Do Spelling Tests Measure the Ability to Spell?

Publisher Website
Google Scholar
Abstract
In order to investigate the validity of spelling tests against the criterion of an estimate of spelling accuracy in a sample of writing, three tests with differing formats were constructed from a common pool of words. Test `A' was a proof reading/correction test. Test `B' followed a multiple choice/cloze format, and Test `C' was a traditional dictated word test. Following the administration of these tests to three Standard III and three Standard IV classes, a sample of writing was collected. Each written script was ranked independently by three judges on the basis of their interpretation of five written criteria. The mean of the three rankings was calculated, and then transformed into a standard score. To help assess the possible impact of sampling on the results of the study, the Standard III and Standard IV groups were randomly divided into two sub-samples of n = 40. There were no significant differences between the mean scores of the two sub-samples at each class level. The median correlation between the rankings of writing and Tests `A' and `C' (the tests emphasising the production of correct spelling) was greater than the corresponding correlation for Test `B' (the test emphasising the recognition of correct spelling). The results of correcting the correlations for attenuation suggested that variations in reliability of the tests and rankings were not responsible for the relationships that had been found.

This publication has 8 references indexed in Scilit: