Method-effect on Test-takers’ Performance and Confidence in Language Tests

Resumo

In a language test, the number of correct answers is commonly used as the only source of information that defines the result. The question raised is the following: do all task types accommodate all test-takers in the same way? This paper attempts to investigate this relationship by relating correctness and confidence in the answers provided with the task typology. Two achievement tests in two versions (same language content – different task type) were used while test-takers were asked to mark their degree of confidence for each item on both versions. Data obtained through this research were: 1) number of correct answers, and 2) degree of confidence in different task types. Through the correlations of the variables examined (task typology, correctness and confidence) several findings were registered concerning the relationship between accuracy and confidence in different task types. The task type typology may be seen as a moderating factor which could be responsible for possible variations in accuracy and confidence scores.