A STUDY OF STANDARDIZED LANGUAGE DOMINANCE TESTS AND THEIR CORRELATION WITH BILINGUAL JUDGES' OPINIONS

This study investigated the specific problems inherent in identifying reliable and efficient means of determining which Hispanic students are Spanish-dominant, English-dominant or bilingual. Specifically, the relationship between bilingual judges' opinions about language dominance and selected measures of language dominance on the following tests was studied: the Crane Oral dominance Test, the accompanying Crane Questionnaire, the Pupil Questionnaire for Language Dominance Survey and the Inter-American Test of General Abilities (Oral Vocabulary subtest).;The investigation studied four question relating to the problem of assessing students' language dominance. They were: (1) What is the relationship between group-administered tests of language dominance and bilingual judges' opinions? (2) What is the relationship between individually-administered tests of language dominance and bilingual judges' opinions? (3) What is the relationship between language survey utilized to determine language dominance and bilingual judges' opinions? (4) What is the relative efficiency of group and individual language dominance tests, and language dominance survey forms a predicting language dominance as classified by bilingual judges?;Relationships between judges' opinions and selected language dominance instruments were analyzed through the use of cross-tabulation, including Lambda and Cramer's V. Similar procedures were utilized in comparisons between instruments and to establish inter-rater reliability among bilingual judges.;Major Findings. According to the results of this study, little relationship exists between individual and group-administered tests of language dominance and sociolinguistic questionnaires when compared to bilingual judges' opinions. Percentages of agreement, predictive association and strength of association were highest between sociolinguistic questionnaires and bilingual judges' opinions. In comparing measures of language dominance to each other little relationship was found to exist. However, both sociolinguistic questionnaires included in the study did reveal the highest percent of agreement, predictive association, and strength of association. The findings suggest that measures of language dominance presently used in the state of Connecticut are not consistent with bilingual judges' opinions regarding students' language dominance. In addition, while paper and pencil tests or sociolinguistic questionnaires measure some language domains, none of the measures included in this study included all domains of language dominance.;The total sample consisted of 123 Hispanic students from fourteen schools. Students were randomly selected from all Hispanic students in grades one through three in each of the school districts. Two urban centers and one rural community participated in this study. Four language dominance measures were administered to participating students. In addition, language samples of each student were assessed by three bilingual judges.