Focus on predictive validity. IELTS research has mainly examined the
test’s predictive validity: the relationship between test scores and
subsequent academic performance, as reflected in test takers’ grade point
average (GPA) at the institutions they go on to attend. This validity
criterion can be problematic for several reasons. As Sawaki and Nissan
(2009, p. 4) note, GPA a) lacks variability, especially in graduate
programs, which attenuates the correlations; b) differs with the different
standards of dissimilar disciplines and institutions; and c) reflects the
“differential values” which lecturers and academicians assign to various
kinds of academic achievement. Sawaki and Nissan (2009) add that the
“summative” criteria of conventional predictive validity studies—which
monitor students’ overall achievement in specific courses—obscure the
actual language performance of ESL students, since academic
performance is a function of much more than language proficiency. In all,
this method of predictive validity research seems to confound the value
implications of test scores, which may help explain why some IELTS
studies have reported weak or even negative correlations between scores
on the listening module of the test and subsequent GPA.