Identifies floor/ceiling effects in PRO data when >15% of responses are at the minimum or maximum possible scale values.
Usage
detect_floor_ceiling(
data,
score_var,
min_score,
max_score,
by_var = NULL,
threshold = 0.15,
subject_var = "USUBJID"
)Arguments
- data
Data frame containing PRO scores
- score_var
Character. Name of the score variable
- min_score
Numeric. Minimum possible score on the scale
- max_score
Numeric. Maximum possible score on the scale
- by_var
Character vector. Variables to group by (e.g., c("VISIT", "TRT01P"))
- threshold
Numeric. Threshold for flagging (default: 0.15 = 15%)
- subject_var
Character. Subject identifier (default: "USUBJID")
Value
A data frame with columns:
Grouping variables
n: number of subjects with non-missing scores
n_floor: count of subjects at minimum
pct_floor: percentage at minimum
floor_flag: TRUE if pct_floor > threshold
n_ceiling: count of subjects at maximum
pct_ceiling: percentage at maximum
ceiling_flag: TRUE if pct_ceiling > threshold
Examples
if (FALSE) { # \dontrun{
# Detect floor/ceiling effects in PRO scores
result <- detect_floor_ceiling(
data = adqs,
score_var = "AVAL",
min_score = 0,
max_score = 100
)
print(result)
} # }