Simple function to return Cronbach's alpha

Usage

getChronbachAlpha(dat, na.rm = TRUE, verbose = TRUE)

Arguments

dat: data, matrix, data frame or tibble and numeric item data as columns
na.rm: logical which causes function to throw error if there is missing data
verbose: logical, if FALSE switches off message about n(items) and number of usable rows

Value

Cronbach as numeric

Background

Very simple function that does a fair bit of sanity checking on the input and returns Cronbach's alpha which was first described comprehensively in the classic paper Cronbach (1951). There is an extensive literature criticising alpha as an index of internal consistency/reliability often, to my reading, failing to note that Cronbach's original paper was quite clear about the issues. Alpha is not a measure of unidimensionality but a measure of shared covariance and there are strong arguments that it is not the best indicator of internal reliability when there are marked differences in variance between items. The various arguments have led many to recommend McDonald's omega as a better measure than alpha. Perhaps partly because the canonical reference to McDonald's omega is to his book (McDonald, 1999) and not to a paper, the literature on this is, to my mind again, a bit complicated. There is good coverage in the help for omega() in Revelle's package psych omega. Despite the complexities alpha remains the most reported statistic for internal relability of multi-item measures with some good reasons: it is simple, robust and never claimed to be doing things it doesn't do!

History/development log

Started before 5.iv.21

References

Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. McDonald R. P. (1999) Test Theory: A Unified Treatment. Mahwah, NJ: Erlbaum

Author

Chris Evans

Examples

if (FALSE) { # \dontrun{
### make some nonsense data
tmpMat <- matrix(rnorm(200), ncol = 10)
### default call
getChronbachAlpha(tmpMat)
### switch off message with dimensions of non-missing data
getChronbachAlpha(tmpMat, verbose = FALSE)
} # }