Predictors of Punjabi, Hindi and English reading comprehension among multilingual children in the Punjab Region of India
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
The research reported in this thesis investigated cognitive-linguistic predictors of reading comprehension (both within and across languages) among multilingual primary school children in the Punjab region of India. The participants of this study learn three languages: Punjabi, Hindi and English; and are exposed to all three from the initial stage of literacy acquisition. Unlike English, the Punjabi and Hindi orthographies are written nonlinearly with a horizontal bar on the top of the aksharas that connects aksharas within a word, and include vowel symbols that have independent and dependent forms. Both Punjabi and Hindi are alphasyllabic orthographies, whereas English is an alphabetic orthography.
Over 400 trilingual school children in Punjab (India) from grades 2 to 5 completed a measure of text reading comprehension that comprised passages followed by questions about details in those passages. Reading comprehension levels were compared to the measures of listening comprehension, phonological processing, orthographic knowledge and speed of processing. Analyses indicated the Punjabi, Hindi and English reading comprehension levels were predicted by measure of listening comprehension and word decoding, with the latter being predicted by phonological and orthographic skills. Such findings were consistent with current models of reading derived from studies of English. However, in contrast to these models, measures of orthographic skills were also predictive of variance in reading comprehension independent of word decoding across Punjabi, Hindi and English models. Contributions of phonological processing and speed of processing were also observed in the English reading comprehension model, again independent of word decoding processes. Overall, Punjabi and Hindi reading comprehension was predicted by similar predictors, with English reading comprehension showing more variations in predictors.
Further analyses investigated the influence of Punjabi and Hindi cognitive-linguistic skills on English reading levels. The findings indicated that, in the younger cohorts of students who are more likely to have less reading experience, the influence of Punjabi and Hindi measures on English was limited to word recognition. However, once these multilingual children acquire more expertise in decoding skills (i.e., in the older cohort), listening comprehension, orthographic knowledge and phonological processing in Punjabi and Hindi influenced levels in English reading comprehension.
The overall findings from this thesis were used to derive three multilingual models of Punjabi, Hindi and English and one cross-linguistic model of English reading comprehension. These models suggest that a simple view of reading could be applied to Punjabi and Hindi orthographies in a similar way to English. However, additional influences of orthographic knowledge for all three languages (Punjabi, Hindi and English) in such multi-literate learners will need to be taken into account. Additionally, the influence of first and second language skills will need to be considered when developing models of third language reading comprehension. The proposed four models that includes the additional factors are discussed in light of previous research and theories/models in the field.