skewness, was not extensive.
What is the weekly rate of student progress and, to what extent, does it vary with grade level? A significant relation between grade and CBM spelling slopes (see ANOVA results in footnotes to Tables 3 and 4) existed for both years, for the graded and common measures, and for the letter sequences and correct words scores. In all but one case (i.e., the Year 1 correct words score for the common measure), both the linear and the quadratic terms contributed significantly to the modeling of the relationship between slope and grade level.
For the primary CBM spelling index (i.e., LS scores on the graded tests), slopes ranged from .21 in Grade 6 to 1.08 in Grade 2 during Year 1; from .46 in Grade 6 to .92 in Grade 2 during Year 2. Follow-up tests to the ANOVA indicated that, for Year 1, the slope of growth was comparable at Grades 5 and 6, but that the Grade 4 slope was reliably greater than slopes at Grades 5 and 6. In a similar way, the slope at Grade 3 was reliably greater than at all subsequent grades, and the Grade 2 slope also was reliably greater than at every higher grade. For Year 2, reliable differences were identical to those found at Year 1, with the exception that the Grades 5 and 6 slopes also were statistically significantly different.
As these findings reveal, LS slope on the graded measure decreased reliably and geometrically with grade level, indicating that students make their most dramatic spelling growth in the early grades and that, although their growth continues, the amount of that growth trails off steadily over time. This inverse relation between LS slope and grade level for the graded CBM spelling test was also evident in the pattern of slopes for the common CBM spelling test (sampling words across Grades 1-6). Corroborating information across years and across types of CBM spelling tests increases confidence in reliability of findings.
Interestingly, the inverse relation between LS slope and grade level also is consonant with theories of developmental spelling growth. These theories propose that children advance through discernable stages of spelling (Henderson, 1980). Most children proceed relatively quickly through the first three stages (i.e., nonphonetic, preconventional phonetic, and conventional phonetic stages) and enter the final stage (i.e., morphemic spelling) by third grade (Beers & Henderson, 1977; Morris, Nelson, & Perney, 1986). This theory suggests that progress occurs relatively quickly during the first 2 to 3 years and then, as students enter the morphemic spelling stage, progress slows. This proposed pattern is mirrored in the pattern of CBM spelling slopes shown in Tables 3 and 4.
In terms of expectations for weekly rates of improvement, at least when monitoring progress on the Harris-Jacobson graded word lists, it appears that realistic and ambitious targets are approximately 1 word and 1.5 words per week at Grade 2; .65 and 1.0 word at Grade 3; .45 and .85 word at Grade 4; .3 and .65 word at Grades 5 and 6.
On a different note, findings reveal a pattern of higher slopes for the LS than for the words score. LS award credit for parts of words spelling correctly. These higher slopes for scores that award partial credit should make LS a more sensitive index of growth, which detects student improvement sooner than would scores requiring students to improve in units of whole words correct (see Deno, 1985 for discussion). Of course, because spelling words, rather than parts of words, is valued in society, it may be important to monitor number of words spelled correctly