Also known as the Spearman-Brown “prediction” or even “correction” formula and apparently there’s an argument that it should be Brown-Spearman, but I’ve never seen it that way round.
Details #
The formula tells you that, if you have the internal reliability of a multi-item measure, what internal reliability, all things being equal, the measure would have had it more or fewer items. Like this:
The formula does make the assumption that the different length instrument would behave the same way which is almost certainly not the case, however, the predictions are probably a sensible guide and the best we can have in the absence of actual empirical data for the other lengths. See my Rblog for a bit more information and shiny app if you want to play with the predictions the formula makes.
Try also #
Cronbach’s coefficient alpha
Internal reliability
Reliability
Chapters #
The issue of lengths of measures is touched on in Chapter 4 but the formula is not discussed: too geeky!
Online resources #
More detail in my Rblog: entry here.
App in my shiny apps which computes the predictions for you here.
Dates #
First created 3.v.24.
