Limits...
Self-concept in fairness and rule establishment during a competitive game: a computational approach.

Lee SH, Kim SP, Cho YS - Front Psychol (2015)

Bottom Line: When the models were fitted to a continuous time window, the parameters of the players in a pair with "switching" and "mutual avoidance" patterns became similar as the game proceeded, suggesting that the players gradually formed a shared rule during the game.In contrast, the difference of parameters between the players in the "unfair" and "mutual rush" patterns did not become stable.The outcomes of the present study showed that people are likely to change their strategy until they reach a mutually beneficial status.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology, Korea University Seoul, South Korea.

ABSTRACT
People consider fairness as well as their own interest when making decisions in economic games. The present study proposes a model that encompasses the self-concept determined by one's own kindness as a factor of fairness. To observe behavioral patterns that reflect self-concept and fairness, a chicken game experiment was conducted. Behavioral data demonstrates four distinct patterns; "switching," "mutual rush," "mutual avoidance," and "unfair" patterns. Model estimation of chicken game data shows that a model with self-concept predicts those behaviors better than previous models of fairness, suggesting that self-concept indeed affects human behavior in competitive economic games. Moreover, a non-stationary parameter analysis revealed the process of reaching consensus between the players in a game. When the models were fitted to a continuous time window, the parameters of the players in a pair with "switching" and "mutual avoidance" patterns became similar as the game proceeded, suggesting that the players gradually formed a shared rule during the game. In contrast, the difference of parameters between the players in the "unfair" and "mutual rush" patterns did not become stable. The outcomes of the present study showed that people are likely to change their strategy until they reach a mutually beneficial status.

No MeSH data available.


BIC scores of the models for each behavioral pattern. Error bars indicate standard errors. Asterisks refer to the significant difference from the baseline at 5% level.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4561810&req=5

Figure 5: BIC scores of the models for each behavioral pattern. Error bars indicate standard errors. Asterisks refer to the significant difference from the baseline at 5% level.

Mentions: Figure 5 shows the BIC scores of the models for each group of pairs with the distinct patterns averaged by the number of pairs. Those averaged BIC scores were compared by the paired t-test. The baseline was defined as the BIC score when the probabilities of actions for every trial were predicted as 50% each, to test if the models' BIC scores are lower than the score at the chance level prediction. The summary of comparison with the baseline is shown in Table 2. The Fehr and Schmidt model showed a low explanatory power for overall behavioral patterns from every pair of participants. Overall BIC score of this model (M = 271.66, SD = 32.17) was not significantly different from the baseline prediction (M = 277.26, SD = 0), p = 0.6012. Cox et al.'s model and the self-concept model both decently predicted all the distinct patterns. They explained better than the Fehr and Schmidt model and the baseline prediction for every distinct pattern (ps < 0.01). The overall BIC score of the self-concept model (M = 204.11, SD = 60.37) was significantly lower than that of Cox et al.'s model (M = 222.42, SD = 55.68), p < 0.01. Specifically, the self-concept model explained the switching pattern and undefined pattern significantly better than Cox et al.'s model (ps < 0.01). The self-concept model also showed a lower BIC score at the mutual rush and avoidance patterns than Cox et al.'s model but the difference was not statistically significant (ps = 0.1347 and 0.2275, respectively). Note, however, that the t-test on the BIC scores of the avoidance pattern was not reliable because there were only two samples. Similarly, the BIC scores of the unfair pattern were inappropriate for the t-test as there was only one pair with an unfair pattern. In addition, the self-concept model was the only model with a BIC score for undefined patterns significantly lower than that of the baseline prediction, p = 0.0434. The Fehr and Schmidt model also showed significant difference in undefined patterns, p = 0.0125, but the BIC score was higher than the baseline, implying that the model predicts worse than at chance level. No other significant difference was found between the models.


Self-concept in fairness and rule establishment during a competitive game: a computational approach.

Lee SH, Kim SP, Cho YS - Front Psychol (2015)

BIC scores of the models for each behavioral pattern. Error bars indicate standard errors. Asterisks refer to the significant difference from the baseline at 5% level.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4561810&req=5

Figure 5: BIC scores of the models for each behavioral pattern. Error bars indicate standard errors. Asterisks refer to the significant difference from the baseline at 5% level.
Mentions: Figure 5 shows the BIC scores of the models for each group of pairs with the distinct patterns averaged by the number of pairs. Those averaged BIC scores were compared by the paired t-test. The baseline was defined as the BIC score when the probabilities of actions for every trial were predicted as 50% each, to test if the models' BIC scores are lower than the score at the chance level prediction. The summary of comparison with the baseline is shown in Table 2. The Fehr and Schmidt model showed a low explanatory power for overall behavioral patterns from every pair of participants. Overall BIC score of this model (M = 271.66, SD = 32.17) was not significantly different from the baseline prediction (M = 277.26, SD = 0), p = 0.6012. Cox et al.'s model and the self-concept model both decently predicted all the distinct patterns. They explained better than the Fehr and Schmidt model and the baseline prediction for every distinct pattern (ps < 0.01). The overall BIC score of the self-concept model (M = 204.11, SD = 60.37) was significantly lower than that of Cox et al.'s model (M = 222.42, SD = 55.68), p < 0.01. Specifically, the self-concept model explained the switching pattern and undefined pattern significantly better than Cox et al.'s model (ps < 0.01). The self-concept model also showed a lower BIC score at the mutual rush and avoidance patterns than Cox et al.'s model but the difference was not statistically significant (ps = 0.1347 and 0.2275, respectively). Note, however, that the t-test on the BIC scores of the avoidance pattern was not reliable because there were only two samples. Similarly, the BIC scores of the unfair pattern were inappropriate for the t-test as there was only one pair with an unfair pattern. In addition, the self-concept model was the only model with a BIC score for undefined patterns significantly lower than that of the baseline prediction, p = 0.0434. The Fehr and Schmidt model also showed significant difference in undefined patterns, p = 0.0125, but the BIC score was higher than the baseline, implying that the model predicts worse than at chance level. No other significant difference was found between the models.

Bottom Line: When the models were fitted to a continuous time window, the parameters of the players in a pair with "switching" and "mutual avoidance" patterns became similar as the game proceeded, suggesting that the players gradually formed a shared rule during the game.In contrast, the difference of parameters between the players in the "unfair" and "mutual rush" patterns did not become stable.The outcomes of the present study showed that people are likely to change their strategy until they reach a mutually beneficial status.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology, Korea University Seoul, South Korea.

ABSTRACT
People consider fairness as well as their own interest when making decisions in economic games. The present study proposes a model that encompasses the self-concept determined by one's own kindness as a factor of fairness. To observe behavioral patterns that reflect self-concept and fairness, a chicken game experiment was conducted. Behavioral data demonstrates four distinct patterns; "switching," "mutual rush," "mutual avoidance," and "unfair" patterns. Model estimation of chicken game data shows that a model with self-concept predicts those behaviors better than previous models of fairness, suggesting that self-concept indeed affects human behavior in competitive economic games. Moreover, a non-stationary parameter analysis revealed the process of reaching consensus between the players in a game. When the models were fitted to a continuous time window, the parameters of the players in a pair with "switching" and "mutual avoidance" patterns became similar as the game proceeded, suggesting that the players gradually formed a shared rule during the game. In contrast, the difference of parameters between the players in the "unfair" and "mutual rush" patterns did not become stable. The outcomes of the present study showed that people are likely to change their strategy until they reach a mutually beneficial status.

No MeSH data available.