Close on the heels of the latest ABS publication on Working Time Arrangements, the subject of a long blog discussion on Andrew Nortons site, the ABS has followed up with Preferred Working Hours of Wage and Salary Earners, Queensland.
This Survey found that, of people surveyed in Queensland, only about half were happy with the hours they worked. About 14 per cent preferred to work more hours and a whopping one third preferred to work fewer hours. In that last category, however, only about one in five was prepared to work fewer hours for less pay the majority wanted to work fewer hours for the same pay.
In the course of perusing the publication, I was reminded that when the ABS carries out its household surveys it asks only one person (any responsible adult) the survey questions about all of the relevant people in the household. The ABS explains that this means that
It is possible that some answers supplied by the responding person may be different to what would have been answered by the in-scope person themselvesalthough it has been estimated that the error resulting from this occurrence is small
Now, I can understand that when it comes to some survey variables, eg gender, age, relationship with other members of the household, the answers provided by someone else can be expected to be pretty reliable. But the further you get away from this into areas that the responsible adult may or may not know, the more error is likely to creep in. This is why, for example, most people would prefer to use data on earnings collected from employers rather than data on earnings collected through the ABS household survey.
Now, when it comes to preferred hours of work, we are really getting into peoples preferences and attitudes. I was thinking about how this might play out in a stereotypical household where the husband is married to the job and this has been a longstanding bone of contention with his wife.
Next, if it were LSW, how would she be likely to answer the question about whether WH preferred the long hours he worked? Well, hadnt he always told her that he only worked those hours because it was expected of him, and of course he would much rather spend more time at home with the family? You can see where Im going with this, cant you?
But, of course, the amount of respondent error attached to data collected like this would be small, wouldnt it? Just wondering