The Data

The New Zealand Census collects data on every resident who was in New Zealand on Census night, in this case, Tuesday 7 March 2006. The Census thus provides the most inclusive sample of data available.

While the Census is the best data set for examining a profile of long hours workers, it is not without problems. The most significant of these for this project relates to the way data on working hours is collected. The Census asks respondents “How many hours, to the nearest hour, do you usually work each week?” Unlike the Household Labour Force Survey, which asks respondents to report their actual working hours for each of the last seven days, the Census requires respondents to estimate the average hours they “usually” work. It is thus likely that at least some responses will vary from actual hours worked, as people round their working hours up or down, report hours that cluster around common standards (such as “40” or “50” hours a week), or forget to report increased or decreased hours that have occurred or will occur in the future (such as including overtime).

In addition, there is no guarantee that the census questionnaire is completed by the member of the household for whom the data is gathered, a problem with all self-completion surveys. If the form was completed by someone else in the household, the accuracy of reported working hours may be in question.

Finally, despite escaping sampling errors and bias that may exist with other data sets, the Census still does not represent a fully complete picture of all New Zealanders. [1] Furthermore, like any self-completion survey, there are respondents whose written answers are illegible, or who perhaps do not fully understand the question being asked and thus provide an answer “outside the possible”.

While these issues need to be kept in mind when analysing the data, the Census still provides the most complete picture of working hours amongst New Zealand workers. This paper is based on the analysis of the data of the 1,832,490 people who report working at least one hour a week. As such, the data that follows does not include those who are not in paid work. When couples are described, calculations are based on couples where both are in paid work for at least one hour each week.

This paper does not attempt to explain patterns in the data, or to give reasons why workers with particular characteristics are more likely to work long hours; rather, the paper provides an elementary analysis of the variables relevant to long hours and how they inform a profile of who works long hours in New Zealand.

---