Testing the statistical learning of spelling patterns by manipulating semantic and orthographic frequency

This study tested the diverging predictions of recent theories of children's learning of spelling regularities. We asked younger (Grades 1 and 2) and older (Grades 3 and 4) elementary school-aged children to choose the correct endings for words that varied in their morphological structure. We tested the impacts of semantic frequency by including three types of words ending in -er: derived and inflected forms, the first of which are far more frequent across the language, and one-morpheme control forms. Both younger and older children were more likely to choose the correct ending for derived forms over one-morpheme control words. This difference emerged for inflected forms only for the older children. We also tested the impacts of orthographic frequency by contrasting performance on the two derived allomorphs -er and -or, the first of which is far more frequent, and comparison one-morpheme forms. Both younger and older children were more likely to choose the correct spelling for the derivational -er over the same letter pattern in control words. This difference did not emerge in either group for the -or ending. The implications of these findings for current models of children's spelling development are discussed.