Executive Summary

This is a report of Phase II of a project that had two general aims:

1. To investigate the role of language and motivational factors in beginning reading development, with particular emphasis on determining the specific language-related factors at school entry that place some children at risk for failure.
2. To work with new entrant teachers to adapt, develop and test supplementary teaching strategies and materials designed to help all children, but especially those at risk for failure, to derive greater benefit from regular classroom instruction.

The introduction to the report provides the background to the project, the specific aims of each of the four phases of the project, the theoretical framework that guided the research undertaken in the project, and the predictive correlations between measures taken at the beginning, middle and end of Year 1 and reading achievement measures taken at the middle of Year 3. Following the introduction are five studies that examine in greater depth the structure of relationships between language-related factors, achievement-related beliefs, gender and beginning reading achievement. Although each study is presented as a self-contained report, the data discussed in each report are from the same longitudinal study in which the literacy development of 152 new school entrants (90 boys, 62 girls) in 16 schools was closely followed from school entry to the middle of Year 3.

The aim of the first study was to test a model of the relationships among the major learning task, learning strategies and cognitive prerequisites of beginning reading development. Path analyses were used to examine data collected toward the end of Year 1 (when the mean age of 141 target children was 5 years, 9 months), at the middle of Year 2 (when the mean age of the children was 6 years, 5 months), and at the end of Year 2 (when the mean age of the children was 6 years, 10 months). Results indicated that the ability to use letter-sound patterns and the ability to use sentence context made the strongest independent contributions to variance in early reading achievement; that each of these abilities appeared to influence the development of the other; that the use of letter-sound patterns exerted a stronger influence than the use of sentence context in relation to both early reading development and the reciprocal causation between these two factors; and that when extraneous variables and autoregressive effects were controlled, phonological sensitivity was the major factor influencing the ability to use letter-sound patterns, and grammatical sensitivity was the major factor influencing the ability to use sentence context. These findings suggest that when beginning readers encounter unfamiliar words in text, they should be encouraged to use letter-sound patterns first and then to use sentence context, but only to confirm hypotheses about what unfamiliar words might be, based on information from partial decoding attempts.

In the second study beginning readers' reported word identification strategies for identifying unfamiliar words in text were examined in relation to reading achievement, reading-related skills, and academic self-perceptions. Children who participated in the three-year longitudinal study of reading development were placed in two groups according to their reported word identification strategies obtained towards the end of their first year of schooling. Results indicated that children who reported using word-based strategies (e.g. "sound it out," "think of the sounds," "hear all the letters," "listen to what the letters are") showed superior reading and reading-related performance, and reported more positive self-efficacy beliefs in reading and more positive academic self-concepts than children who reported using text-based strategies (e.g. "guess", "think, guess what the word is," "read it over again," "read on," "have a look at the picture," "put my finger on the book and try other words and get a word that makes sense," "miss it out and go to the end and go back and guess a word that makes sense"). The results are discussed in terms of predictions stemming from the different theoretical assumptions about reading acquisition that underlie the code-emphasis and whole language approaches to beginning reading instruction.

The aim of the third study was to examine the emerging casual interplay between reading self-concept and beginning reading performance. Although achievement-related perceptions are casually related to academic performance, it is not clear at what age this relationship starts to form, especially in terms of learning to read. Path analyses of the data from the three-year longitudinal study were used to examine the relationships between reading-related skills and reading self-concept at the start of Year 1, the middle of Year 2, and the middle of Year 3. Results indicated that reading performance emerged as casually predominant over reading self-concept between the middle of Year 2 and the middle of Year 3. The findings suggest that the initial reading-related experiences in school are associated with the development of reading self-concepts within the first two and a half years of schooling. This period may mark the time during which negative Matthew (poor-get-poorer) effects develop for those who experience initial difficulties in learning to read.

In the fourth study the relations between academic self-concept and measures of reading-related performance and self-concept were examined in beginning school children who, after two years of schooling, were assessed as having positive, negative, or typical academic self-concepts. The data examined were obtained soon after school entry, toward the end of Years 1 and 2, and during the middle of Year 3. Children with negative academic self-concepts performed poorly on reading-related tasks and reported more negative reading self-concepts than did children with positive or typical academic self-concepts. Reading was also highly predictive of negative and positive academic self-concept group membership, but not of typical academic self-concept group membership. These findings suggest that past studies of relations between academic self-concept and achievement involving full-range samples of young children have underestimated the point in time when these factors become causally related to each other. The consequences of young children developing patterns of difficulty in learning to read are discussed.

The fifth study concerned early literacy achievement and gender. On the basis of the findings that boys consistently achieve lower than girls on School Certificate results, the Educational Review Office (1999) concluded that "boys and girls learn and respond in different ways and achieve best with different teaching styles". In the fifth study an alternative view of the differential performance of boys and girls is presented based on the absence of early literacy performance differences. Data from the longitudinal study of beginning literacy development were used to investigate bias in school-based identification of boys due to their tendency to display externalising classroom behaviours. Results indicated that the performance of boys and girls was not significantly different on all measures of early literacy given over the three-year period. Nonetheless, boys were identified for reading remediation proportionately twice as often as girls. Finally, teachers rated boys selected for Reading Recovery as exhibiting significantly more inappropriate externalising behaviours, a finding consistent with the literature suggesting an identification bias for underachievement based on inappropriate classroom behaviours.