Fewer respondents put Swedish statistics accuracy in jeopardy

A sharp decline in the number of Swedes who choose to participate in telephone polls is threatening the quality of official statistics, Sweden Statistics has warned.

The number of Swedes who decline requests to participate in Sweden Statistics polls has doubled in eight years, from 18.9 percent to 38.8 percent, Swedish national daily newspaper Dagens Nyheter reported on Sunday.

"This is one of the biggest problems we have. We are constantly working to reverse this trend. It becomes increasingly hard to make conclusions based on the statistics. The uncertainty rate and margins increase," Peter Beijron, statistician at Sweden Statistics, told the newspaper.

Theories to explain the decline include an "inflation" in polls, people are growing tired of an increasing number of poll phone calls.

The decline is especially high in certain groups where unemployment is higher, for example among foreign-born residents and young people. That means an even higher risk that statistics do not reflect reality.

"If the decline continues at the same rate for a few more years, it could, hypothetically, mean we would not be able to provide certain charts on a monthly basis. When we break down the result into different groups, we risk a situation where we do not have enough respondents to provide the statistics," Beijron said.

Bengt Swensson, statistics professor at Uppsala and Orebro universities, is concerned the development means official unemployment statistics cannot be trusted. Endit