Package index
Checking one study
Run once per cleaned experimental dataset. Each check takes a study_id and appends it to the output, so results can be stacked across studies.
-
check_y_bounds() - Check that outcome variables are within [0, 1]
-
check_missingness_nona() - Check covariate missingness and imputed-version coverage
-
check_balance() - Check covariate balance across treatment conditions
-
check_smd() - Standardized mean differences between treatment arms
-
check_attrition() - Check differential attrition across treatment conditions
-
check_covariate_missingness() - Summarize and visualize missingness of covariates
Checking many studies
Collapse per-study results and read them without over-reading them. Balance p-values are uniform under a valid design, so a count of flags is the wrong summary; the distribution is the right one.
-
stack_checks() - Stack per-study check results into one list of tibbles
-
report_checks() - Triage a stacked set of design checks
-
summarize_check_pvalues() - Summarize a set of design-check p-values against the uniform reference
-
plot_check_pvalues() - Plot design-check p-values against the uniform reference
Data shape
Check that a cleaned dataset has the structure a pipeline expects, as distinct from checking that the experiment behind it was sound.
-
check_schema() - Check that a cleaned data frame conforms to its schema
-
assert_schema() - Assert a cleaned data frame conforms to its schema
-
assert_key_unique() - Assert that a data frame's row key is unique
-
assert_no_labelled() - Assert that no column is a leaked haven_labelled
Cleaning helpers
Utilities for building cleaned data, and code generators that emit copy-pasteable cleaning and checking blocks.
-
stat_mode() - Statistical mode
-
scale_by_control() - Scale outcome variables by control-group standard deviation (Glass's delta)
-
check_s_scaling() - Check that
_s(Glass's delta) variables are correctly standardized -
write_covariate_imputation_code() - Write covariate imputation code
-
write_outcome_missingness_dummies_code() - Write outcome missingness code
-
write_balance_check_code() - Write balance check code
-
write_attrition_check_code() - Write attrition check code
-
excheckrexcheckr-package - excheckr: Tools for Exploring and Checking Experimental Data