from The Vancouver Sun

OTTAWA — Statistics Canada is cautioning its new, voluntary National Household Survey released Wednesday contains incomplete and possibly incorrect data in a number of areas, including Aboriginals, languages, immigrants and visible minorities.

The federal agency warns the voluntary nature of the 2011 National Household Survey (NHS) — which replaced the mandatory long-form census — means there’s “non-response bias” in the results for certain populations and geographic areas.

Various other problems and the non-response bias in the NHS means the population counts of a number of groups are either overestimated or underestimated.

The issue of non-response bias is a concern because the characteristics of people who responded to the survey are different than those who refused to participate. That means there are gaps in the data from groups and communities that tend not to respond to voluntary surveys — including Aboriginals, new immigrants and others.

The risk of the non-response bias quickly increases as the response rate declines, so the smaller the group surveyed, the less reliable the information.

“It is impossible to definitively determine how much the NHS may be affected by non-response bias,” Statistics Canada explains in the documents. “Generally, the risk of error increases for lower levels of geography and for smaller populations.”

Some data, whose quality was not considered satisfactory, were suppressed (not included) for certain small population or geographic areas, but were included at the national level.

Statistics Canada also encountered other problems that prevented it from surveying groups of Canadians such as First Nations.

A majority of Indian reserves participated in the 2011 National Household Survey, however, 36 of the 863 inhabited reserves were incompletely counted because enumeration was not permitted on reserves, was interrupted before completion or because of natural events such as forest fires for 13 reserves in Northern Ontario. (Officials were able to return later to some reserves to complete the household survey, but the data were not ready for release on Wednesday).

“The issue of incompletely enumerated reserves has the most impact on HNS estimates for the First Nations population living on reserve as well as the Registered or Treaty Indian population living on reserve. Estimates for other populations will also be affected,” says the NHS.

In 2006, there were 22 Indian reserves and settlements that were incompletely enumerated in the census.

The population counts of a number of groups are overestimated or underestimated, Statistics Canada warns.

For example, estimates and trends from other data sources suggest the Inuit population living outside of Inuit Nunangat is overestimated at the national level. (Inuit Nunangat includes the four Inuit regions of Northern coastal Labrador, Northern Quebec, territory of Nunavut and parts of the Northwest Territories).

As well, analysis of the data on same-sex married couples show this family type and marital status may be overestimated.

The 2011 NHS shows 63,920 same-sex couples in Canada, of which 20,280 are married couples. Both of these counts are believed to be overestimated at the national level somewhere between 0 and 3,800.

Statistics Canada also says evidence suggests the NHS estimates for the population born in the Philippines is overestimated at the national level.

Population estimates suggest the number of immigrants from the Philippines who entered Canada between January 2006 and June 2011 is 141,502, while the NHS estimate between January 2006 and the survey date of May 10, 2011 is larger (152,270).

Also, the population born in Pakistan is suggested to be underestimated in the 2011 household survey, as well as the Arab population.

Other problems were found in determining languages in Canada.

For example, evidence suggests the population who reported Malayo-Polynesian languages as their mother tongues is overestimated at the national level, while the population having “Romance languages” (Italian, Portuguese, Romanian, Spanish, Catalan and others) as their mother tongue appears to be underestimated, says Statistics Canada.

The National Household Survey was sent to approximately 4.5 million randomly selected households, or roughly 30 per cent of all Canadian dwellings.

Approximately 2.65 million Canadian households participated in the survey, for a response rate of 68.6 per cent among those homes that received the survey, or approximately 20 per cent of all households in Canada in 2011.

In 2006, the mandatory long-form census was sent to approximately one in five Canadian households, with a response rate of 93.5 per cent.

By JASON FEKETE This e-mail address is being protected from spambots. You need JavaScript enabled to view it , POSTMEDIA NEWS