pickGroupMethod: pickGroupMethod

Usage Arguments


pickGroupMethod(data, continList = NULL, catList = NULL, grouping = NULL,
  minCount = NULL, minExpectedCountProportion = 0.75,
  maxLevelsCutoff = NULL, dropOneRow = T, dropNLessThan = 1,
  testNormality = T, minDiscreteValues = NULL, varEqual = T,
  verbose = T)



data.frame containing all data


continList Vector of continuous variable names


catList Vector of categorical variable names


grouping Grouping variable to test for equal variance and other multiple group based assumptions


Minimum total count


Minimum proportion of cells that have an expected count of 5 or more to still use Chi.square test. Default of 75


maxLevelsCutoffMaximum number of unique levels to a variable to still analyze. If the variable exceeds this cutoff it will be dropped from the analysis lists.


dropOneRowIf TRUE, any variables with only 1 unique level will be dropped from analysis lists.


dropNLessThanMinimum number non-NA responses a variable must have before being dropped from the analysis lists. Default of 1


testNormalityIf TRUE, normality assumptions are checked for continuous variables


minDiscreteValuesminDiscreteValues Minimum number of discrete values of a continuous variable to still be tested with parametric methods.


varEqualIf TRUE, and grouping is non-NULL, equal variance assumptions will be tested


verboseIf TRUE, messages about decision criteria will be printed to the console.

List of vectors indicating whether variances should be tested parametrically, non-parametrically or dropped, for continuous and categorical lists separately pickGroupMethod will test common assumptions for parametric and non-parametric methods for 2 + group methods (e.g. t.test, mann-whitney U, chisq, fisher, ANOVA) and will decide which method to use for each variable. #NULL

TaylorAndrew/atPrepAnalyze documentation built on May 9, 2019, 4:23 p.m.