You are on page 1of 1

Making decisions about learners, teachers, programmes and policies 19

observe the input – test instructions and questions, for example – and the output
– test takers’ responses – and then make inferences about the nature of the system
that produced the output.
There may be some confusion about the idea of indirect measurement owing to
references in the language testing literature to a distinction between ‘direct’ and
‘indirect’ tests. For example Hughes (2003: 17) asserts the following:

‘Testing is said to be direct when it requires the candidate to perform precisely


the skill that we wish to measure. If we want to know how well candidates can
write compositions, we get them to write compositions.’

Such views seem to me to confuse what it is that is being measured. Hughes goes on,
‘The very acts of speaking and writing provide us with information about the can-
didate’s ability’. It is not clear whether Hughes is talking about measuring skills or
ability. Adding to the confusion is the fact that language tests have traditionally been
categorised by skill: reading tests, writing tests, speaking tests, and listening tests. The
assumption appears to be that we can measure these skills separately and give test
takers a score for ‘reading’ or ‘speaking.’ C ertainly we do frequently engage in these
skills in isolation: we often just read for pleasure, listen to a talk show on the radio, write
a letter to the editor of a newspaper, or give an extemporaneous lecture. However, just
as often we combine these skills in communicative language use: we read a chemistry
textbook, taking notes for later use in oral discussion; we engage in rapid listening/
speaking give and take in a telephone conversation; we read an email message and
immediately compose a reply – and the resulting message may have more in common
with informal spoken language than with written language. Importantly, too, it’s dif-
ficult to conceive of using language in a vacuum – we never just ‘read’ – we read for
a purpose, in a context, with some goal in mind; we don’t just ‘speak’ – we speak to
someone, about something, with some communicative intent. The main point is that
it is language ability that underlies each of these complex manifestations of language
use, and it is the nature of that ability that we attempt to measure and make inferences
about when we observe learners’ performances on language tests. Thus, to conclude
this section, my own view, and that of many other language testers (e.g. Bachman and
Palmer 1996, Fulcher and Davidson 2007), is that the object of our measurement is
language ability, which is manifested through the skills of reading, writing, speaking,
and listening, and that is the principle that will guide the discussion of language testing
for the remainder of this chapter and indeed the entire book.

2.2 MAKING DECISIONS ABOUT LEARNERS, TEACHERS, PROGRAMMES


AND POLICIES

Having made an interpretation of test takers’ language ability, based on their test
performance, usually the next step is to make use of that measurement as input for
decisions: are the learners ready to go on to the next unit of instruction? Which of
them should be promoted to the next level? What areas of the language are learners

You might also like