RESEARCH QUESTION
Are individual differences in conditionability using evaluative conditioning (i.e., association between constructs via cognitive processing) related to degree of risk-taking during gambling on a simulated slot machine?

PURPOSE
The aim of the present study was to investigate whether positive and negative evaluative conditioning were related to risk-taking on a simulated slot machine. In addition, the relationship between behavioural inhibition/behavioural activation and risk taking on the simulated slot machine was investigated.

HYPOTHESIS
None stated.

PARTICIPANTS
Participants were 100 undergraduate students (average age = 21 years; 49% males) at the University of Bergen, none of whom were present or previous possible pathological gamblers (i.e., they had South Oaks Gambling Screen-Revised scores < 5). 

PROCEDURE
Upon arriving at the laboratory, participants were asked to complete questionnaires related to behaviour inhibition and activation. Following the questionnaires, each participant completed a computerized evaluative conditioning paradigm and a simulated slot machine task.

MAIN OUTCOME MEASURES
The behavioural inhibition scale assessed predisposition to avoid threatening or punishing stimuli and was parsed into the fight-flight-freeze system and behavioural inhibition anxiety. The behavioural activation scale assessed predisposition to approach appetitive stimuli and was subdivided into drive, fun seeking, and reward responsiveness. For the Evaluative Conditioning (EC) Paradigm, 18 pictures from the International Affective Picture System were presented on the computer screen, 12 of which were pre-rated as neutral in valence and arousal (e.g., shadow of person), 3 of which were pre-rated with high positive valence and high arousal (e.g., children on roller coaster), and 3 of which were pre-rated with high negative valence and high arousal (e.g., crying baby). During the pre-evaluation phase, participants were instructed to look at each picture. For the pre-conditioning phases, participants rated how much they liked or disliked each picture using a visual analog scale. Nine picture pairs were assembled for use in the conditioning phase: 3 neutral-negative, 3 neutral-positive, and 3 neutral–neutral. Each picture pair was presented 5 times. The conditioning phase was followed by the post-conditioning phase, which was identical to the pre-conditioning phase. For the simulated slot machine task participants were informed that they were going to play a game that resembled a slot machine, that they would play 100 mandatory trials, that they would start with 3000 NOK (i.e., Norwegian Kroner; approx. $540 Canadian) and that they would receive a payment that equalled 10% of the amount that remained in the bank after the session. Participants were instructed to choose bet-size (big win was bet-size times 4.5, and had a 10% chance of occurring; small win was bet-size times 2.25, and had a 20% chance of occurring. If neither a big or small win was obtained, the win was zero). Average bet-size across the 100 trials was calculated for use in the analyses.

KEY RESULTS
The results of the EC paradigm showed a flooring effect where a large proportion of participants did not show negative EC or positive EC. Among the participants who did show negative EC or positive EC, the scores lacked a normal distribution. Therefore, negative EC and positive EC were coded as dichotomous (1 = yes, 0 = no) variables. Participants who did not show negative evaluative conditioning or positive evaluative conditioning gambled with lower average bet-size compared to the other participants. Scores on the fight-flight-freeze system subscale of behavioral inhibition and scores on the reward responsiveness subscale of behavioral approach were positively associated with average bet-size. Participants who scored low on both behavioural approach reward responsiveness and behavioural inhibition of fight-flight-freeze system had lower bet-sizes compared to the other participants. Overall, the findings suggest that conditionability and reinforcement sensitivity play a role in gambling behavior.

LIMITATIONS
The study results may have limited generalizability because only healthy participants with a restricted age range were included. The study may also have low ecological validity since studies have shown that gambling in a laboratory and gambling in a casino may be experienced differently. A total of 18 participants needed to be excluded from the evaluative conditioning analysis because they did not rate the positive unconditioned stimuli more positively or the negative unconditioned stimuli more negatively than the conditioned stimuli they were paired with.

CONCLUSIONS
Despite the fact that conditioning is regarded as fundamental in the development of gambling problems, the present study was the first to investigate the role of evaluative conditioning in risk-taking during gambling. Overall, the results suggested that conditionability and reinforcement sensitivity play a role in gambling behavior. Future research should investigate whether individual differences in conditionability are associated with an increased vulnerability for developing gambling problems.