Spencer, M., & Wagner, R. K. (2018). The comprehension problems of children with poor reading comprehension despite adequate decoding: A meta-analysis. Review of Educational Research, 88, 366–400.
Students with specific reading comprehension deficits (SCD) exhibit adequate decoding skills but marked difficulties in comprehending text. Previous research indicates that SCD affects approximately 8% of students ages 9 to 14, which amounts to 2 students in a class of 24. Deficits in reading comprehension put students at risk for overall academic failure due to the critical role of text reading in content-area learning as students advance through the upper-elementary grades and beyond. The source of reading comprehension difficulty may lie in a number of cognitive skills required to make meaning of text, such as memory, inferencing, background knowledge, vocabulary, executive function, and oral language. The purpose of this study was to synthesize the results of research on SCD to advance the field’s understanding of the underlying difficulties students with SCD may have in order to improve identification and intervention.
Before synthesizing the research on SCD, Spencer and Wagner (2018) present a brief overview of four theories of reading comprehension:
One unresolved issue in the study of SCD is the reason for the reading comprehension difficulties that students experience. Previous research has differed in concluding that the source of the reading comprehension difficulty is a developmental delay in reading skills or a developmental deficit in the progress of skill growth. If students with SCD are experiencing a delay, they would be on the same developmental trajectory of reading skill growth as typical readers but progressing at a slower rate. If instead they are experiencing a developmental deficit, they would be on a different trajectory. Spencer and Wagner matched students with SCD with a younger group of students with similar levels of comprehension skills to explore whether the patterns of differences fit a developmental delay or developmental deficit pattern.
Another issue Spencer and Wagner confronted is the definition of SCD. Researchers in this area have generally agreed that SCD is characterized by a discrepancy between decoding skills and comprehension skills. Where they have differed is in the level of decoding skill students must have to be characterized as having SCD. Some researchers have identified students as having SCD based on any discrepancy while others have added requirements that the students have decoding skills in the average range, that they have a discrepancy based on their chronological age, or that their comprehension skills fall below a criterion score.
Spencer and Wagner chose to conduct a meta-analysis to address their research questions about SCD. Meta-analysis is a method for synthesizing the findings of multiple studies of a particular topic. By performing a “study of studies,” a meta-analysis provides a systematic way of bringing together data across many prior studies to test hypotheses. In this meta-analysis, Spencer and Wagner tested the following hypotheses:
These hypotheses were explored using sets of studies that have compared students with SCD and typical readers or younger students with similar reading comprehension skills. In addition, Spencer and Wagner also determined if results differed based on student age, type of oral language measure used, or type of publication (which was included to determine if results that were published in journals differed from results that were not).
In their meta-analysis, Spencer and Wagner included 86 studies conducted between 1970 and 2016. All studies met the following criteria:
The results of the meta-analysis are reported using a metric known as an effect size. Effect sizes allow for synthesis of results across studies by reporting results on the same scale, which is the number of standard deviation units (SDs) that separate groups of students.
Results Comparing Students With SCD to Typical Readers
Across studies between typical readers and students with SCD, the difference in reading comprehension scores was 2.78 SDs. However, the difference in scores on measures of oral language, 0.78 SDs, was considerably smaller. The difference in scores for verbal working memory, 0.77 SDs, was similar in magnitude. These results indicate, as expected, that students with SCD are more like students with typical reading skills in their language skills but show a marked deficit in reading comprehension. Age was a significant moderator of the difference in effect size between students with SCD and typical readers, with studies of younger students showing larger differences between groups than studies involving older students. No differences were found based on the type of oral language measure used or the publication type.
Results Comparing Students With SCD to Younger Students With Similar Comprehension Skills
In their second analysis, Spencer and Wagner synthesized the results of studies in which students with SCD were matched with younger students who had reading comprehension skills that were at a similar level. They refer to these younger students as “comprehension age-matched readers.” On average they were two years younger than the students with SCD. The difference in comprehension scores between the groups was not statistically significant, meaning it can be seen as not different from zero. The comparison of the two groups’ oral language skills also was not statistically significant. Based on these results, Spencer and Wagner concluded that students with SCD have a developmental delay in their oral language skills, which are more like the skills of younger students with similar comprehension abilities than like their same-age peers. No differences between studies were found for type of oral language measure used.
Results From Within-Student Comparison of Comprehension and Oral Language Skills
Spencer and Wagner also computed an average effect size for the difference between the reading comprehension scores and the oral language scores of students with SCD. The average difference was 0.84 SDs, showing a significant deficit in reading comprehension relative to oral language in students with SCD. Study type and type of oral language measure used were significant predictors of the magnitude of the difference between comprehension and oral language scores.
Based on the results of their study, Spencer and Wagner concluded that the data support their third hypothesis, that students with SCD have larger deficits in comprehension than in oral language but have some impairment in both domains. Additionally, they concluded that the nature of the reading comprehension difficulties of students with SCD are consistent with a developmental delay rather than a developmental deficit. In other words, these students are on a similar trajectory of growth in reading comprehension skills as typical readers but are growing at a slower rate. They pointed to a lack of sufficient background knowledge to enable comprehension and less automaticity in decoding leading to fewer cognitive resources being available for comprehension as possible explanations for the difficulties experienced by students with SCD. They called for future research examining decoding fluency and background knowledge in these students.
Spencer and Wagner noted several important implications of their results for teachers and other professionals who work with students with SCD: