Check for duplicate rows in response data

Duplicate rows in response data almost always indicate a merge bug or an accidental double-write. This check flags two distinct cases:

Usage

check_duplicates(
  responses,
  method = c("2ifc", "briefrc"),
  col_participant = "participant_id",
  col_stimulus = "stimulus",
  ...
)

Arguments

responses

A data frame of trial-level responses.

method

Either "2ifc" or "briefrc". Currently informational.

col_participant, col_stimulus

Column names.

...

Unused.

Value

An rcdiag_result() object. The data element contains full_duplicates and pair_duplicates data frames of the flagged rows.

Details

1. Fully duplicated rows (identical across every column).

2. Duplicated (participant, stimulus) pairs where the response or RT differs. These are harder to interpret: either the design legitimately repeats stimuli (in which case the check should be skipped), or the data pipeline glued together two runs for the same participant.

Examples

responses <- data.frame(
  participant_id = c(1, 1, 1, 2),
  stimulus       = c(1, 2, 2, 1),
  response       = c(1, -1, -1, 1)
)
check_duplicates(responses)
#> [WARN] Duplicates
#>   1 of 4 rows are fully duplicated.
#>   1 rows share a (participant, stimulus) pair with differing response or RT.