Kees P. van den Bos

Effects of Listening Comprehension Training on Listening and Reading

In this study the effects of providing text strategy instruction in a listening mode on listening and reading comprehension of experimental and control groups of 9- to 11-year-old poor readers were examined. All students were very poor in decoding and poor in reading comprehension. In addition, half of the group were also poor listeners, whereas the other half consisted of students with normal listening test scores. The experimental program employed auditorily presented texts, and the students were trained in the strategies of clarifying, summarizing, predicting, and questioning. The programs teaching format consisted of a combination of reciprocal teaching and direct instruction. Measurements occurred according to a pretest-posttest-retention test design. The results indicated significant program effects on strategic listening and strategic reading posttests, and the results were maintained on retention tests that were administered 3 months after termination of the program. However, transfer in terms of an improved performance on more general and standardized listening and reading comprehension tests was not demonstrated.

Text comprehension strategy instruction with poor readers

The goal of this study was to investigate the effects of teachingtext comprehension strategies to children with decoding andreading comprehension problems and with a poor or normallistening ability. Two experiments are reported. Four textcomprehension strategies, viz., question generation, summarizing,clarification, and predicting were taught through directinstruction and reciprocal teaching. In both experiments, effectswere measured according to a pretest – posttest – retentiontest – control group design. Dependent variables wereexperimenter-developed strategic reading and listening tests, andstandardized reading and listening comprehension tests. In thefirst experiment the subjects were 9 to 11-year-old poor readersfrom special schools for children with learning disabilities. Inthis experiment, the intervention programs texts and strategyinstructions were presented in listening settings only. Thesubjects in the second experiment were 10-year-old children fromregular elementary schools and 9 to 11-year-old children fromspecial schools. They were also poor readers but their decodingperformance was not as poor as in the subjects in experiment 1.In experiment 2, the intervention program involved textpresentations in alternating reading and listening lessons.Although in general, normal listeners performed better on allcomprehension tests than poor listeners, there were nodifferential program effects for the two listening levels. Cleareffects of both programs were found on strategic reading andlistening tests administered directly after the interventions. Inthe first experiment, maintenance test performance showedprolonged program benefits, whereas in the second experimentthese maintenance effects were blurred by unexpected gains of thecontrol groups of students, especially from regular schools.Finally, apart from some local successes, neither of the twoexperiments offered stable evidence of transfer of comprehensionstrategy training to standardized general listening and readingcomprehension tests.

Relationship Between Cognitive Development, Decoding Skill, and Reading Comprehension in Learning-Disabled Dutch Children

About a year ago, a popular Dutch magazine (Vrij Nederland) published an article about the largest type of special school we have in the Netherlands: Schools for Children with (Specific) Learning Disability (L. D.) and Pedagogical Problems. There are about 40,000 such children in the age range of 6–13 years in these schools. This constitutes about 3.5% of the total Dutch children in that age range. The article focused on one randomlv chosen school, and students, parents, teachers, and school inspectors were interviewed. Among the many Intriguing observations was the following statement: Studying the progress reports, one is not impressed by the results . . . Compared to the test findings two years ago, his verbal intelligence has declined. . . now functions at low-average level. . . According to the head-master, the gradual IQ-decline of many children can be attributed to their inability to read. Intelligence tests for a ten-year-old presuppose a considerable amount of knowledge. This knowledge could not be acquired because they read so badly. Anyone who, as a ten-year-old, stumbles through the Sleeping Beauty, cannot do much with geography or history books. . . (Free translation)

The nonword-reading deficit of disabled readers: a developmental interpretation.

This article presents empirical evidence challenging the received wisdom that a nonword-reading deficit is a characteristic trait of disabled readers. On the basis of 2 large-scale empirical studies using the reading-level match design, we argue that a nonword-reading deficit is the consequence of normal developmental differences in word-specific knowledge between disabled readers and younger normal readers (both groups being matched on real-word reading). The first study shows that the nonword-reading deficit varies as a function of age and reading level and that this deficit is not typical for disabled readers. The second study demonstrates that a nonword-reading deficit crucially depends on the sensitivity of the matching word reading task to detect age-related differences in word-specific knowledge between disabled and normal readers. We clarify how these findings can be interpreted within the current framework of the phonological deficit hypothesis and discuss implications for theories of reading development.

Rapid naming and phonemic awareness in children with reading disabilities and/or specific language impairment: Differentiating processes?

PURPOSE The objective of this study was to assess and compare the predictive values of group membership for rapid automatized naming (RAN) and phonemic awareness (PA) in Dutch school children with and without reading disabilities (RD) or specific language impairment (SLI). METHOD A composite word reading index and a formal SLI diagnosis were used to classify a total of 1,267 children aged 8 to 13 years old either as RD-only (n = 126), SLI-only (n = 21), comorbid (RD+SLI; n = 30), or typically developing (n = 1,090). RAN and PA were assessed with 4 standardized subtests. The clinical subgroups were compared to each other and contrasted with the control group. RESULTS For each subgroup, results indicate substantial effect sizes of RAN and PA. However, the RD-only group seems to be more affected by poor RAN than the SLI-only group, whereas the 2 groups perform equally poorly on PA. The comorbid group was revealed as most severely impaired on all measurements. CONCLUSIONS In studying RD and SLI, this research indicates that it is important to distinguish between RD-only, SLI-only, and comorbid groups. The comorbid group shows additive effects of both disorders.

Item type comparisons of language comprehension tests

Several test procedures are available to measure language comprehension. In this article two tests with different item types are compared. The tests are the Dutch Reynell test - in which the subject has to carry out verbal instructions - and the BELL test - in which the subject has to select which of four pictures best represents a given statement. We studied the contribution of the type of item, and the linguistic features of the item to the prediction of item difficulty. Both tests were administered to 64 Dutch kindergarten children with an average age of 70.3 months. The simple correlation between type of item and item difficulty is relatively high (.40). However, regression analyses indicate that type of item does not contribute significantly to the prediction of item difficulty. The linguistic features of the item, in contrast, do. The two tests cover successive age ranges. The results therefore indicate that the BELL test can be used as a follow up test of the Reynell test for language comprehension.

Phonological Processing and Word Reading in Typically Developing and Reading Disabled Children: Severity Matters

In this study word reading (WR) fluency was used to dichotomously classify 1,598 Dutch children at different cutoffs, indicating (very) poor or (very) good reading performance. Analysis of variance and receiver operating characteristics were used to investigate the effects of rapid automatized naming (RAN) and phonemic awareness (PA) in predicting group membership. The highest predictive values were found for the combination of RAN and PA, particularly for the poorest readers. Furthermore, results indicate that with the severity of impairment, WR is more dominated by deficient PA, which is interpreted as an enduring problem with sublexical processing. Another main result is that with the increase of reading skill, the contribution of PA diminishes, whereas the contribution of RAN remains fairly constant for the whole reading fluency continuum. These results warrant the conclusion that whereas PA hallmarks reading disability, RAN appears to be the default predictor for above-average or excellent reading proficiency.

Rapid Naming and Phonemic Awareness in Children With or Without Reading Disabilities and/or ADHD

Employing a large sample of children from Dutch regular elementary schools, this study assessed the contributing and discriminating values of reading disability (RD) and attention-deficit/hyperactivity disorder (ADHD) to two types of phonological processing skills, phonemic awareness (PA) and rapid automatized naming (RAN). A second objective was to investigate whether comorbidity of RD and ADHD should be considered as an additive phenomenon as to RAN and PA. A total of 1,262 children, aged 8 to 13 years, were classified as RD (n = 121), ADHD (n = 17), comorbid (RD+ADHD; n = 16), or control (n = 1,108). Phonological processing was assessed by standardized tests of PA and RAN. Disability groups were compared to each other and contrasted to the control group. Although results indicate substantial effects for all three disability groups on both types of phonological processing, and the RAN/PA compound measure in particular, effect sizes were considerably larger for the RD groups, as compared to the ADHD-only group. Theoretical and practical implications are discussed.

Measuring word identification skills and related variables in Dutch children

In this chapter a simple definition of dyslexia is adopted: Ward identification ability below a certain performance criterion on a suitable wand identification test. In Study 1, the research focus is on two word identification tests, a real-word test (RWT) and a pseudoword test (PWT). The question is: Are these tests differentially sensitive to phonological and lexical ‘routes’ or subcomponents of woad identification ability? Beginning readers are supposed to use the phonological word identification component in both tests, whereas more experienced readers make use of the lexical component while reading the RWT words. Because the PWT words – except parts of them – are new, the lexical component cannot (fully) be used in the PWT. If the two subcomponents exist, a decreasing correlation between RWT and PWT scores is expected with increasing reading performance age. The results of the Study 1 show little evidence in favor of this prediction. Correlations between RWT and PWT scores remain high at elementary grade levels. It is suggested that the Dutch RWT and PWT are not differentially sensitive to different word identification component skills. In Study 2, two recently proposed aspects of underlying reading processes were investigated: Sensitivity at the level of phonological cluster discrimination, and naming speed of symbols. The results of stepwise regression analyses clearly support the predictions that beginning readers and dyslexic readers show the same predictive structure. Naming speed was the only significant predictor of word identification. On the other hand, in more experienced readers the phonological sensitivity was the main predictor variable. Therefore, it was concluded that the reading performance of dyslexic children is mainly limited by mental speed capacity, as it is the case in normal beginning readers.

Fix-Count : A Program for Computing the Error-Success Transitions in Fixed-Trial Two-Stage Markov Learning Experiments

This paper describes a FORTRAN 77 program that computes both the different response success-error patterns and their summary statistics for learning and forgetting in fixed trial experiments in which a two-stage Markov model is used. The frequency counts of the different response success-error patterns are used to estimate the parameter of the two-stage Markov model for fixed trials.