Abstract:
This study was done to compare the content validity of standardized and non-standardized language achievement tests. The sampled non-standardized language achievement tests for this study were taken from Gurage Zone. These tests are teacher-made achievement tests with the purpose of modeling the EGSEC tests. Similar with the EGSEC English tests they contained 80 multiple choice questions. The study was done with the assumption that the two language achievement tests were dissimilar. The study was conducted with mixed design. The collected nominal data from the relevant data sources were analyzed quantitatively and qualitatively. The content compatibility of each expected and observed data was checked with Pearson chi-square analysis. The measures of association between the above variables have also been computed with Contingency formula. The rank and percentage of syllabus objectives in the syllabus and in sampled tests were also judged based on decisions’ of judges. Moreover, data related with the preparation of the two kinds of tests were collected through interviewing two purposively selected experts. The finding of the study has showed that the two kinds of tests were contently not well representative of the contents/objectives in the syllabus. The average chi value of EGSEC tests resulted 50.24 where as the average chi value of zonal tests resulted 139.7.This entails that the zonal test was highly divergent of the syllabus than the EGSEC English tests did. The measures of association of the zonal tests were also weaker than the EGSEC English tests. Thus, the assumption made at the start of the research was confirmed that zonal tests in Gurage zone were contently divergent even more than the standardized English tests. In addition, comparison of the two tests was done based on five judges’ decision of the syllabus objectives and sample tests. 80% of the zonal and 67% of the EGSEC English tests were sampled from cognitive domain though the textbook has given high (60%) for psychomotor domain. In general, in all ways of the comparisons, zonal tests were more contently unrepresentative sample of the textbooks contents and learning objectives where they should be better
