Skip to contents

check_design() validates whether an analysis fits the narrow branching workflow implemented by statsguider. The function checks the presence of required columns, infers simple data properties, and reports issues and warnings before recommendation or execution.

Usage

check_design(
  data,
  outcome,
  group = NULL,
  id = NULL,
  goal = c("difference", "association", "adjusted_effect", "time_to_event", "agreement",
    "equivalence"),
  outcome_type = c("auto", "continuous", "binary", "nominal", "ordinal", "count"),
  paired = c("no", "yes"),
  repeated = c("no", "yes"),
  adjust = c("no", "yes")
)

Arguments

data

A data.frame.

outcome

Name of the outcome column.

group

Optional name of the group column.

id

Optional name of the subject identifier column.

goal

One of "difference", "association", "adjusted_effect", "time_to_event", "agreement", or "equivalence".

outcome_type

One of "auto", "continuous", "binary", "nominal", "ordinal", or "count".

paired

"yes" or "no".

repeated

"yes" or "no".

adjust

"yes" or "no".

Value

A named list with ok, issues, warnings, and inputs.