Outline

• The Idea
• My Motivation
• My Methods
• Current Results
• Discussion

23
Traditional Results
24
Traditional Results
25
Latent Class Analysis Results
26
Latent Class Analysis Results
27
Latent Class Analysis Results
28
Latent Class Analysis Results
?
29
Outline

• The Idea
• My Motivation
• My Methods
• Current Results
• Discussion

30
Question 1

• Are there identifiable types of gamblers? If so,
  are these types different for men and women?

31
Question 1

• 3-class model best describes the gambling
  behavior of participants 3 identifiable types
  of gamblers
• Non-problem gamblers
• No diagnostic criteria endorsed
• Preoccupied gamblers
• Moderate endorsement of being preoccupied with
  gambling
• Pathological gamblers
• Endorse
• Being preoccupied with gambling
• Needing to gamble with increasing amounts of
  money
• Not being able to control/cut back/stop gambling
• Resorting to chasing behavior to win back
  losses
• Lying to others to conceal extent of gambling
  involvement

32
Question 2

• Is latent class analysis (LCA) a more useful
  method than traditional approaches when
  classifying gambling behavior?

33
Question 2

• An alternative approach to classifying
  individuals simply based on the total number of
  diagnostic criteria met
• LCA can help identify types of gambles with
  similar patterns of behavior
• May be helpful when designing targeted prevention
  programs

34
Question 3

• What can LCA tell us about the performance of the
  diagnostic criteria when they are used to
  identify individuals with non-clinical levels of
  problem gambling for targeted prevention
  programs?

35
Question 3

• ? parameters for latent class 2 (preoccupied
  gamblers) suggest a lot of heterogeneity
• Suggests possible need for other types of
  indicators or criteria in order to understand
  behavior at this level of development

36
Our Themes

• Gambling as a Public Health Concern
• Etiology of Gambling and Problem Gambling
• Implications for Problem Gambling Prevention

37
Where Do We Go From Here?

• Further investigation of the categorical latent
  structure of gambling behavior
• Posterior predictive check distribution for model
  selection
• Power of hypothesis tests
• Include other indicators of gambling behavior
  that move beyond the diagnostic criteria
• Include other important grouping variables
• Race / ethnicity, Age, Income, Religion

38
Where Do We Go From Here?

• Include predictors of latent class membership
• Alcohol use
• Other substance use
• Psychiatric and psychological disorders
• Depression
• Anxiety
• Extend longitudinally to address change in latent
  class membership
• LTA, ALTA

39
References

• NESARC Website http//niaaa.census.gov
• Shaffer, H. J., LaBrie, R. A., LaPlante, D. A.,
  Nelson, S. E., and Stanton, M. V. (2004). The
  road less traveled Moving from distribution to
  determinants in the study of gambling
  epidemiology. Canadian Journal of Psychiatry,
  49, 8, 504-516.