IAT

```{library(plyr) # must import before dplyr library(tidyverse) library(data.table) library(pim)

Data acquisition and cleaning

Set the working directory in which to look for data files, then read them in.

files <- list.files(pattern = “\.csv$”) # Create a list of the csv files in this folder df <- tbl_df(rbind.fill(lapply(files, fread, header=TRUE)))

Make some variable names more transparent, plus rectify the the accuracy variable

df <- rename(df, block = blocks.thisN, # current block trial = trials.thisN, # trial order of presentation within each block rt = requiredResponse.rt, # first correct response rt accuracy = feedbackResponse.corr, # accuracy of the first response (currently reversed: 0 = correct) trial = trials.thisN) %>% # trial by order of presentation within each block mutate(accuracy = abs(accuracy - 1), # recitfies the direction of accuracy so that 0 = error and 1 = correct. block = block + 1) # recifies block to be 1 to 7 rather than 0-6

Extract demographics data

demographics_data <- select(df, participant,

     expName,
     date) %>%

distinct(participant)

IAT D1 scores

D1_df <-
group_by(df, participant) %>% filter(rt <= 10000) %>% summarize(RT_block3_m = mean(rt[block == 3]), # blocks 3 and 6 RT_block6_m = mean(rt[block == 6]), RT_block3and6_SD = sd(rt[block == 3 | block == 6]), RT_block4_m = mean(rt[block == 4]), # blocks 4 and 7 RT_block7_m = mean(rt[block == 7]), RT_block4and7_SD = sd(rt[block == 4 | block == 7])) %>% mutate(difference_a = RT_block6_m - RT_block3_m, IAT_D1a = difference_a / RT_block3and6_SD, difference_b = RT_block7_m - RT_block4_m, IAT_D1b = difference_b / RT_block4and7_SD, IAT_D1 = round((IAT_D1a + IAT_D1b)/2, 2), block_order = (participant - 1) %% 2 + 1, # set block order based on modulus of the participant code, as in the psychopy script IAT_D1 = ifelse(block_order == 1, IAT_D1, # D1 score rectifications for block order. if ==2, invert D1 score ifelse(block_order == 2, IAT_D1 * -1, “error”)), IAT_D1a = ifelse(block_order == 1, IAT_D1a, # D1 score rectifications for block order. if ==2, invert D1 score ifelse(block_order == 2, IAT_D1a * -1, “error”)), IAT_D1b = ifelse(block_order == 1, IAT_D1b, # D1 score rectifications for block order. if ==2, invert D1 score ifelse(block_order == 2, IAT_D1b * -1, “error”))) %>% select(participant, IAT_D1, IAT_D1a, IAT_D1b)

Accuracy and latency stats

accuracy_summary_df <- group_by(df, participant) %>% summarize(critical_blocks_percentage_accuracy = round(sum(accuracy[block == 3 | block == 4 | block == 6 | block == 7]) / 120, 2)) # 120 critical trials in the IAT

df\(fast_trial <- ifelse(df\)rt < .3, 1, 0) # add new column that records if RT < 300. Should probably be done using mutate() for code consistency, but this works fine.

latency_summary_df <- group_by(df, participant) %>% filter(block == 3 | block == 4 | block == 6 | block == 7) %>% summarize(critical_blocks_mean_rt = round(mean(rt), 2), percent_fast_trials = sum(fast_trial)/120) %>% # 120 critical trials mutate(exclude_based_on_fast_trials = ifelse(percent_fast_trials>=0.1, TRUE, FALSE)) %>% # exclusions based on too many fast trials is part of the D1 algorithm select(participant, critical_blocks_mean_rt, exclude_based_on_fast_trials)

Join data frames

all_tasks_df <- join_all(list(demographics_data, accuracy_summary_df, latency_summary_df, D1_df), by = “participant”, type = “full”)

}

```