Eileen Harris wrote her dissertation on a study she conducted with 6 fourth grade, below grade level readers entitled, Evaluating the Efficacy of Reading Fluency Instruction, for the University of Southern Maine in August of 2010. She selected students who were performing between 25th and 50th percentile on the AIMSweb and MAP assessments given to students in the school. Students who qualified for other reading supports were excluded from her study.
Of note is the fact that these students would be classified as "average" and "low average" students. Students read at Lexile levels between 200L and 600L. When reading material was selected it was apparent that only quantitative measures were used, discounting the qualitative measures that are identified as equally important to identifying if readings are appropriate for a group. While the author does cite research that indicates that poor readers do better with word overlap and stronger readers do better with content overlap (Daly, Martens, et al. 1999, p. 34), she does not consider content, reader background knowledge or text features not measured in Lexile such as the use of figurative language and symbolism.
The program was delivered by paraprofessionals and consisted of five major steps. First, oral reading. Students each read aloud for one minute, rotating through the group until the passage for the day was read. During this time corrective error correction was engaged in (teacher correctly says the word, student repeats and moves on). This was followed by about 10 minutes of word study in which 6-10 words from the passage representing the six syllable types were taught and practiced in isolation. The third step was oral rereading guided practice in which the passage was again read in one minute round robin fashion. During this reread phrase drill error correction was used (teacher corrects word, student is asked to repeat word, then the student is asked to read the phrase with the error over three times correctly). The lesson ended with metacognitive feedback in which the teacher facilitated reflection on the word study and reread. Altogether the lesson took about thirty minutes.
Five of the six students the students in the study made gains which is not unpredictable based on the fact that up to 22 sessions were implemented. We do not know if they were daily or not. Even daily instruction would represent a month of school during which students were receiving their in class reading instruction as well as the intervention. One of the participants dropped out of the study because of a special education referral. Three of the six students were near the 50th percentile rate at the end of the intervention. These three were, however, the ones who had been higher achieving at the beginning of the study. The students who made the most progress are the ones who had decoding weaknesses. The phonics instruction related to syllable type helped them understand the code and thus increase their performance. Also unsurprising is the conclusion that weaker readers at the beginning made less progress. One other data skewing fact is that the student who made the most progress received afterschool tutoring. Whether the progress was due to one, the other or the combination of interventions is unknown.
While the author notes that her intervention could be used in a classroom so that students would not need to be pulled from instruction, pulling aside a group in a fourth grade that does not have reading groups, as is the case in many schools, would still be removing them from "instruction" just not from the room. If the teacher used this process instead of the regular reading instruction, it is unknown if the result would be the same. Further, implementing a program of repeated reading in a separate room with limited distractions will likely have different results that working in a classroom full of students and activities.
The author admits that what appears to be most true is that more individualization might have helped the students achieve more. The student who showed regression might have been better served if he had been pulled and given a different intervention. The student who made the most progress received additional time for instruction. This might ultimately be the largest solution. If students learn along a bell curve, some will need more time to learn material than others. Those in the below average group are fully capable of learning to read fluently, but they may require more time for high quality, one-on-one intervention.