The influence of immediate gamble-information feedback in the Iowa gambling task - Reevaluating the Maia & McClelland (2004) study

• Main Authors: Shi-Shih Yen, 顏世旭
• Other Authors: Yao-Chu Chiu
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/96n3jc

碩士 === 東吳大學 === 心理學系 === 101 === Background: Bechara et al., (1997) and Maia & McClelland (2004) possessed controversial prospects on the knowledge penetration (late vs. early) of gamble structure (conscious to gamble structure). Nevertheless, both research teams congruently demonstrated that most normal subjects approached the advantageous cards and figured out the task structure eventually during playing the Iowa Gambling Task (IGT) -This is so called the viewpoint of expected value (EV). However, in recent years, a few research groups (Chiu et al., 2008； Fernie & Tunney, 2006； Fum et al., 2008；Lin et al., 2007; Steingroever et al., 2013; Wilder et al., 1998) found a confounding variable, namely the gain-loss frequency (GLF) rather than EV mostly dominated the choice behavior in the IGT. Notably, this argument between GLF vs. EV only involved choice-behavior level. Nevertheless, few researches verifiy the consistency between choice behavior and knowledge of gamble structure based on the two viewpoints (GLF vs. EV). Therefore, this study provided three experiments to evaluate the consistency between participants’ choice behavior and their knowledge for gamble structure. Method: Experiment 1 aims to further evaluate the critical experiment conducted by Maia & McClelland (2004). Namely, we are curious about what kind of gamble knowledge (GLF vs. EV) mainly guide the subjects' choice behavior in IGT & reverse version IGT (rIGT). Experiment 2 & 3 extended Experiment 1 and the questionnaire used by Maia & McClelland (2004) for evaluating subjects’ knowledge on gamble structure was replaced with the immediate gamble-information feedback (Including the information of immediate gain/loss and cumulative gain/loss in each deck). This manipulation is to eliminate the shortage by using the questionnaire (Maia & McClelland, 2004). Result: The result of Experiment 1 demonstrated that GLF rather than EV mostly guided the subjects’ choice behavior. Moreover, in Experiment 2 & 3, the immediate gamble-information was further provided to subjects, they were still unable to penetrate the EV of each deck and approach the positive-outcome choice. The present findings are inconsistent with Maia & McClelland (2004) and Bechara et al. (1997) observation, but largely consistent with the viewpoint of GLF.