The Real-World Reading and Writing U.S. Children Need

Abstract

The 2001 Progress in International Reading Literacy (PIRLS) study--a study of the reading achievement of mostly 10-year-old students in 35 nations--found that U.S. students had the largest gap between literary reading achievement and informational reading achievement of any nation studied (Mullis et al. 2003), with the gap favoring literary reading. In the 2006 study, this gap narrowed somewhat, but average scores of U.S. students were statistically unchanged (Mullis et al. 2007). What causes this? In this age, why are U.S. students relatively poor at reading to acquire and use real-world information (informational texts) when compared to reading to experience imagined worlds (literary texts)? Some might say that informational reading is inherently more difficult than literary reading, but many nations had comparable literary and informational reading achievement, and some even had stronger informational achievement, so that explanation seems insufficient. Others might say that it's the nature of the particular reading assessment used in the PIRLS study, but studies using other assessments have also shown that U.S. students struggle with comprehension of informational text as compared to narrative or literary reading. A more plausible explanation is that U.S. students have little experience with informational text in the early years of schooling. In fact, they get very little. In a study I conducted in U.S. 1st-grade classrooms (Duke 2000), I found that only 9.8% of the books and other materials in classroom libraries and only 2.6% of the materials on classroom walls and other surfaces were informational text. More important, important, only 3.6 minutes of the classroom day, on average, were devoted to informational text--only 1.9 minutes in classrooms in low-socioeconomic-status settings. Studies that have looked at this issue in other ways--for example, by examining basal reading programs--have also found a relative lack of informational text (e.g., Moss and Newton 2002). When I tell others about this finding, a common response is that the focus of primary grade education should be on developing decoding and fluency, not comprehension and certainly not comprehension of informational text. However, we found that 1st graders whose teachers provided greater experience with informational text had decoding and spelling skills comparable to those of other students (Duke et al. 2009). Another response is that informational text is simply too hard, perhaps even developmentally inappropriate, for young children. This stance has long been noted, and countered, in research (Duke and Kays 1998; Newkirk 1987; Pappas 1993). Young children can be successful with informational text--in fact, some studies in this area (Maduram 2000; Tower 2002) have included three-and four-year-old children! There are some indications that attention to informational text in early schooling is increasing. For example, in 2008, Moss found more informational text in basal reading programs than in the past. This is encouraging-informational reading ability is arguably more important now than it ever has been. Reflecting this, the 2009 National Assessment of Educational Progress (NAEP) Framework calls for 50% informational passages by 4th grade, 55% by 8th grade, and 70% by 12th grade (National Assessment Governing Board 2007). In sum, there is little question about whether we should involve informational text in early schooling; the question now is how. THE USUAL SUSPECTS So far, most instructional practices shown to help develop young children's comprehension of informational text come as no surprise. For example, just as teaching young children the structure of stories has long been shown to improve their comprehension of stories, so, too, does teaching young children the structures of informational text (specifically, compare-contrast and cause-effect) improve their comprehension of these texts (Williams 2005, with 2nd graders). …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 26 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.