This study explores the validity and reliability of three measures of achievement: a schema-based cloze multiple choice item test (MCIT), a Persian to English MCIT and an English to Persian open-ended translation examination (OETE). The results obtained on the performance of 89 undergraduate university students show that the schema-based cloze MCIT has the highest reliability (α = 0.93) followed by Persian to English MCIT (α = 0.80). Although English to Persian OETEs suffer from not yielding to objective measures of reliability, they correlate significantly with schema-based cloze MCITs and Persian to English MCITs and can therefore be used as valid measures of achievement in general English courses. Since the three tests correlate significantly with each other they establish themselves as empirically valid measures of translation ability, too. The multivariate analysis of variance, however, reveals that the three measures have significantly different means (Wilks\\\\\\\\\\\\\\\' Lambda F= 953.14, p<0.001). Since the schema-based cloze MCIT enjoys the highest index of reliability and its significant correlation with the English to Persian OETE (r = 0.75) is higher than the significant correlation between Persian to English MCIT and English to Persian OETE (r = 0.61), it is suggested that schema-based cloze MCITs be used as reliable and valid measures of both translation ability and achievement.

Keywords