Group Two: Studies without Comparison Groups

Sugarman and Howard (2002) studied eleven two-way programs, and reported that Spanish speakers in these programs improved on different tests of English from grade three to five. There was, however, no comparison group and the measures used were not standardized. Thus, there was no way to compare the two-way childrens’ progress to children in other programs. Sugarman and Howard’s results, however, provide a positive answer to the first of the three questions posed in the introduction; they show that children in two-way programs make progress in acquiring English.

Several studies provide data using standardized tests without comparison groups. Before proceeding with a discussion of these studies, it should be noted that results in these studies are reported as NCE (Normal Curve Equivalent) scores. NCE scores are very useful when comparing across studies: they are constructed so that the mean is 50 and the standard deviation is 21.06.

NCE scores are not the same as ordinary percentiles, except at 1, 50, and 99. Of special interest is the fact that when scores are below the 50th percentile, NCE scores are higher than percentiles: A percentile rank of 29, for example, is equalivant to an NCE score of 38. Above the 50th percentile, NCE scores are lower than percentiles. A percentile rank of 76 is equivalent to an NCE score of about 65. Thus, in discussing the studies to follow, both NCE and percentiles will be included. (I thank Jim Crawford for pointing out the necessity of doing this.)

de Jong (2002) reported that different cohorts of Spanish-dominant children in a two-way program consistently attained NCE (Normal Curve Equivalent) scores between 38 (29th percentile) and 43 (37th percentile) on the Stanford Reading Test in grade five between 1995 and 2000. Sample sizes were small, however, ranging from 13 to 18 students. A possible problem with this study is that “Spanish-dominant” does not necessarily mean limited English proficient; no information was provided on subjects’ initial English proficiency.

In Lindholm (2002), on tests of English reading, English learners in three two- way programs achieved NCE scores of 34 (23rd percentile), 35 (24th percentile) and 39 (30th percentile) by grade six, with one group achieving an NCE score of 43 (37th percentile) by grade seven (see table 10.6, p. 217). (One group, however, attained only the 9th percentile in grade five, see discussion below.) A subgroup studied longitudinally (n = 149) attained an NCE score of 41 in grade seven (34th percentile). Because few subjects were classified as orally proficient in English in kindergarten (Lindholm, 2002; p. 190) we can be confident that most subjects in this group were in fact limited in English when they began the program (see footnote 1 for studies in which this was not the case).

2