At issue in this debate is whether the relative productivity question in the Australian Workplace Industrial Relations Survey (AWIRS) can give reliable data for the dependant variable in Crockett, Dawkins and Mulvey’s equations. We claim it cannot; they claim it can.

Their Comment suggests the argument turns on four points:

1. (i) Does the distribution of responses suggest effects which might pollute the data?

2. (ii) Did the ‘subjective’ nature of the productivity estimates leave scope for ‘error’?

3. (iii) Are the likely statistical effects of this ‘error’ important?

4. (iv) Do independent and informed authorities support their judgement that the data is useable?