Summary and discussion
In conclusion, the results provide reasonably strong support for the hypothesis: A more
precise estimate of English proficiency in a specific language domain is possible by
assessing skills not only in that domain but in other related domains as well. We qualify
this assertion by noting that the increased precision is due mainly to the first additional
TOEIC measure that is considered. The returns diminish substantially when further
measures are considered.
The additional variance explained ranged from 2% to nearly 7% for the four domains.
Although characterizing the magnitude of effect sizes is not without some controversy,
one traditional social science norm (Cohen, 1988) has been to regard effects of this size
as being small to medium. Critics of standardized tests often cite such low correlations
and small proportions of variance explained as evidence that tests lack utility in academic
or workplace contexts. Others, however (e.g., Cho & Bridgeman, 2012), have