Canadian officials have recently decided to make the long-form of their census voluntary. While the move is being made to protect the privacy of citizens, some people are not happy. The Wall Street Journal describes some of the protests:
Statisticians have protested, arguing that fewer people will respond to a voluntary survey, which will make the results less representative and reduce the government’s knowledge about its populace.
The article goes on to talk about how participation and response rates are dropping for many private surveys. If this is the case, might not the information from voluntary long-form be biased? A study done by the US Census Bureau in the early 2000s showed that changing a government survey from mandatory to voluntary dropped response rates by 20%.
Generally, social scientists operate under the assumption that research participation is voluntary. However, governments have the ability to require participation, say in areas like taxes or the census. Additionally, the statistics collected by the census have broad implications for funding, research, and knowledge about a country.
While some may not like such information being in the hands of government, do they not think their bank or credit card company already knows a lot about them?