Maria Anna Donati ,

Francesca Chiesi , & Caterina Primi

U N I V E R S I T Y O F F L O R E N C E

I T A L Y

Gambling in youths:

A new scale to measure

pathological gambling behavior

Gambling in adolescence

• A large number of adolescents are involved in

gambling activities (e.g., Scholes-Balog et al., 2014).

• Between 60 and 99% of young people aged 12–20

years gambled over the past year (Splevins et al., 2010).

• Despite boys gambled more than girls, female

adolescent gamblers are increasing (Bastiani et al., 2010).

Youth gambling problems

• A high proportion of young people gamble

excessively, developing a wide array of

psychological, social, and economic problems (e.g.,

see Ariyabuddhiphongs, 2013).

• Problem gambling among adolescents is four or five times

higher than among adults (e.g., Olason & Gretarsson, 2009).

• Gambling as a central public health issue of

prevention research and practice (see Blinn-Pike et al.,

2010).

The measurement

• Much attention has been paid to the issue of

measurement (for reviews, see Derevensky & Gupta, 2004; 2006;

Stinchfield, 2010).

• The 3 most commonly used instruments are:

• South Oaks Gambling Screen-Revised for Adolescents (SOGS-RA) (Winters et al., 1993);

• DSM-IV-Juvenile (DSM-IV-J) and its revision, DSM-IV-

Multiple Response-Juvenile (DSM- IV-MR-J) (Fisher, 1992; 2000);

• Massachusetts Adolescent Gambling Screen (MAGS)

(Shaffer et al., 1994).

Limitations in the measurement

• Adaptation from adult instruments, with minor changes

in the wording of some items (Chiesi et al., 2013);

• Dichotomous response format (or it becomes

dichotomous in the scoring procedure), while a Likert-

type format offers more advantages psychometrically

(Muňiz et al., 2005);

• No evidence for equivalence among male/female

(Derevensky & Gupta, 2004) and younger/older

respondents;

• No screens based on the new DSM-5.

Gambling Disorder in DSM-5

From DSM-IV to DSM-5

From Pathological Gambling to Disordered Gambling;

From Impulse-Control Disorders not elsewhere classified to

Substance-Related and Addictive Disorders;

From ten to nine criteria: One criterion (to commit illegal

acts) has been eliminated;

From five to four criteria for diagnosis: lower threshold;

Different forms of the disorder are specified (mild, moderate,

severe);

Specified time period: symptoms must be present during a

12 month time period.

Aim

• To develop a new brief scale to assess

pathological gambling among adolescents

(Gambling Behavior Scale for Adolescents, GBSA)

referring to the DSM-5:

o Adapting the criteria to adolescent lives

o Using a multiple-response format

o Applying Item Response Theory (IRT) • Testing the equivalence for male/female and younger/older respondents

o Providing validity evidence.

Item Response Theory (IRT)

IRT aims to show the probabilistic relationship

between pattern responses, latent characteristics of

respondents, and item properties.

Advantages of IRT: Fit of the measurement model

Parameter estimates of items that define their ability to

discriminate and to measure different levels of the latent trait

(θ)

Precision of the test at different levels of θ (Reliability)

Differential Item Functioning (DIF) across different groups.

Method

Participants

o 1723 adolescents attending middle and high school

(56% males, mean age=15.64 years, SD=1.79, 11-23

years):

• 265 (15%) middle school (II year);

• 1458 (85%) high school (23% I year; 30% II year; 19% III

year; 28% IV year).

Method

Measures • Gambling Behavior Scale for Adolescents (GBSA): 33

items - referred to the last 12 months - with a 5 point scale (1=never, 2 = rarely,

3 = sometimes, 4 = often, 5 = very often).

o This version was

obtained starting from

a preliminary set of 45

items tested in pilot

studies at both

qualitative and

quantitative levels.

DSM-5

Criteria

Step 0:

45 items

Step 1:

33 items

Tolerance 5 3

Withdrawal 5 4

Loss of Control 5 4

Preoccupation 5 3

Escape 5 3

Chasing 5 4

Lying 5 4

Risk/lose relationships/opportunities

5 5

Bailout 5 3

Method

Measures

• Questionnaire on gambling behavior in the last 12

months (activities and relative frequency);

• South Oaks Gambling Screen – Revised for

Adolescents (SOGS-RA) (Winters et al., 1993; Italian version:

Colasante et al., 2013)

• Illusion of Control (GRCS-IC) (Raylu & Oei, 2004; Italian

version: Donati et al., 2014)

• Gambling Attitude Scale (GAS) (Delfabbro & Thrupp,

2003; Italian version: Primi et al., 2013)

Results – Preliminary Analysis

o Analyses were conducted with adolescents who

gambled

• N=1187 (69%) (59% males, mean age=15.66 years,

SD=1.71).

o Since item response distributions were strongly skewed,

we converted the 5-point scale into a 3-point scale

(never = 1, rarely/sometimes = 2, often/very often = 3).

o The 33 item-version resulted to be unidimensional.

Results – Item selection and calibration

• The fit of the unidimensional

two parameter model was

tested: • Seven items were eliminated since they

did not fit to the model.

• We performed the IRT

calibration with the remaining

26 items: • For each criterion we selected the

best functioning items relying on IRT

discrimination parameters.

DSM-5

Criteria

Step 1:

33

items

Tolerance 3

Withdrawal 4

Loss of

Control 4

Preoccupation 3

Escape 3

Chasing 4

Lying 4

Risk/lose relationships/opportun

ities

5

Bailout 3

DSM-5

Criteria

Step 1:

33

items

Step 2:

26

items

Tolerance 3 3

Withdrawal 4 4

Loss of

Control 4 4

Preoccupation 3 1

Escape 3 2

Chasing 4 1

Lying 4 4

Risk/lose relationships/opportun

ities

5 5

Bailout 3 2

DSM-5

Criteria

Step 1:

33

items

Step 2:

26

items

Step 3:

9

items

Tolerance 3 3 1

Withdrawal 4 4 1

Loss of

Control 4 4 1

Preoccupation 3 1 1

Escape 3 2 1

Chasing 4 1 1

Lying 4 4 1

Risk/lose relationships/opportun

ities

5 5 1

Bailout 3 2 1

Item DSM-5 a b1 b2

1 Tolerance .67 3.50 (0.34) 1.66 (0.10) 2.83 (0.22)

2 Withdrawal .69 2.98 (0.29) 1.55 (0.09) 2.35 (0.18)

3 Loss of Control .63 3.02 (0.38) 1.72 (0.11) 2.52 (0.18)

4 Preoccupation .52 1.89 (0.19) 1.35 (0.10) 2.21 (0.17)

5 Escape .68 2.29 (0.25) 1.62 (0.11) 3.02 (0.24)

6 Chasing .53 1.66 (0.16) 0.82 (0.07) 2.80 (0.21)

7 Lying .58 2.53 (0.30) 1.56 (0.10) 2.56 (0.19)

8 Risk/lose relationships/

opportunities

.68 3.19 (0.38) 1.65 (0.10) 2.65 (0.17)

9 Bailout .70 1.84 (0.19) 1.52 (0.11) 2.87 (0.24)

Factor loadings and item parameters

Test Information Function (TIF)

o The TIF shows the precision of the test at

different levels of the latent trait

M(trait)

Differential Item Functioning (DIF) across gender and age

An item is considered to exhibit DIF if respondents

of two different groups who have equal levels of the

trait do not have the same probability of responding

affirmatively to that item.

Comparing male and female gamblers and younger

and older gamblers (based on the median age of

around 16 years old), no items exhibited DIF with

respect to:

Discrimination parameters (a)

Severity parameters (b)

Validity

GBSA-total

score

Gambling Frequency .44***

Gambling Versatility .33***

Illusion of Control .48***

Perception of gambling profitability .30***

Problem gambling .55***

*** p<.001

Discussion

• Referring to the DSM-5 definition of GD and

using the IRT approach, this scale seems to be a

reliable and valid brief tool to accurately

measure medium and high levels of gambling

problems among adolescents.

It could be especially useful to identify severe levels

of gambling problems in adolescents, regardless of

their gender and age.

Conclusion

• The GBSA is likely to be a useful screening

and diagnostic instrument for gambling

research and practice with adolescents.

• Future research is needed to:

• Test its psychometric properties with

adolescents who are in treatment for their

gambling problems (clinical samples).

• Develop new IRT-based cut-offs.

• Conduct cross-cultural studies.

References

Ariyabuddhiphongs, V. (2013). Problem gambling prevention: before, during, and after measures.

International Journal of Mental Health and Addiction, 11(5), 568-582.

Bastiani, L., Curzio, O., Gori, M., Colasante, E., Siciliano, V., Panini, R., et al. (2010). L’Italia che gioca

[Gambling population in Italy]. Retrieved 21.07.11, from.

http://www.andinrete.it/portale/documenti/pdf/IFC-CNR%20giovani%202009.pdf.

Blinn-Pike, L.,Worthy, S. L., & Jonkman, J. N. (2010). Adolescent gambling: a review of an emerging

field of research. Journal of Adolescent Health, 47, 223–236.

http://dx.doi.org/10.1016/j.jadohealth.2010.05.003.

Chiesi, F, Donati, M. A., Galli, S., & Primi, C. (2013). The Suitability of the SOGS-RA as Screening Tool:

Item Response Theory-Based evidence. Psychology of Addictive Behaviors, 27(1), 287-293. doi:

10.1037/a0029987.

Colasante, E., Gori, M., Bastiani, L., Scalese, M., Siciliano, V., & Molinaro, S. (2013). Italian adolescent

gambling behaviour: Psychometric evaluation of the South Oaks gambling screen—Revised for

adolescents (SOGS-RA) among a sample of Italian students. Journal of Gambling Studies,. doi:10.

1007/s10899-013-9385-6.

Delfabbro, P., & Thrupp, L. (2003). The social determinants of youth gambling in South Australian

adolescents. Journal of Adolescence, 26, 313–330. doi:10.1016/S0140-1971(03)00013-7.

Derevensky, J. L., & Gupta, R. (2004). Adolescents with gambling problems: A synopsis of our current

knowledge. eGambling. The Electronic Journal of Gambling Issues, 10, 1–22. Retrieved from

http://www.camh.net/egambling/archive/pdf/EJGI-issue10/EJGI-Issue10-derevenskygupta.pdf

References

Derevensky, J. L., & Gupta, R. (2006). Measuring gambling problems among adolescents: Current

status and future directions. International Gambling Studies, 6, 201–215.

Donati, M. A.., Ancona, F., Chiesi, F., & Primi, C. (in preparation). Psychometric properties of the

Gambling Related Cognitions Scale (GRCS) in young Italian gamblers.

Fisher, S. (1992). Measuring pathological gambling in children: The case of fruit machines in the U.K.

Journal of Gambling Studies, 8, 263–285.

Fisher, S. (2000). Developing the DSM-IV-MR-J criteria to identify adolescent problem gambling in

non-clinical populations. Journal of Gambling Studies, 16, 253–273.

Muñiz, J., Garcı́a-Cueto, E., & Lozano, L. M. (2005). Item format and the psychometric properties of

the Eysenck Personality Questionnaire. Personality and individual differences, 38(1), 61-69.

Olason, D. T., & Gretarsson, S. J. (2009). Iceland. In Problem Gambling in Europe (pp. 137-151). Springer

New York.

Primi, C., Donati, M. A., Bellini, I., Busdraghi, C., & Chiesi, F. (2013). Measuring the attitude towards

the profitability of gambling: The psychometrics properties of the Gambling Attitude Scale. Bollettino

di Psicologia Applicata, 266, 49-58.

Raylu, N., & Oei, T. P. (2004). The Gambling Related Cognitions Scale (GRCS): development,

confirmatory factor validation and psychometric properties. Addiction, 99, 757-769.

Shaffer, H. J., LaBrie, R. A., Scanlan, K. M., & Cummings, T. H. (1994). Pathological gambling among

adolescents: Massachusetts gambling screen (MAGS). Journal of Gambling Studies, 10, 339–362.

References

Scholes-Balog, K. E., Hemphill, S. A., Dowling, N. A., & Toumbourou, J. W. (2014). A prospective study

of adolescent risk and protective factors for problem gambling among young adults. Journal of

Adolescence, 37(2), 215–224. doi:10.1016/j.adolescence.2013.12.006

Splevins, K., Mireskandari, S., Clayton, K., & Blaszczynski, A. (2010). Prevalence of adolescent

problem gambling, related harms and help-seeking behaviours among an Australian population.

Journal of Gambling Studies, 26, 189–204. http://dx.doi.org/10.1007/s10899-009-9169-1.

Stinchfield, R. (2010). A critical review of adolescent problem gambling assessment instruments.

International Journal of Adolescent Medicine and Health, 22(1), 77-93.

Winters, K. C., Stinchfield, R. D., & Fulkerson, J. (1993). Toward the development of an adolescent

gambling problem severity scale. Journal of Gambling Studies, 9, 63–84.

Acknowledgments

mariaanna.donati@unifi.it