Dissertation Defenses

Doctoral Dissertation Announcement

---

Candidate: Andrew Ellis Brandt

Degree of: Doctor of Philosophy

Department: Psychology

Title: Experimental Analysis of Gambling Using a Concurrent-Schedules Procedure

Committee:
Dr. Cynthia J. Pietras, Chair
Dr. R. Wayne Fuqua
Dr. Scott Gaynor
Dr. Dorothea C. Lerman

Date: Monday, March 8, 2010 9:30 a.m. to 11:30 a.m.
3723 Wood Hall

Abstract:
Gambling has been experimentally investigated using various laboratory gambling procedures. Two features that are common to these procedures may enhance risk taking and therefore influence gambling: 1) the availability of only a single response option, and 2) the use of a participant stake (i.e., tokens given freely to the participant prior to a session). The present research investigates whether these procedural variables affect the probability of gambling using college student participants playing a newly-developed simulation, which uses a concurrent gamble no-gamble procedure. In all experiments, the gamble option produces token gains and losses probabilistically and tokens are exchangeable for entries into a $50 lottery. Experiments 1 and 2 compared gambling levels when only a single option was available (gamble-only conditions) and when both a gamble and a no-gamble option were available (concurrent conditions). In Experiment 1, during concurrent conditions choice of the no-gamble option was followed by a brief timeout and the start of the next trial. Gambling levels were similar across gamble-only and concurrent conditions. In Experiment 2, during concurrent conditions choice of the no-gamble option produced a fixed number of tokens with certainty. Gambling was lower in the concurrent conditions than gamble-only conditions. Experiment 3 investigated gambling levels under the concurrent conditions when a participant stake was or was not used. Under no-stake conditions, participants started the session without tokens but could earn them during the session by choosing the no-gamble option. Preference for the gamble option was higher under stake than no-stake conditions, but only during the first exposure to the task. Experiment 4 investigated gambling levels under the concurrent conditions when the number of tokens produced on the no-gamble option varied. Gambling levels were similar across all conditions. Overall, these experiments demonstrate that more gambling occurs when only a single-gamble option is available compared to when a gamble and a no-gamble option are available, and when participants gamble with staked tokens compared to earned tokens. These findings suggest that these features of existing laboratory gambling procedures may elevate risk taking and that a concurrent gamble no-gamble procedure may therefore be more useful for investigating gambling.