Insight

Scotland predictions

Scotland predictions

But aside from diligently towing the company line, what do the Opinium team think will actually happen? With the entire polling industry potentially having ostrich sized egg on its face on September 19th here are some predictions to add a little more potential embarrassment once final results are in.

Adam Drummond?s prediction

Starting with the obvious, a referendum is totally different to an election and all indications are that hundreds of thousands of people who don?t normally vote in elections are going to take part in the referendum. This means that all our polls, which are based on likely voters and systematically undercount these people, could end up being completely wrong.

That said, I think the end result will be fairly similar to our two polls with a narrow win for the No camp. There are two scenarios I can think of that would cause the actual result to be significantly different to the polls.

The first is that mentioned above; unprecedentedly high turnout from normally low turnout groups which could have a disproportionate influence in such a close race. Whether these people lean Yes or No is difficult to say but the conventional wisdom is that they are more disillusioned with the political status quo and more likely to want to shake everything up and vote Yes. However, in our Observer poll, people who say they didn?t vote in the 2010 Westminster OR 2011 Holyrood elections break for ?No? by 63% to 37%. The sample is very small (75 people) and the poll was conducted online and many non-voters do not have internet access but it suggests that the conventional wisdom, if not wrong, should at least be treated with caution.

The second is the possibility of a ?shy No? vote, similar to the Conservatives in 1992 where it?s more fashionable to vote Yes and Yes voters are more likely to take part in phone polls and click on the link in online polls.

I suspect these two factors will probably come into play but the lack of an overwhelming Yes slant among our ?non-voters? makes me believe that if the result is different to my prediction then it will either be an extremely narrow Yes win or a much larger win for No. It sounds like I?m hedging my bets so I?ll put a number out there and say 53% for No and 47% for Yes.

James Crouch?s prediction

The fact is that referendums in Britain are still such an unknown quantity ? this will be only the third Scotland-wide referendum in the best part of four decades, which doesn?t allow for a huge amount of a comparative analysis.

Most opinion polls have shown a very clear lead for the unionist campaign right up until the last few weeks, with only recent polls appearing to show the gap narrowing. What people say in answer to the pollsters in the final few weeks of a hotly debated election can sometimes say more about the campaign than what voters will actually decide in the polling booth. The final results of the last general election show how fluctuating polls in the immediate run up to an election may obscure the polling data that had been consistent for months beforehand.

Finally, there is always the problem of what the ?undecideds? are going to do on the day itself. In a general election, the split of don?t knows in three or even four ways has perhaps less of an effect than you might think on headline poll figures. In a race like this, 6% or 8% don?t knows could have a stunning ability to tip the scales one way or the other. An example of what I?d expect to happen to these is the same as in the AV referendum. For months polls showed roughly a third for, a third against, and a third who didn?t know ? for me the final results of a two-thirds ?No? and a third ?Yes? speak for themselves regarding how these ?undecideds? will split.

With both of these factors in mind, my educated conjecture would be for a ?No? win, but not quite as narrow as we might expect from the final polls. If I had to put a number to it, I?d go with a generous 55% for No and 45% for Yes.