What the 1975 Public Affairs Act tells us about opinion polling

To recap: to demonstrate the tendency of people to answer survey questions even when they have no idea about the answer, there's an experiment about the "1975 Public Affairs Act" where survey respondents are asked if they think it should be repealed or not.

When the experiment was first run by George Bishop, between 20 and 40 percent of Americans were willing to offer an opinion on this Act, saying that they supported or opposed repealing it.

The problem is that the 1975 Public Affairs Act is fictional so obviously the correct answer is "the 1975 Public Affairs Act is fictional, so stop wasting my time and move onto which celebrity I'd like to have tea with!"

This was originally done by telephone poll and face-to-face but YouGov and the Huffington Post recently repeated it online.

A significantly smaller proportion of people in this survey offered an opinion but then they're then told that the Republican party sports the repeal while President Obama opposes it and asked what they think. While the vast majority still did not offer an opinion, those who did spilt along partisan lines, supporting the Obama or Republican position depending on their side.

This experiment shows us the effect of social pressures on survey answers as people will often feel embarrassed about their supposed ignorance on a subject and offer an opinion anyway to .

What it really shows us though is a fairly basic lesson in survey design which is that you have to consider whether respondents will reasonably be able to answer your question. No doubt you could run the same experiment about an obscure piece of existing legislation and respondents will be just as ignorant.

It also shows it's important to have a "don't know" option in surveys. When Bishop and his colleagues offered this in their survey, the proportion offering an opinion dropped from 24% to 4%. When YouGov asked online, this figure was around 15% which may be because online respondents are more used to seeing this opt out option in surveys.

This brings us to the second part of the experiment which is that people who support a particular party are more likely to agree with a position if, absent any other information, the leaders of that party also support it. In the US it was Barack Obama and the Republican leaders in congress but no doubt it would be similar here with David Cameron and Ed Miliband.

This has been portrayed as showing the ignorance and idiocy of the modern voter but it's actually just the opposite.

Most people, if they follow politics at all, only do so at a very general level when it doesn't impact directly on their interests. They don't have much information about particular bits of legislation and therefore don't have particularly strong views on a given law.
However, they do know that they generally agree with, for example, the Conservative party on most issues and without any other information (which, when describing a fictional law, is likely), what David Cameron thinks is a logical thing to consider when deciding their own position.
This isn't ignorance but an understandable and logical approach to politics. The world is a complicated place and people have a lot going on in their lives without having to keep track of every white paper and proposal. That's why we have a representative democracy rather than the ancient Athenian ideal of direct participation. There's a reason why you had to have a self sufficient household to be a full and active participant in Athenian democracy because it required substantial time and effort that's not possible when you also have to put food on the table.

As Jonathan Bernstein says, fortunately representative democracy still works pretty well even when we outsource the day to day stuff to MPs and vote every so often.