Motivated Reasoning in an Explore-Exploit Task

Authors: Caddick, Z. A., Rottman, B. M.

Lab: Causal Learning and Decision Making Lab

Paper Citation: Caddick, Z.A., Rottman, B.M. (2021). Motivated reasoning in an explore-exploit task. Cognitive Science, 45(8), e13018. doi:10.1111/cogs.13018

For help or more information contact caddickzac@gmail.com.

Note: Due to the complexity of the data and sheer volume of analyses conducted, we follow the organization presented in the published paper and organized the datasets and R scripts by analysis.

R Scripts and Data

Study 1

3.2.1 Choices in the Learning Task.

3.2.1.1 Switching Multiple Policies to the Preferred Option at the Beginning of Learning (Table 1, Figures 4 and 5).

Counts of Testing Instances by Type (Table 1) & Number of controlled and confounded changes to policies per trial (Figure 4). {R script, data (.csv)}

Number of Trials Until Testing by Preference (Figure 5). {R script, data (.csv)}

3.2.1.2 Changing a Policy and Holding others Stable for Periods of Time (Table 2). {R script, data (.csv)}

3.2.1.3 Never Testing Bias by Preference (Figure 6). {R script, data (.csv)}

3.2.1.4 Percent of Trials During which the Preferred Option Was Selected (Figure 7). {R script, data (.csv)}

3.2.1.5 Percent of Trials the Optimal Policy was Selected by Preference (Figure 8). {R script, data (.csv)}

3.2.2 Judgments of Policy Efficacy after the Learning Task.

3.2.2.1 Causal Functions (Figure 9). {R script, data (.csv)}

3.2.2.2 Non-Causal Functions (Figure 10). {R script, data (.csv)}

3.2.3 Function Identification.

3.2.3.1 Causal Functions (Table 3). {R script, data (.csv)}

3.2.3.2 Non-Causal Functions (Table 4). {R script, data (.csv)}

Study 2A (MTurk sample)

4.2.1 Choices in the Learning Task.

4.2.1.1 Switching Multiple Policies to the Preferred Option at the Beginning of Learning (Table 1, Figures 4 and 5).

Counts of Testing Instances by Type (Table 1). & Number of controlled and confounded changes to policies per trial (Figure 4). {R script, data (.csv)}

Number of Trials Until Testing by Preference (Figure 5). {R script, data (.csv)}

4.2.1.2 Changing a Policy and Holding others Stable for Periods of Time (Table 2). {R script, data (.csv)}

4.2.1.3 Never Testing Bias by Preference (Figure 6). {R script, data (.csv)}

4.2.1.4 Percent of Trials During which the Preferred Option Was Selected (Figure 7). {R script, data (.csv)}.

4.2.1.5 Percent of Trials the Optimal Policy was Selected by Preference (Figure 8). {R script, analysis 1 data (.csv), analysis 2 data (.csv)}

4.2.2 Judgments of Policy Efficacy after the Learning Task.

4.2.2.1 Causal Functions (Figure 9). {R script, data (.csv)}

4.2.2.2 Non-Causal Function (Figure 10). {R script, data (.csv)}

4.2.3 Function Identification.

4.2.3.1 Causal Functions (Table 3). {R script, data (.csv)}

4.2.3.2 Non-Causal Functions (Table 4). {R script, data (.csv)}

Study 2B (Intro. Psych sample)

4.2.1 Choices in the Learning Task.

4.2.1.1 Switching Multiple Policies to the Preferred Option at the Beginning of Learning (Table 1, Figures 4 and 5).

Counts of Testing Instances by Type (Table 1). & Number of controlled and confounded changes to policies per trial (Figure 4). {R script, data (.csv)}

Number of Trials Until Testing by Preference (Figure 5). {R script, data (.csv)}

4.2.1.2 Changing a Policy and Holding others Stable for Periods of Time (Table 2). {R script, data (.csv)}

4.2.1.3 Never Testing Bias by Preference (Figure 6). {R script, data (.csv)}

4.2.1.4 Percent of Trials During which the Preferred Option Was Selected (Figure 7). {R script, data (.csv)}

4.2.1.5 Percent of Trials the Optimal Policy was Selected by Preference (Figure 8). {R script, analysis 1 data (.csv), analysis 2 data (.csv)}

4.2.2 Judgments of Policy Efficacy after the Learning Task.

4.2.2.1 Causal Functions (Figure 9). {R script, data (.csv)}

4.2.2.2 Non-Causal Function (Figure 10). {R script, data (.csv)}

4.2.3 Function Identification.

4.2.3.1 Causal Functions (Table 3). {R script, data (.csv)}

4.2.3.2 Non-Causal Functions (Table 4). {R script, data (.csv)}

Study 3

5.2.1 Choices in the Learning Task.

5.2.1.1 Switching Multiple Policies to the Preferred Option at the Beginning of Learning (Table 1, Figures 4 and 5).

Counts of Testing Instances by Type (Table 1). & Number of controlled and confounded changes to policies per trial (Figure 4). {R script, data (.csv)}

Number of Trials Until Testing by Preference (Figure 5). {R script, data (.csv)}

5.2.1.2 Changing a Policy and Holding others Stable for Periods of Time (Table 2). {R script, data (.csv)}

5.2.1.3 Never Testing Bias by Preference (Figure 6). {R script, data (.csv)}

5.2.1.4 Percent of Trials During which the Preferred Option Was Selected (Figure 7). {R script, data (.csv)}

5.2.1.5 5 Percent of Trials the Optimal Policy was Selected by Preference (Figure 8). {R script, data (.csv)}

5.2.2 Judgments of Policy Efficacy after the Learning Task.

5.2.2.1 Causal Functions (Figure 9). {R script, data (.csv)}

5.2.2.2 Non-Causal Functions (Figure 10). {R script, data (.csv)}

5.2.3 Function Identification.

5.2.3.1 Causal Functions (Table 3). {R script, data (.csv)}

5.2.3.2 Non-Causal Functions (Table 4). {R script, data (.csv)}

Figures

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Tables

Table 1

Table 2

Table 3

Table 4