YouGov are carrying out genuine daily voting intention polls throughout the conference season. The first week’s figures for the Lib Dem conference are up on the YouGov website here now (hopefully – it’s IP blocked until we pass the embargo at 10.15!), and for the next two weeks they’ll go out daily there and on Sky News at 5pm. As well as voting intention, there will be daily topical questions on whatever issues arise at the conferences.

In the past daily polling has almost always been a “rolling poll”, so each day’s sample was actually made up of a couple of days small samples, with the oldest one dropping off each day to be replaced by new data. Each day’s YouGov poll is a completely new sample of 1000 people (less than the normal YouGov sample of 2000, to make sure we don’t start asking people too regularly). They are also being carried out in just 24 hours, so the data published on Monday was all collected on Sunday and Monday morning, Tuesday’s data was collected between Monday and Tuesday (for those of who interested in the mechanics, in the past we were wary of really quick polls in case people who responded fast to surveys were different from people who responded slowly. We’ve got round this using a different method of inviting people to surveys, without going into details the end result is that some of the people answering the daily surveys will have been invited a day or two before the survey even began, meaning you still get a mix of fast and slow respondents).

graph

The latest voting intention figures are CON 40%, LAB 24%, LDEM 21%. As you can see from the graph, the Lib Dems peaked straight after Nick Clegg’s conference speech when they reached 23%, but only for a day. Their Friday score of 21% is still significantly up on the 17% they scored on Monday at the beginning of conference season. The Conservatives went down to 38% at the same point, but ended the first week of conference season back at 40% (still down on the last YouGov poll from before conference which had them at 41%).

The big losers so far as Labour, pushed down to 24%. Of course, starting today we will have the first daily poll conducted during the Labour conference, and in the next few days we may (or may not) see Labour’s totals climbing.


18 Responses to “Daily YouGov polls for the Conference season”

  1. Voting intention.
    38 – 23 = 14
    dont think so!
    Laz

  2. Gah! I can’t change that when I’m at work either. It’ll just have to stay up all day and embarrass me ;)

    [edit – right. Managed to change it. Down the memory hole with that!]

  3. What is interesting is that pre-LibDem conference, Labour polled 30%!! That has to be a high in recent months/years?

  4. Sorry, but this is pretty obvious rubbish. And I’m surprised to see it’s inclusion on this blog. (Especially given the careful nature in which most polls are examined).

    Take the following Sky statement:

    “Labour are down from 30% on the 21st September to just 24% now”.

    Where on earth did SKY get the idea that Labour were on 30% just a week ago?

    Labour were on about 26% then and are on about 26% now.

    The other parties have similar faults.

    Put this dreadful thing to sleep and let’s stick to pollsters who at least have an idea as to what they are doing!

  5. They are sampled the same as normal YouGov polls, using the same wording as normal YouGov polls, processed and weighted in the same way as normal YouGov polls.

    The only differences are sample size, which is smaller, but still a perfectly good size (indeed, it is ICM’s standard size), and the compacted time for the fieldwork.

    The main potential problem with that would have been if people who responded slowly were different from people who responded quickly. That isn’t an issue here, since each survey has a normal spread of slow and fast respondents, because some people were invited prior to the survey going live.

    The only other potential snag is whether a survey done only on a Sunday might behave strangely – we’ll see. Certainly for the other days of the week these should be treated as normal YouGov polls.

  6. er.. yeah. I’m really interested in this, please keep it up!

  7. Since they are only polling half of their normal sample size these snapshpts are bound to be erratic if not inaccurate, The meejah will like it and YouGov will make some dosh but I’ll stick to properly weighted and constructed major polls thanks.

    How in the time can they possibly apply their long established weighting criteria to a truncated sample ?

  8. How in the time can they possibly apply their long established weighting criteria to a truncated sample ?

    Since they are only polling half of their normal sample size these snapshots are bound to be erratic if not inaccurate, The meejah will like it and YouGov will make some dosh but I’ll stick to properly weighted and constructed major polls thanks.

  9. Whoops ! Sorry!

  10. …but if you don’t read *too* much into the movements you see on individual days, and rather look at the general trends, then over any given period of several days you’re seeing a much *larger* sample than normal.

  11. WEs – But “How in the time can they possibly apply their long established weighting criteria to a truncated sample ?”

  12. Anthony,

    Are you going to include these polls in your VI list (http://ukpollingreport.co.uk/blog/voting-intention)?

  13. Christina – in exactly the same way. Within reason (clearly no amount of weighting is going to save a sample of 100), target weights are not dependent at all on sample size. You weight a sample of 1000 to the same targets as one of 10000.

    LondonStatto – Yep, but I have to decide how to do it without swamping it. That said, I’ve already got adjustments in there to weight down multiple polls from the same company, so I’ll probably put them in in the normal way.

  14. Anthony –

    The post above (paragraph 2) uses “we” to refer to YouGov, and I notice that the site is now “in association with” YouGov too – just wondering what this all means. Are you working for/with YouGov on the design/analysis of any of these polls?

  15. Yariv… sign up to YouGov and you will see Anthony beaming down upon you with Peter Kellner and John Humphries…

  16. Yariv.

    UKPollingReport is all owned by me, but my day job is working for YouGov. ” In association with YouGov” is really just to flag up the connection.