Question 33 was never designed to cover every aspect of unpaid work. Unpaid work probably accounts for about one-half of all human work hours in Canada. But unlike paid work, which is covered in 12 questions over many, many pages in every census, unpaid work has been segregated into one single question. So it has been framed as what is referred to as a symbolic question, which gets at the three most important unpaid activities people engage in.
It's a very wide-ranging framing of the question, and I'd just like to emphasize that it does not ask about unpaid work; it asks about unpaid activities. That's why the census data really are the place to start, because they reach every part of Canada, every community, every income category, and every cultural and linguistic group. No one is left out.
In a quick comparison of just how the two samples, the GSS versus question 33, compare, I discovered that in the last round of census data collection, 40.9% of all the women who responded to the census—and there were 25.5 million responses to the census—indicated they did some unpaid activities relating to children.
The GSS for the year just before that, for 2005, which only sampled 19,500 people in the whole country, or 2,000 per province, was only able to find 20% of all female participants in its sample reporting any child care work. So there are significant data limitations with the GSS.
The census is absolutely crucial with respect to question 33, even though we could make it a lot better than it is.