Abstract
There were two purposes for this study: 1) to determine whether children's spelling strategies changed according to the grade level of the child and the particular demands of the kind of word the child was trying to spell, and 2) to examine the relationship between children's spelling strategics and their cognitive development measured in terms of Piaget's notion of decentralized.
Two lists of words, each of which contained three exemplars of five spelling categories (Lax Vowel, Tense Vowel, Past Tense Marker, Consonant Doubling, and Vowel Extension) were administered to 15 children each from grades one through four. Appropriate scores based on specific spelling strategies were assigned to each child's attempts. A two-way hierarchically partitioned analysis of variance was performed on these scores to determine the effects of grade, spelling category, and their interaction. Frequency counts of the use of specific strategies were tabulated by category for each grade in order to provide more descriptive data about what the children were trying to do.
A decentration test battery (specially constructed for this study) was also administered to the 60 children. This instrument was divided into seven specific areas: conservation of number, mass, continuous quantity, weight and volume, class inclusion, and the <i>Picture Integration Test</i> (Elkind, in press). A canonical correlation and two R factor analyses (one with grade level controlled) were performed on the children's five spelling category scores and their scores on the seven decentration tasks in order to determine if there was a relationship between these two sets of variables.
The results of the analysis of variance indicated that the effects for grade, category, and their interaction were significant (<i>p</i><,01). The findings suggested a progressive pattern of spelling strategies dependent upon both the child's familiarity with written language and his intellectual maturity. However, little evidence was found for some of the transitional spelling strategies noted by previous investigators. This may have been due to the immaturity of the second grade sample.
The canonical correlation between the spelling and decentration variables was significant (r = .67, <i>p<i/> <.01). The first factor analysis clearly indicated a two factor solution. All the decentration variables loaded significantly on the first factor, while none of the spelling variables loaded on this factor. The pattern of the second factor was exactly the reverse. However, the two factors were significantly correlated (r = .56, <i>p<i/> < .01). Partial correlations with grade level control was used to general a second factor analysis. The factor patterns remained the same, and the correlation between the two factors, though somewhat reduced remained significant (r = .36, <i>p<i/> <, 01).
