November 3, 2010

Reducing Poverty Through Statistical Trickery?

There are different ways to try to eradicate poverty. One which has been adopted in European Union countries and seven Canadian provinces is to create a poverty reduction strategy with a goal, targets, timelines and an action plan.

Another way to reduce poverty is through statistical sleight of hand; simply stop counting some of the poor.

A real poverty reduction strategy does not simply choose to not count some people who live in poverty. Rather it relies on sound data to help identify where to focus attention.

When the Federal Government decided to scrap the mandatory long-form census and replace it with a voluntary National Household Survey, it created the conditions to undercount the number of people living in poverty and undermine the efforts of provincial government and local communities with poverty reduction strategies.

Statistics Canada explains how the shift from the mandatory long-form census to the voluntary National Household Survey will lead to bias in the data it collects. That bias results from the fact that some groups -- including people living in low income -- will be less likely to fill out a voluntary survery than a mandatory census. is believed that the most significant source of non-sampling error for the National Household Survey will be non-response bias. All surveys are subject to non-response bias, even a Census with a 98% response rate. The risk of non-response bias quickly increases as the response rate declines. This is because, in general, non-respondents tend to have characteristics that are different than those of the respondents and thus the results are not representative of the true population. Given that the National Household Survey is anticipated to achieve a response rate of only 50% there is a substantial risk of non-response bias.
The Statistics Canada analysis shows that the number of households with incomes below $1,000 or with no income would have been 4% lower in Toronto if the voluntary National Household Survey rather than the mandatory long-form census had been used for the 2005 census. Similar results would have occured in Winnipeg and Bathurst, New Brunswick. By contast the voluntary survey would have shown more people with incomes above $50,000.  The voluntary survey would have skewed figures for various demographic groups.

While Statistics Canada will be making an effort to overcome some of the problems it has identified with the sudden replacement of the mandatory long-form census, it still means the data will be less reliable than before. And it will be impossible to make accurate comparisons with data from previous censuses.  That means it will be harder to see whether Provincial and local poverty reduction strategies are working or to get an accurate sense of where priority for updated action plans should be placed.

The Federal Government has made no commitment to create a Canadian Poverty Reduction Strategy; changing the census does not count as a poverty reduction strategy.

No comments:

Post a Comment