Why does UKIP only score well in online polls?
Farage’s party at just 3% in ComRes phone poll
There’s a new phone poll out from ComRes for the Independent. The figures are, with changes on the firm’s last phone poll, CON 37%(+3): LAB 39%(+1): LD 10%(-4): OTH 14%.
This is the first telephone national voting intention poll since the Survation survey on Saturday which had UKIP on 11% – just one point behind the Lib Dems. Although the Lib Dems slipped back sharply in this latest survey they still had a lead of 7 points over UKIP who were down at 3%.
Interestingly the ComRes phone poll UKIP total is very similar to the 4% that Survation itself found in a phone survey a fortnight ago.
Looking across the phone pollsters UKIP struggles with Populus (3%), ICM (4%) and Ipsos-MORI (3%). The big numbers for the anti-EU party come from the internet pollsters – Survation (11%), Angus Reid (7%) and YouGov (upto 7%).
A lot of this is probably down to prompting. With an online poll the parties have to be listed on the screen while with phone polls only the main parties generally get mentioned at first. If the respondent says he/she will vote for another party then the full list is read out.
YouGov has found in the past that if all the minor parties are listed on the front VI page then they do better.
So which form of polling is better? After the experience of the 2010 general election I regard phone polls as being in the top tier to be taken more seriously.
This was the May 2010 polling accuracy table. As can be seen five out of the top six were phone surveys. Five of the bottom six were carried out on-line.
| Rank | Pollster | CON | LAB | LD | Error |
|---|---|---|---|---|---|
| 1 | RNB India: Phone | 37 (0) | 28 (-1.7) | 26 (+2.4) | 4.1 |
| 2= | ICM phone | 36 (-1) | 28 (-1.7) | 26 (+2.4) | 5.1 |
| 2= | Ipsos-MORI: phone | 36 (-1) | 29 (-0.7) | 27 (+3.4) | 5.1 |
| 2= | Populus: phone | 37 (0) | 28 (-1.7) | 27 (+3.4) | 5.1 |
| 5= | ComRes: phone | 37 (0) | 28 (-1.7) | 28 (+4.4) | 6.1 |
| 5= | Harris: Online | 35 (-2) | 29 (-0.7) | 27 (+3.4) | 6.1 |
| 7 | Opinium: online | 35 (-2) | 27 (-2.7) | 26 (+2.4) | 7.1 |
| 8 | YouGov: online | 35 (-2) | 28 (-1.7) | 28 (+4.4) | 8.1 |
| 9= | Angus Reid: online | 36 (-1) | 24 (-5.7) | 29 (+5.4) | 12,1 |
| 9= | BPIX: online | 34 (-3) | 27 (-2.7) | 30 (+6.4) | 12.1 |
| 9= | TNS-BMRB: face to face | 33 (-4) | 27 (-2.7) | 29 (+5.4) | 12.1 |
| 12 | OnePoll: online | 30 (-7) | 21 (-8.7) | 32 (+8.4) | 24 |
| – | Actual GB share | 37% | 29.7% | 23.6% | – |
In May this year the top pollsters for the AV referendum and the Holyrood elections in Scotland carried out their fieldwork by phone.
Mike Smithson @MikeSmithsonPB
