Exclude participants who fail too many catch trials

Catch trials present the target image as one of the two choices; a correct response requires selecting the choice that matches the target. This function removes any participant whose proportion of incorrect catch-trial responses exceeds max_prop_wrong.

Usage

filter_failed_catch(d, max_prop_wrong = 0.2)

Arguments

d

Data frame with columns worker_id, head, winner, and loser.

max_prop_wrong

Numeric between 0 and 1. Participants with a proportion of wrong catch-trial responses strictly greater than this value are removed. Default: 0.2.

Value

The input data frame with rows belonging to excluded participants removed.

Details

A catch trial is identified by head == winner (correct response) or head == loser (incorrect response). Only rows that satisfy one of these conditions contribute to the catch-trial counts; other trial types are ignored.

Examples

if (FALSE) { # \dontrun{
d <- data.frame(
  worker_id = rep(c("p1", "p2"), each = 5),
  head      = c("cat","cat","dog","dog","cat",  "bird","bird","fish","fish","bird"),
  winner    = c("cat","dog","cat","dog","fox",   "bird","crow","fish","tuna","crow"),
  loser     = c("dog","cat","dog","cat","bear",  "crow","bird","tuna","fish","bird")
)
filter_failed_catch(d, max_prop_wrong = 0.4)
} # }