The «Yellow-Red test» : A closer look at data from a longitudinal pilot project with Norwegian children
Abstract
The intention of this report has been to identify interesting questions for further work on the Yellow/Red test battery. Rather simple analyses have been carried out, partly due to the limited size of the sample (N=148).
The test scores consistently improve over time, as does the percentage of correct responses. Also, items within the same game elicit rather different responses, and item differences remain rather constant across replications. However, the original summed scales may have room for some improvement in both reliability and validity.
Consequently, alternate coding paradigms are suggested. They are neither proven nor disproven, however, by our initial recoding attempts. With larger samples, the recoding ideas may be put to a better test.