Method
Participants
This study took place in eight Metropolitan Nashville schools participating in a large-scale investigation of the First-Grade PALS reading program from October, 2000, to April, 2001, (Fuchs, Fuchs, Yen et al., 2001). Four of the schools were high-poverty "Title I" schools, and four were middle-class "Non-Title I" schools. Thirty-three first-grade teachers volunteered to participate. The teachers were stratified by school type (Title I or Non-Title I) and assigned randomly within school to one of three conditions. In the large-scale study, 11 classrooms used a standard version of the PALS program (Standard PALS), 11 classrooms used a fluency-building version (PALS + Fluency), and 11 classrooms served as "no-treatment" controls. Pre- and posttreatment data were available for 496 students (168 in Standard PALS classes, 155 in PALS + Fluency classes, and 173 in control classes). The control classrooms did not participate in the nonresponder portion of the study. A three-step process guided selection of the nonresponders from the 22 Standard PALS and PALS + Fluency classrooms:
- selecting students at risk for unresponsiveness to PALS,
- monitoring the at-risk students' progress, and
- identifying nonresponders among the at-risk group.
Selecting the risk pool. In October, 2000, written parental consent was obtained for students to participate in the large-scale study. These students were given a Rapid Letter Naming test, an effective predictor of future reading achievement (Torgesen et al., 1997). Within each classroom, students' were rank-ordered from highest- to lowest-performing based on their Rapid Letter Naming scores. Adjustments to these rankings were made based on teacher judgment. For example, if a student's Rapid Letter Naming score was low, but the teacher believed the student to be an average reader, the teacher's judgment overrode the Rapid Letter Naming ranking. Based on the adjusted rankings, 4 average-performing and 8 lowest-performing students were identified in each classroom. The 8 lowest-performing students from each of the 22 Standard PALS and PALS + Fluency classrooms were considered "at risk" to be unresponsive to the PALS program (n = 176). The 4 average-performing students per class served as a comparison group (n = 88).
Monitoring progress. From October to December, the at-risk and average-performing students' reading progress was monitored weekly. Monitoring measures included "chapter tests," which were criterion-referenced measures of students' progress in PALS, and two word-level reading CBM measures. These were Nonword Fluency probes from the Dynamic Indicators of Basic Early Literacy Skills (DIBELS; Good & Kaminski, 2001), and Dolch word probes consisting of high-frequency pre-primer, primer, and first-grade-level words.
Identifying nonresponders. After seven weeks of PALS participation, complete monitoring data were available for 166 at-risk and 87 average students (representing an attrition rate of 6% and 1% for at-risk and average students, respectively). At this point, nonresponders were identified in accordance with a five-step process. First, the CBM levels and growth rates (i.e., slopes) of each of the at-risk and average students were calculated. Level indicated the number of correct words per min the student read at the end of the monitoring period. It was calculated as the mean of each student's last two scores. Slope, by contrast, indicated how many more correct words per min students read each time they were monitored. For example, a slope of 1 indicated a growth rate of one word read correctly per min for each monitoring session. It was calculated using a least squares regression between monitoring scores and calendar days.
Next, means and standard deviations (SD) of the average-performing students' CBM levels and slopes were calculated. Third, all students from each classroom were identified who had scored 90% or less on the last chapter test, or who were the lowest-scoring in their class on this measure (n = 97). Fourth, z-scores were calculated separately on the CBM levels and slopes of these 97 students, using the means and SDs of the average-performing students. Finally, we examined the z-scores of each and every low-performing student. Students were identified as nonresponders if they scored more than 0.5 SD below the average readers in terms of both level and slope on the CBM measures (n= 66).
Assigning nonresponders to groups. Within the Standard PALS and PALS + Fluency classes, the 66 nonresponders were stratified in terms of "low" (n = 28) vs. "very low" (n = 38) status based on CBM levels and slopes. Then, they were assigned randomly to one of three treatments: PALS, Modified PALS, or Tutoring. Each of the three treatments was nested within the PALS and PALS + Fluency treatments, such that Modified PALS and Tutoring students received either a Standard or Fluency version of these programs, depending on their initial PALS condition.
Twenty-two students were assigned randomly to each of the three treatments. Due to attrition, there were 21, 15, and 20 students in the PALS, Modified PALS, and Tutoring treatments, respectively, for a total of 56 nonresponders, at the end of the study. Analyses of variance (ANOVAs) showed no statistically significant differences across the three groups in terms of Dolch level (F [2, 53] = 0.37, p = 0.70), Dolch slope (F [2, 53] = 1.09, p = 0.34), Nonword Fluency level (F [2, 53] = 0.24, p = .79), and Nonword Fluency slope (F [2, 53] = 0.94, p = 0.40) before Modified PALS and Tutoring began. Students in the three conditions were also compared on several demographic variables, including sex, ethnicity, English Language Learner (ELL) status, socioeconomic status, and special education status. Chi-square tests indicated no statistically significant differences across conditions on these variables.
Previous Page | Next Page
(Introduction) | (Method: Treatments)
