Research: Word Identification Strategy Study
All students entering ninth grade in two high schools who earned decoding scores that were two or more grade levels below the 9th-grade level participated in this study (Deshler, Schumaker & Woodruff, 2002). The identified students in the experimental school (N = 62) received instruction in the Word Identification Strategy, a learning strategy for decoding multi-syllabic words (Lenz & Hughes, 1990; Lenz, Schumaker, Deshler, & Beals, 1984), for four to eight weeks (depending on the time required for each student to master the strategy) in groups of four to six students. They were taken out of their English classes in order to receive the instruction. Students in the comparison school, who were matched to the students in the experimental school by age, ethnicity, gender, and pretest decoding scores, received traditional reading instruction in their English classes.
Posttest results revealed that students in the experimental school, including those with LD, had gained
an average of 3.4 grade levels in reading decoding skills. All of the experimental students had gained at
least one grade level in decoding skills. Matched students in the comparison school made an average gain of
.2 of a grade level. On average, the students in the experimental school were decoding three years below
grade level during the pretest (at the 5.9 grade level) and at grade level (at the 9.3 grade level) during
the posttest. Students in the comparison school were decoding at the 6th -grade level during the pretest
(6.1) and also during the posttest (6.3). (See Figures 1 and 2.) The posttest decoding raw scores for the
two groups on the Slossen Oral Reading Test (Slossen & Nicholson, 1992) were found to be statistically
different using an ANCOVA, F (1,121) = 31.078, p <.001, 
2 = .692.
This is considered to be a large effect size according to Cohen's descriptions (1988).
Similar results were achieved with the students with learning disabilities (LD) within the larger groups.
On average, the students with LD in the experimental school were decoding four years below grade level during
the pretest (at the 4.9 grade level) and close to grade level (at the 8.8 grade level) during the posttest.
Students with LD in the comparison school were decoding at the high 5th -grade level during the pretest (5.9)
and at the low sixth-grade level during the posttest (6.3). (See Figure 3.) The posttest decoding raw scores
for the two groups on the Slossen Oral Reading Test (Slossen & Nicholson, 1992) were found to be
statistically different using an ANCOVA, F (1,19) = 29.673, MSE = 43.94, p <.001, 2 = .610. This is considered to be a large effect size according to Cohen's descriptions (1988).
Figures 4 and 5 display the results for male and female students, respectively. Each graph also displays the results for African Americans, Hispanics, and Caucasians. The results in both figures show that, regardless of gender and ethnicity, students in the experimental group made large gains and that students in the comparison school did not.
This same Word Identification Strategy course has been taught to ninth graders in the same high school since 1995. Figure 6 shows the outcomes of the course for eight years for all students. (1998-99 is the year of the comparison-group study described above.) The results show that substantial gains have been made each year by students taking the course. Figure 7 shows similar results for students with LD across eight years. During each year, student scores have indicated substantial gains in decoding skills.
References
Deshler, Schumaker, & Woodruff (2002)
Lenz & Hughes, 1990
Lenz, Schumaker, Deshler, & Beals, 1984
Slossen & Nicholson, 1992
Cohen (1988)
