Shy Tory Factor

The Shy Tory Factor is a name given by British opinion polling companies to a phenomenon first observed by psephologists in the early 1990s. They observed that the share of the electoral vote won by the Conservative Party (known colloquially as the 'Tories') was significantly higher than the equivalent share in opinion polls. The accepted explanation was that so-called "shy Tories" were voting Conservative after telling pollsters they would not. It has only been present in the 1992 General Election, but has been discussed in subsequent elections where the Conservatives did unexpectedly well.

In the 1992 general election, the final opinion polls gave the Conservatives between 38% and 39% of the vote, about 1% behind the Labour Party – suggesting that the election would produce a hung parliament or a narrow Labour majority and end 13 years of Tory rule. In the final results, the Conservatives received almost 42% (a lead of 7.6% over Labour) and won their fourth successive general election, though they now had a 21-seat majority compared to the 102-seat majority they had gained in the election five years previously. As a result of this failure to 'predict' the result, the Market Research Society held an inquiry into the reasons why the polls had been so much at variance with actual public opinion. The report found that 2% of the 8.5% error could be explained by Conservative supporters refusing to disclose their voting intentions; it cited as evidence the fact that exit polls on election day also underestimated the Conservative lead.

After the 1992 election, most opinion pollsters altered their methodology to try to correct for this observed behaviour of the electorate. The methods varied for different companies. Some, including Populus, YouGov and ICM Research, have adopted the tactic of asking their interviewees how they had voted at the previous election, and then assumed that they would vote that way again at a discounted rate. Others weighted their panel so that their past vote was exactly in line with the actual result of the election. For a time, opinion poll results were published both for unadjusted and adjusted methods. Polling companies have found that telephone and personal interviews are more likely to generate a shy response than automated calling or internet polls.

...
Wikipedia