Reading
Nonparametric statistical techniques were used to analyze the effects
of the training on both reading measures (i.e., Woodcock Word Identification
Subtest and the posttest only phonetically regular word list). The
means and standard deviations for these posttests are also found in Table
IV. To evaluate group differences on the Woodcock posttest, scores were
categorized according to our pretreatment sample selection criteria. That
is, children with raw scores of 3 or less were considered to be nonreaders
and children reading four or more words on the WRMT were considered
to be readers. (Prior to the intervention, children reading four or more
words on the WRMT were eliminated from the initial subject pool because
they were considered to be readers.) It should also be noted that prior to
the intervention only children in the original sample of 90 were able to
read one, two, or three words on the WRMT, and all seven children were
distributed through random assignment to the two control groups. All
other children had scores of 0 on the Woodcock Word Identification pretest.
Table VI shows the breakdown of posttreatment Woodcock scores for
each treatment group using this a priori cut-off decision. Differences
among the groups on this measure were significant, ×2 (2) = 8.4, p = .015.
It is of clinical interest that over 34 percent of the treatment group were
able to read four or more words on the WRMT after the segmentation intervention,
as opposed to 13 percent for control group I (language activities
group), and only 7 percent for control group II (no intervention).