We used a mixed factorial design with the between-subjects variables of: a) language background of the students based on teacher ratings (E.O., E.D., Eq.D., S.D., S.O.) and b) reading stage based on teacher ratings on a scale of 1 to 5, from low to high ability. The with-in subjects variables were a) Mann PST (English or Spanish adaptation); b) Yopp-Singer PST (English or Spanish adaptation); c) Peabody Picture Vocabulary Test (English or Spanish adaptation), and; d) the Woodcock Reading Mastery Tests (Word Identification or WordAttack). All subjects, whether English- or Spanish-speaking, were given both the English and Spanish versions of all the tests, with the exception of the E.O. subjects, who were administered only the English PPVT-R. The dependent variables were student performance. Half the children received first, the PPVT-R (either in Spanish or English), second, the Woodcock measures, third, the Mann PST (either the Spanish or English version), fourth, the Yopp-Singer PST (either in Spanish or English), and fifth, the PPVT-R (in whichever language was not used to administer the test previously). The other half of the children received the tests in the same order, except that the Yopp-Singer and Mann PST order was interchanged. The language of test administration, both Spanish and English, was counterbalanced for all tests.
Between February and April 1998, the participants were tested individually, either in the library storage/computer room or in a room off of the teacher's lounge at the school. Within one testing session, the students were administered the PPVT-R (in English or both Spanish and English), the Woodcock subtests (in English), and the phoneme awareness measures (in both Spanish and English). While administering the English versions of all tests, the examiner read the test items orally. In all Spanish versions of the tests, the target words were read by a native speaker of Spanish (reproduced on a tape player). A break of approximately three to five minutes was given to the children in mid-session. The one-time assessment took approximately 45 minutes to administer per child.
Table 1 presents the descriptive data on all language, vocabulary, and phonological awareness measures as well as age, language background and reading stage values. All significant effects are p < .01 unless otherwise specified.
The Analysis of variance showed a significant effect for reading group, F (3,39) = 9.182, p < .0001. Figure 1 shows mean performance on the phoneme awareness tests as a function of reading stage. Overall, the children who differed in the reading ability did differ in level of performance on the phoneme awareness tasks. There was a significant difference for reading group and the Mann English test, F (3,48) = 6.932, p < .001 as well as for the Spanish version, F (3,48) = 5.147, p < .004. Significant differences were also found for the Yopp-Singer tests, English, F (3,48) = 9.267, p < .0001, and Spanish, F (3,48) = 7.913, p < .0001. This replicates previous studies' results that showed that phoneme awareness is highly related to reading ability. No other significant differences were found for reading stage and any other variable.
There were no significant differences for English and Spanish versions of the phoneme awareness tests. On the average, language of presentation did not affect the students' overall level of performance, p>.01. Additionally, no overall differences were found for language background of the children such that children who differ in language control of Spanish and English differed in overall performance. Thus, monolinguals did not perform better than bilinguals or vice versa, in general.