My Pollbase database of national voting intention polls, going back to 1943, seeks to be the largest such collection. That’s been helped by the kind cooperation of various pollsters and academics over the years, sharing data from their own collation efforts.
For recent years, my data set if very similar to that of other collections you’ll find, though I usually have a few extra polls that others miss. (Such as, in this Parliament, the very first Techne polls, which were not put on general release but the firm has kindly given me data for.)
But I’ve always known that despite being able to say my dataset is the most comprehensive, that’s not to say it’s complete. A good example is the 1950-1951 Parliament, where I have a Daily Express poll from the general election campaign itself – when records are usually at their best – but not from earlier in the Parliament. That’s a good hint that there are other polls out there, which haven’t made it into previous records and which are waiting to be found in the archives.
So on my to do list for this year has been getting access to newspaper archives and starting to hunt down the missing polls. I’ve done a few mini tests, resulting in extra, if rather confusing, information about Gallup poll from March 1961, for example.
But more excitingly* has been a little test with Dailly Express polling. Just the one newspaper article has already given much of the information about six extra voting intention polls that had been ‘lost’. Previously, we only had 25 polls for the 1950-51 Parliament in the records, so adding an extra six is already quite a boost to the historical record.
It looks like there’s a rich seam of old polls to rescue awaiting in the archives. If I can find the time…
Updates as new information is found will be included in each quarterly update of Pollbase. The next one will include those six extra polls as well as sundry other details picked up from the tests so far.
* Look, if you’ve read as far as this, admit it, you’re excited too.