Individual differences in reading aloud: A mega-study,item effects,and some models

Normal individual differences are rarely considered in the modelling of visual word recognition – with item response time effects and neuropsychological disorders being given more emphasis – but such individual differences can inform and test accounts of the processes of reading. We thus had 100 participants read aloud words selected to assess theoretically important item response time effects on an individual basis. Using two major models of reading aloud – DRC and CDP+ – we estimated numerical parameters to best model each individual’s response times to see if this would allow the models to capture the effects, individual differences in them and the correlations among these individual differences. It did not. We therefore created an alternative model, the DRC-FC, which successfully captured more of the correlations among individual differences, by modifying the locus of the frequency effect. Overall, our analyses indicate that (i) even after accounting for individual differences in general speed, several other individual difference in reading remain significant; and (ii) these individual differences provide critical tests of models of reading aloud. The database thus offers a set of important constraints for future modelling of visual word recognition, and is a step towards integrating such models with other knowledge about individual differences in reading.     

Understanding reading as a form of language-use: A language game hypothesis

Reading research and research on conversation have followed different paths: While the research program for reading committed itself to a relatively static view of language, where objective text properties serve to elicit specific effects on cognition and behavior of a reader, research on conversation has embraced a language-use perspective, where language is primarily seen as a dynamic, context dependent process. In this essay I contrast these two perspectives, and argue that in order to reach a unified understanding of natural language – be it reading, talking, or conversing – one needs to adopt a language-use perspective. Furthermore, I describe how reading can be seen as a form of language-use, and how the current landscape of research on reading can be re-interpreted in terms of a dynamic, context-sensitive perspective on language. In particular, I propose that the concept of ‘language games’ serves as a good starting point to conceive reading as a form of language-use, describe how one can derive first concrete hypotheses by re-interpreting reading in terms of language games, and show how they can be readily operationalized using tools from dynamic systems analysis.     

A Computational and Empirical Investigation of Graphemes in Reading

It is often assumed that graphemes are a crucial level of orthographic representation above letters. Current connectionist models of reading, however, do not address how the mapping from letters to graphemes is learned. One major challenge for computational modeling is therefore developing a model that learns this mapping and can assign the graphemes to linguistically meaningful categories such as the onset, vowel, and coda of a syllable. Here, we present a model that learns to do this in English for strings of any letter length and any number of syllables. The model is evaluated on error rates and further validated on the results of a behavioral experiment designed to examine ambiguities in the processing of graphemes. The results show that the model (a) chooses graphemes from letter strings with a high level of accuracy, even when trained on only a small portion of the English lexicon; (b) chooses a similar set of graphemes as people do in situations where different graphemes can potentially be selected; (c) predicts orthographic effects on segmentation which are found in human data; and (d) can be readily integrated into a full‐blown model of multi‐syllabic reading aloud such as CDP++ (Perry, Ziegler, & Zorzi, 2010). Altogether, these results suggest that the model provides a plausible hypothesis for the kind of computations that underlie the use of graphemes in skilled reading.     

Remembering,understanding, and representation

Starting with the facts that not everything that is understood is remembered, and that not everything that is remembered is understood. this paper urges that models of language processing should be able to make a distinction between comprehension and memory. To this end. a case is made for a spreading activation process as being the essential ingredient of the comprehension process. It is argued that concepts activated during comprehension not only restrict the search set for candidate concepts to be used in a top-down fashion, they also constitute part of an episodic representation that can come to be p e of long-term memory. The way in which these representations atrophy is discussed, as is the way in which their idiosyncratic components are eliminated in producing representations in semantic memory. Some observations on the comprehension and memory of text are made and arguments are presented to show how intrusions and omissions in recall can be handled. Some existing experimental data is reanalyzed in terms of the proposed model and alternative interpretations consistent with the model are shown to be possible.     

Highlights as an Early Predictor of Student Comprehension and Interests

When engaging with a textbook, students are inclined to highlight key content. Although students believe that highlighting and subsequent review of the highlights will further their educational goals, the psychological literature provides little evidence of benefits. Nonetheless, a student’s choice of text for highlighting may serve as a window into her mental state—her level of comprehension, grasp of the key ideas, reading goals, and so on. We explore this hypothesis via an experiment in which 400 participants read three sections from a college-level biology text, briefly reviewed the text, and then took a quiz on the material. During initial reading, participants were able to highlight words, phrases, and sentences, and these highlights were displayed along with the complete text during the subsequent review. Consistent with past research, the amount of highlighted material is unrelated to quiz performance. Nonetheless, highlighting patterns may allow us to infer reader comprehension and interests. Using multiple representations of the highlighting patterns, we built probabilistic models to predict quiz performance and matrix factorization models to predict what content would be highlighted in one passage from highlights in other passages. We find that quiz score prediction accuracy reliably improves with the inclusion of highlighting data (by about 1%–2%), both for held-out students and for held-out student questions (i.e., questions selected randomly for each student), but not for held-out questions. Furthermore, an individual’s highlighting pattern is informative of what she highlights elsewhere. Our long-term goal is to design digital textbooks that serve not only as conduits of information into the reader’s mind but also allow us to draw inferences about the reader at a point where interventions may increase the effectiveness of the material.