Assessing the Construct Validity of the Gambling Functional Assessment
Attempts to understand the rationale for why people gamble when the odds of winning are against them have included a myriad of self-report assessments. The Gambling Functional Assessment (GFA) is a twenty item questionnaire assessing four maintaining variables of gambling behaviors: social attention, psychological/physical escape, access to tangible rewards, and sensory stimulation. To date, validity of the GFA has not been assessed with a clinical gambling population. The purpose of the current study was to assess the clinical validity of the GFA as a measure of functional relations between environmental events and gambling behaviors.
Method:
One hundred and seventy five participants recruited across the US completed a demographics questionnaire, were screened for potential gambling pathology using the South Oaks Gambling Screen (SOGS), and completed the GFA. The order of the surveys was counterbalanced across participants to control for sequence effects. Gamblers (N=175; 28% female) who were over 18 and scored above 3 on the SOGS participated in this study.
Results:
All GFA responses were subjected to principal axis factoring to assess the construct validity. An initial exploratory analysis was conducted where items were included into a confirmatory analysis if: eigenvalues > 1, and logical item loading. The first exploratory analysis yielded five components with eigenvalues greater than 1, although only slightly for the last two (1.3 and 1.09 respectively). The shape of the scree plot was suggestive of a 4-factor solution, accounting for 56.42% of GFA item variance. Rotated item scores were then examined and ranked based on fit in the category and variance score. Items were excluded based on the following criteria: a) component matrix loading below .4; b) component matrix loading across three or more factors; c) highest component score at least .10 units higher than other component scores; d) component item ranking of 1. Five total items were removed, creating a fifteen item GFA-II. The remaining fifteen items served as latent variables for a confirmatory analysis, to determine if a four component model would fit the second data set. The model of best fit was analyzed, and was determined by using absolute (e.g. Goodness-of-Fit Index), Standardized Root Mean Squared Residual, and incremental fit statists. Fit indices suggest the items on the GFA-II provided a good fit to the data, χ²(141) = 269.08, p < .01, χ²/df= 1.9, RMSEA = .07, CFI = .9, and accounted for 64.25% of the total variance.
Implications:
The results of these analyses suggest that the GFA-II may be a clinically relevant psychometric for identifying the function of gambling behavior. Identifying variables that maintain gambling behaviors is considerably important when developing function-based treatments. Clinicians can use the GFA-II to identify functional relations for individual clients, as a way to make data based decisions on how to develop treatment.