The Rationale for and Sample of the First-Grade Text Complexity Project

• In Event: 70.021 - Multiple Perspectives on the Complexity of First-Grade Texts Over the Past 50 Years

Abstract

PURPOSE: The purpose of this presentation is to describe: (a) the problem that the session addresses, (b) the rationale for selecting the texts and six text analysis and (c) the features of the shared dataset used in the 6 text analyses.

PERSPECTIVE: A critical educational policy, which is the case with increased expectations for first-grade reading levels, requires a strong evidence based. This project on first-grade text complexity addresses the assumption of decelerated text complexity at grade 1 from multiple perspectives. By corroborating findings from multiple systems—each of which examines unique aspects of text complexity— policy-makers and practitioners can have comprehensive evidence in making decisions about and selections of text.

METHODS: A first choice in designing the project was to focus on quantitative systems. The CCSS proposes a tripartite system for analyzing text complexity that includes qualitative and reader-task as well as quantitative measures. The few qualitative and reader-task systems that are available (e.g., Fountas & Pinnell, 2012) have yet to provide data on the validity or the reliability of the system. Consequently, the project focused on the currently available quantitative systems.

The second step was to invite individuals who have developed and/or use unique text analytic systems to analyze a shared dataset. Invitations and information were communicated to individuals without knowledge of other participating teams. Each of the six analytic approaches was applied to the dataset independently.

DATA SOURCES: Text analytic systems were identified through reviews of recent studies of text complexity (e.g., Nelson, Perfetti, Liben, & Liben, 2012). Further information on the identification of text analytic systems will be included in the presentation but, for purposes of blind review, will not be described in this proposal.

The cost of producing core-reading programs has meant that only three core-reading programs are available in 2014, down from a high of approximately 12 programs in the early 1980s (Tyson-Bernstein, 1988): Scott Foresman/Pearson, McGraw-Hill, and Harcourt Houghton Mifflin. For this phase of the project, the Scott Foresman/Pearson program was chosen because it was one of two mainstream reading programs in Chall’s (1967) seminal analysis of first-grade texts. The other mainstream program, Ginn, was purchased by Pearson in 1998 and was incorporated into the Scott Foresman program.

The dataset consisted of all Scott Foresman first-grade core-reading texts from 7 editions: 1962, 1971, 1983, 1993, 2000, 2007, and 2013. This period includes the 50 years during which the CCSS cited a decline in text complexity and the four years subsequent to the publication of the Standards in 2010 (the sub-title of the 2013 Scott Foresman program is Common Core edition). The number of texts and words for each of the 7 editions is given in Table 1.

SCIENTIFIC SIGNFICANCE: Both the methods of this project (i.e., a shared dataset analyzed independently by experts in available text analytic systems) and its focus on a critical educational policy will add considerably to the literature on text complexity.

Authors

• Elfrieda H. Hiebert, TextProject

• Leigh Ann Martin, University of Michigan