Measuring the alphabetic principle: Mapping behaviors onto theory
Laugle, Kelly M.
Research suggests that development of the alphabetic principle is a critical factor in learning to recognize words and becoming a successful reader. The alphabetic principle encompasses both the understanding that relationships exist between letters and sounds and the application of these relationships to reading words. This study investigated the degree to which different measures of the alphabetic principle were predictive of later reading development. These measures were examined in the context of Ehri's phase theory of sight word development to investigate how different behaviors associated with the alphabetic principle fit within a developmental framework. Two cohorts of students (109 kindergarteners, 212 first graders) participated in this study from spring of 2007 until late fall of 2008 (58 second graders, 121 third graders). The predictive powers of single and combined measures of the alphabetic principle were analyzed using sequential regression. Results indicated that each measure explained significant between-student variation in performance on measures of word reading fluency, oral reading fluency (ORF), vocabulary, and reading comprehension. A measure of letter-sounds embedded in nonsense words appeared to have more utility for the prediction of reading outcomes than a measure of letter-sounds presented in isolation. Additionally, including a measure of nonsense words with a measure of letter-sounds embedded in nonsense words increased the predictive power of the model over and above the predictive power of letter sounds alone. Growth on ORF served as an additional criterion for the purpose of investigating the methodology of measuring growth. Two conceptualizations of growth were explored: raw score change over time and individual rates of growth over time (slope). Correlations and sequential regression were used to evaluate the relationship between raw score change and measures of the alphabetic principle. Hierarchical Linear Modeling (HLM) was used to model individual slopes on Lexile measures of ORF (LORF). In general, raw score change appeared largely unrelated to measures of the alphabetic principle. HLM analyses revealed that individual differences in slope on LORF were minimal and not very reliable, making the prediction of these differences difficult. Recommendations for future research and implications for practice are discussed.