PROJECT OVERVIEW: Background: The National Research Council report, Preventing Reading Difficulties in Young Children (Snow, Burns, & Griffin, 1998), urged researchers to better understand how software could benefit literacy learning, stating, "To date, a great deal of educational software design is commercial art rather than instructional science: it needs to be both." Our project took up this challenge. We assembled of team of learning scientists, computer and graphic artists, and commercial artists with the goal of designing and scientifically evaluating scalable technology tools for beginning readers that would help teachers achieve a balanced reading program and help parents support beginning readers at home.
Purpose: The primary purpose of the research studies in this project was to co-create and evaluate with teachers different models of software and video support for beginning readers in grades one and two.
Intervention: Our Take Me To Your Readers software contains 43 units (each with a book and three software sessions, each unit focusing on approximately 20 words). Our video series includes a set of 12 animated videos (four each for grades K 2) that provide instructional information for parents and children in a fun context, organized around these themes: Reading Books, Talking about Books, Building Vocabulary and Knowledge, and Cracking the Code. The prototype Palm Pilot tool added in the final year of the project allows children to type in an unfamiliar word and simultaneously see and hear the word's graphemes and phonemes.
Setting: Urban schools in a large Southeastern city, from 1999-2003.
Research Design: For our studies, we used first and second grade students in inner-city schools in a large Southeastern city. The project started September 1, 1999 and ended August 31, 2003:
In Year 1, 133 first graders and 36 second graders participated in with-in and between-subject experimental studies to evaluate and refine our software design.
In Year 2, 162 first graders and 91 second graders participated in within-class, between-subjects experiments to evaluate program effectiveness outside of teacher effects and whole-class implementations to examine the impact of teacher support. This pointed us toward refining our assessments and emphasized the importance of whole-class implementations to best integrate the program into classroom use.
In Year 3, 266 first graders and 91 second graders participated in between-class, between-subjects experiments to evaluate program effectiveness in natural classroom settings. This provided evidence for program effectiveness on studied words, directedus to adjust usage and add materials in order to aim for more transfer in word learning, and narrowed our focus to first grade.
In Year 4, 241 first graders participated in between-class, between-subjects experiments to replicate and extend the implementation model at grade one. This enabled development and exploration of Palm Pilot support in addition to computer support materials.
Students in experimental conditions used our materials, students in comparison conditions used other software chosen for the study, and students in control conditions used only their normal classroom materials.
We measured outcomes using standardized reading assessments (aligned with Fletcher and Mathes' IERI project) and assessments with computer-generated word lists for each student, based on a sample of words that each student had practiced in the software. We yoked experimental students to control students who received the same lists. We also evaluated the implementation using teacher observations, teacher surveys, and parent surveys developed during the project.
Findings: Overall the results from our work validated that although the word-reading practice in our software helped children learn to read the words in our program, the prime developmental window for this type of practice is likely earlier and shorter than the entire first grade of year of use that we tested. We suspect that optimal benefits for individual word practice like our program provides may occur beginning in late kindergarten and may end by the middle of first grade, at which time most children would likely benefit more from reading connected text than from continued practice on activities like ours. Fueled by these findings, our initial development of Palm Pilot technology support for connected text reading in first grade suggested a promising avenue for future development and research.
Our work also highlighted important aspects of children's decoding ability. Results from studies in Years 1 and 2 imply that word recognition fluency is a multi dimensional construct. Children can exhibit "receptive fluency" or the ability to match a printed word to a spoken one before they exhibit "productive fluency," the ability to look at a printed word and produce the spoken equivalent. Many typical software activities require only receptive fluency. Teachers are often surprised at how well the students do on computer activities yet how poorly they do when actually asked to read from a book. Our results suggest that it is important to construct software activities that promote productive fluency and to give teachers better information about children's productive fluency. They also underscore the importance of ensuring that software-based assessments are not misleading. Finally, these results may have implications for how cognitive theorists describe the relationship between mental representations of words and their access during different word reading tasks.
PROJECT PUBLICATIONS: Sharp, D., & The Cognition and Technology Group at Vanderbilt. (in press, 2006). From mental models to lexicons: The case of early literacy software. In Kinzer, C., & Verhoven, L. (Eds.), Interactive Literacy Education: Facilitating literacy environments through technology. NJ: Erlbaum.