obviously depending on what topics
within the target construct are being tested. Hughes (2003: 31) goes so far as to suggest that
defining construct validity is not necessary for direct tests of what are sometimes seen to be
common-sense constructs, which he names as reading and writing.