And the recurring problem with the NILT Survey (and it hasn’t gone away you know)?
Political party support (%)
DUP/Democratic Unionist Party 17
Sinn Fein 12
Ulster Unionist Party (UUP) 10
Social Democratic and Labour Party (SDLP) 13
Alliance Party 8
Other Party (please specify) 2
None of these 24
Other answer (please specify) 5
Don’t know 8
For comparison, including only those with a political preference from the above, this would translate as follows (NILT figures in bold: electoral figures in brackets are for 2011 Assembly – 2011 Local Government – 2010 Westminister).
DUP: 27.4% (30%-27.2%-25%)
SF: 19.4% (26.9%-24.8%-25.5%)
UUP: 16.1% (13.2%-15.2%-15.2%)
SDLP: 21% (14.2%-15%-16.5%)
Alliance: 12.9% (7.7%-7.4%-6.3%)
Others: 3.2% (7.9%-10.4%-11.5%)
The only control measure available to calibrate NILT is the return for party political support which can be compared to electoral results as shown above. And it continues to fail to match. And the past defence on the basis of NILT (a survey) being accurate and the dismissal of electoral results as tactical is pure exceptionalism. [For slow learners, exceptionalism is where you want to believe that, eg, a survey is more accurate than election results, where people have actually voted, in mapping political preference]
The failure to accurately map political trends also undermines the credibility of other statistics where political sentiment would be influential (e.g. identity, constitutional preference etc). So there is an apparent problem in structure somewhere, either in the form of questions, survey approach, interpretations, or, in understanding how to match it against community attitudes towards responding to such surveys. Given that this has been a recurring and commented upon feature of NILT, it must be presumed that the disparity is somehow acceptable. Remarkably, then, there is a significant difference in the political drift that seems to appear in NILT and other ‘surveys’ as opposed to empirical exercises like censuses or actual elections.
In a region with a significant historical problem with official disinformation, NILT looks suspiciously less like a survey and more like another political tool.