The Effect of Format and Language on the Test Scores of Bilinguals

As American colleges become more diverse, cultural
differences among students raise important concerns
about large scale assessment of students whose
primary language is not English. Are the scores
obtained by these students comparable to the scores
of U.S. students for whom the assessments are
designed? Score comparability ensures that the
meaning and interpretation of test scores are the
same for all groups of students. The purpose of the
present study was to examine the effects of lingual
background and item format on the test performance
of bilinguals. Contrary to what was expected, format
and lingual background did not collectively affect
the performance of bilinguals on the construction
items to a larger extent than on multiple-choice
items. The majority (88%) of the bilinguals used in
this study were students born and schooled in Puerto
Rico, where the constructed-response format is
traditionally used to test students throughout their
schooling. This study confirmed the findings of
Cooley & Leinhart (as cited in O Leary, 2001) and
O Leary (2001), that frequent exposure to a test
format will make a difference to performance.

Autorentext

Juan L. Trujillo, PhD, is an Assistant Professor of Measurement and Statistics at Albany State University. His research interests include psychometrics and applied statistics. He received his PhD in Measurement and Statistics at the Florida State University. Dr. Trujillo was born in San Juan, Puerto Rico.

Klappentext

As American colleges become more diverse, cultural differences among students raise important concerns about large scale assessment of students whose primary language is not English. Are the scores obtained by these students comparable to the scores of U.S. students for whom the assessments are designed? Score comparability ensures that the meaning and interpretation of test scores are the same for all groups of students. The purpose of the present study was to examine the effects of lingual background and item format on the test performance of bilinguals. Contrary to what was expected, format and lingual background did not collectively affect the performance of bilinguals on the construction items to a larger extent than on multiple-choice items. The majority (88%) of the bilinguals used in this study were students born and schooled in Puerto Rico, where the constructed-response format is traditionally used to test students throughout their schooling. This study confirmed the findings of Cooley & Leinhart (as cited in O'Leary, 2001) and O'Leary (2001), that frequent exposure to a test format will make a difference to performance.