On Cronbach’s Alpha as the Mean of All Split-Half Reliabilities
A famous description of Cronbach’s alpha is that it is the mean of all (Flanagan–Rulon) split-half reliabilities. The result is exact if the test is split into two halves that are equal in size. This requires that the number of items is even, since odd numbers cannot be split into two groups of equal size. In this chapter it is shown that alpha is approximately identical to the mean of all split-half reliabilities, if a test consists of an odd number of items and has at least eleven items.
KeywordsSplit-half reliability Spearman–Brown prophecy formula Flanagan–Rulon split-half
This research was done while the author was funded by the Netherlands Organisation for Scientific Research, Veni project 451-11-026.
- Abramowitz M, Stegun IA (1970) Handbook of mathematical functions (with formulas, graphs and mathematical tables). Dover, New YorkGoogle Scholar
- Brown W (1910) Some experimental results in the correlation of mental abilities. Br J Psychol 3:296–322Google Scholar
- Field A (2009) Discovering statistics using SPSS, 3rd edn. Sage, Los AngelesGoogle Scholar
- Furr RM, Bacharach VR (2008) Psychometrics. an introduction. Sage, Los AngelesGoogle Scholar
- Lord FM, Novick MR (1968) Statistical theories of mental test scores (with contributions by A. Birnbaum). Addison-Wesley, ReadingGoogle Scholar
- McDonald RP (1999) Test theory: a unified treatment. Erlbaum, MahwahGoogle Scholar
- Rulon PJ (1939) A simplified procedure for determining the reliability of a test by split-halves. Harvard Educ Rev 9:99–103Google Scholar
- Spearman C (1910) Correlation calculated from faulty data. Br J Psychol 3:271–295Google Scholar
- Thorndike RL (1971) Educational measurement. American Council on Education, WashingtonGoogle Scholar