Can strategic thinking be taught or is it a matter of cognitive ability and personality?

The intervention for the treatment group is reading brief information on game theory and seeing a picture with text providing basic logic of strategic thinking.

2021-06-02

2021-06-13

Answers in p-beauty contest game, which is a number between 0-100, and variable sum game, which is either row T or B. Answers will be analyzed based on theoretical predictions, which are Nash equilibrium in both games and additionally Level-k in p-beauty contest game. In the p-beauty contest game, the empirical performance against other players is also analyzed.

The level-k analysis is based on the Level-0 player answering randomly. To determine the personality, the test subjects do the Big 5 personality test from the parts to determine conscientiousness, openness to experience, and agreeableness. To determine cognitive ability, the test subjects do a short Raven's Progressive Matrices test including 12 questions.

Variables for age, sex, education level, math background, and whether the test person has familiarized with game theory before the test will also be collected.

The design involves test subjects doing the Big 5 personality test including questions determining conscientiousness, openness to experience, and agreeableness, and a short Raven’s Progressive Matrices test for cognitive ability. The test subjects in the treatment group get treatment of reading brief basic information of game theory and seeing a picture with text providing basic logic of strategic thinking. All players play the p-beauty contest game guessing any number between 0 and 100 and a variable-sum game choosing between two rows T and B.

My hypotheses are the following.

Hypothesis 1: Independent of the cognitive ability or personality individual can learn strategic thinking through an information nudge. Each individual has the inherent ability to learn new things when the proper information and guidance are provided.

Hypothesis 2: Individuals with cognitive ability above-median benefit more from the treatment answering closer to theoretical predictions and receiving higher payoffs. Individuals with higher cognitive ability have a faster processing speed implying that they can understand and apply the information more frequently under the relatively strict time limitation.

Hypothesis 3: Individuals with higher openness to experience, agreeableness, and conscientiousness benefit more from the treatment answering closer to theoretical predictions and receiving higher payoffs. Individuals high on conscientiousness have a strong will to succeed in given tasks and want to do their best to solve the problems at hand. Individuals high on openness to experience are curious and have high motivation to learn new things. Individuals high on agreeableness follow given guidelines and do their best to solve the ordered tasks in the experiment.

The hypotheses are tested by using Mann–Whitney U test.

The test subjects are recruited via Amazon Mechanical Turk.

Randomization by computer.

Experimental sessions.

No

2 experimental "sessions" over different time periods.

Ca. 120 test subjects from Amazon Mechanical Turk.

Ca. 60 individuals in the control group and ca. 60 individuals in the treatment group.

There no previous data available to do the power calculations based on a similar experiment. I have used the basic rules of thumb to determine the effect size. For both calculations, I used the power of 0.8. In the p-beauty contest game, the logic is relatively simple and I assume that the treatment gives a slightly below-average treatment effect. Thus, I have used Cohen's d of 0.4 implying required total sample size of 52 individuals. In the variable sum game, the logic is slightly more difficult to understand and I have chosen to use Cohen's d of 0.3 implying a required sample size of 90.