Reviews of Read Naturally

Review of Kemp Study Deceives Educators

Susan Kemp conducted a study that used Read Naturally passages but not the Read Naturally strategy. It was a study of SSR (sustained silent reading). In fact, 70% of the participants in the Kemp study were above the 50th percentile in fluency. In her dissertation, Kemp states: "Another limitation [of the study] is that Read Naturally was not just used with at-risk readers. The Read Naturally program was developed to help students who are at-risk. Typically this would represent students who are scoring below the 25th percentile.* Students participating in this study did not fit the description of at-risk."

Susan Kemp conducted a study that used Read Naturally passages but not the Read Naturally strategy. It was a study of SSR (Sustained Silent Reading). Susan Kemp stated that the purpose of her study was not to evaluate Read Naturally. Because the Kemp study did not intend to evaluate Read Naturally, its results cannot be interpreted as evidence of the program's effectiveness.

Notable discrepancies between the Kemp study and the intended use of Read Naturally include the following:

• Most students in the Kemp study were above the 50th percentile in reading fluency. Of the 157 students in the study, only 47 were at or below the 50th percentile in oral reading fluency as determined by the Hasbrouck/Tindal oral reading fluency norms. The other 110 students—70 percent of the study participants—were above the 50th percentile in fluency. Read Naturally is designed for students who score 10 or more words below the 50th percentile in fluency. Seventy percent of those students studied were not a group for whom the Read Naturally intervention is intended.
• Susan Kemp acknowledges that the study participants did not need Read Naturally. Kemp states on page 74 of her dissertation that Read Naturally is intended for at-risk readers, which is not the population she studied:
• Kemp states: “Another limitation [of the study] is that Read Naturally was not just used with at-risk readers. The Read Naturally program was developed to help students who are at-risk. Typically this would represent students who are scoring well below the 25th percentile.* Students participating in this study did not fit the description of at-risk. Many scored well above the 25th percentile. Perhaps, if this study was only completed with at-risk students, results would have been different. In addition, because two of the participating schools were California Distinguished schools, the quality of instruction of the core program could have influenced results as well.”
• The critical pass step was not implemented correctly. Read Naturally recommends a teacher-to-student ratio of 1 to 6. Kemp’s study had a teacher-to-student ratio of 1 to 20. Consequently, the pass step of the Read Naturally process could not be implemented with fidelity. Holding the students accountable at the pass step is critical for student success in Read Naturally.

In addition, the Kemp study is flawed as a review of the Read Naturally strategy due to the following issues:

• Special Education students were excluded from the study. These students, a target demographic for Read Naturally, would have likely performed poorly in SSR and performed well in Read Naturally.
• Disparate groups were compared. When comparing the Kemp study’s Read Naturally group with the SSR group, the mean pretest scores in fluency for the native English-speaking students in the “poor readers” category were substantially different. The mean score for native English-speaking “poor readers” in the Read Naturally group was 40.5 WCPM, and the mean score for native English-speaking “poor readers” in the SSR group was 75.2 WCPM. Nonetheless, these disparate groups were compared.

In spite of the Kemp study's design flaws, the group using Read Naturally passages did make significant gains in fluency. These students averaged a gain of 1.5 WCPM (words correct per minute) each week. The Hasbrouck/Tindal data shows that the average fluency gain per week for third grade students is 1.1 WCPM. This data indicates that the students in the Read Naturally group were accelerating their progress.

Nonetheless, because the Kemp study did not follow the Read Naturally strategy and was not intended as an evaluation of Read Naturally, WWC should not have reviewed it as though it was. Any conclusions drawn about Read Naturally's effectiveness on the basis of this study are therefore inaccurate.

Furthermore, WWC combines the Kemp study and the Denton study in the same ELL analysis. The Denton study was designed to evaluate Read Well, not Read Naturally. This study used phonemic decoding, which is not part of the Read Naturally strategy, as an assessment measure. The Denton study did not use fluency as an assessment measure, even though Read Naturally is a fluency-building program.

*Read Naturally Strategy programs are intended for students who score 10 or more words below the 50th percentile in fluency.