Diagnostic Theory of Information as a Testing Technique to Ascertain Spanish Reading Proficiency

Abstract

The purpose of this study was to investigate the use of information theory as a testing reading proficiency of bilingual students. This study is not an attempt to analyze mistakes or to debate which reading proficiency test is the best in testing Spanish language abilities or expertise. This test was constructed to determine the readability and the validity of information theory as a testing tool. A second purpose was to determine whether this technique could help increase the student's ability to comprehend reading material at his/her level of language learning. A third purpose was to try to develop a Spanish reading proficiency test that could be used in the classroom. This study of Spanish reading proficiency consisted of three experimental groups. The first was composed of ten fourth, fifth and sixth grade students from Dr. Martin Luther King Jr. #9 School in the city of Rochester. All of the students in this group were raised in the United States. This group was at an intermediate level and has been in a bilingual program no less than three years. They all received an equal amount of English reading and Spanish reading during the day. This group is categorized as English dominant in a bilingual setting. The second group consisted of ten fourth, fifth, and sixth grade students who were more recent arrivals in the United States. This group was also at an intermediate level and the students selected have been in a bilingual program since at least September of 1989. This group is categorized as Spanish dominant in a bilingual setting. Finally, the third group was composed of ten students from Spanish-speaking homes, but are enrolled in an all-day English only classroom. This particular group is also at an intermediate level. Each group was tested with a variety of reading materials in the target language. The materials used were poems with part of the information deleted. The objective of the test was for the subjects to encode the missing letters within a five minute, four minute, and three minute time frame. All three groups were presented with a similar procedure at different levels of difficulty. The passages consisted of approximately 300 to 400 bits of information each.