Table 1 shows good internal correlations between listening, reading and total scores at a
very high level of significance, suggesting that both sections are measuring aspects of the same
thing: 'general proficiency'. However, the data reported in table 2 suffer from various
shortcomings, namely: though high on face validity (section 6.1.4), three of the concurrent tests
were unvalidated and all were scored subjectively; the level of significance is reported for only
one concurrent test, the LPI, making interpretation difficult; the sample sizes, though adequate,
are not substantial; and no 'negative evidence' is offered to show that the test is not testing other,
unrelated abilities, as advised by Bachman (1990: 259). These data should therefore be
interpreted with caution.