The 1992 British general election is firmly lodged in the political opinion polling hall of shame alongside 1970 and 2015 as being one of the occasions when the pollsters got it ignominiously wrong.
The opinion polls put Labour slightly ahead, and even the two that didn’t either had it tied (ICM) or the Tories only ahead by 0.5% (Gallup). When the votes were counted, however, John Major’s Conservatives were 8% ahead.
But as I recently discovered when reading Nick Moon’s venerable Opinion Polls (dated but still full of good information) there was one poll which got it right. What’s more, it was different from the other polls in a way which brings to mind the 2015 polling debacle and the one opinion poll which got it right then too.
The 1992 polling hero was conducted by then regular political pollster NOP. It was done – unusually – over a week, and what’s more face to face, for the Independent Television Commission as part of its media research. Hence the figures were not published before polling day. When they were calculated, the poll had it almost exactly spot on with a 9% Conservative lead.
Similarly in 2015 it was a poll carried out in a slower, more methodical way and on a larger scale than traditional political polling – the British Social Attitudes Survey research – which got the election result right whilst other pollsters got it wrong.
Perhaps there is a lesson there about speed, cost and quality.
What went wrong with the polls in 1992?
Here is the polling post-mortem carried out by the industry.
File via the Archive of Market and Social Research (AMSR).