What’s happened to “other others” at ComRes?
Should we be “recalibrating” its figures?
Something very strange is happening with the ComRes polling numbers which look as though they might be being distorted by a huge increase in “other others”.
This is the category that excludes the three main parties, SNP/PC/UKIP/GRN/BNP and in, for instance, the firm’s February 11th poll was down to just four respondents in a sample of over a thousand.
According to the data from last night’s poll for ITV News/Independent “other others” were recorded as being the choice of 55 respondents giving a share of 8% of those intending to vote – something that we have not seen on this scale from any other firm.
This looks like being a methodology issue and I have emailed ComRes boss, Andrew Hawkins, for his thoughts.
I wonder whether most of the “other others” should be categorised as “don’t knows”. So this is my re-calculation of the ComRes poll with “other others” recorded at the mid-February level.
| ComRes “recalibrated“ | Adjusted | Poll |
|---|---|---|
| CONSERVATIVES | 38.0% | 35% |
| LABOUR | 31.7% | 29% |
| LIB DEMS | 23.0% | 21% |
| LAB to CON swing from 2005 | 4.7% | 4.5% |
We have also got the latest Harris poll for the Metro shown below with changes on the weekend survey by the firm for the Daily Mail. Fieldwork for the poll started last Thursday and finished on Tuesday so was only partially influenced by the manifesto launches.
| Harris (Metro) | Apr 13 | Apr 6 |
|---|---|---|
| CONSERVATIVES | 36% | 37% |
| LABOUR | 27% | 27% |
| LIB DEMS | 23% | 22% |
| LAB to CON swing from 2005 | 6% | 6.5% |
There is a Crosby-Textor marginals poll in the Daily Telegraph where the reported comparison figures don’t appear to make sense. I’ll wait for further information before covering it.