How much does the referendum question matter?
Lorraine Murray explains why it matters that the referendum question doesn’t matter
Download the article as a PDF 
In the Autumn of 2014, we will be asked what is, perhaps, the most important question in our democratic history. The wording probably won’t affect the outcome. We all know what the question is about. And most of us already know what our answer will be.
So if the question wording is unlikely to affect the outcome, does it matter? Yes. It matters that arguments about the wording of the question do not detract from the debate about the real issues. And it matters that people have faith in the democratic process and accept the outcome as being a fair reflection of the will of the Scottish people. It matters that the question doesn’t matter.
For a referendum question to work it must pass two tests. Firstly, voters must understand what they are being asked and find it easy to answer the question in a way that matches their views on the issue. Secondly, voters must perceive the question to be neutral and not to encourage a particular response. Thorough testing with voters is the only way to determine whether a question works.
Note that testing can only examine the perceptions of the neutrality of a question. There is no way to test whether a referendum question actually is neutral. Large scale surveys of the population could be used to test different versions of a question and see which wording was more likely to lead to a particular result and how much difference it was likely to make. But there is no way of knowing the ‘true’ result because all we can do is ask the question in different ways – all of which are potentially flawed or biased in some way. It’s true that a version that led to a wildly different result from other versions (or from recent polling on the issue) would seem suspect, but assuming that the wording isn’t blatantly leading and makes a relatively small difference, we would have no way of knowing which of our versions came closest to the ‘real’ level of support/opposition.
The question recommended by the Electoral Commission and accepted by the Scottish Government - “Should Scotland be an independent country?” - has been thoroughly tested with voters. People understand what they are being asked. The question does its job. It can do no more. What’s needed now is more information and debate on the real issues so that Scotland can better decide how it wants to answer.
This article originally appeared in The Times
Further details about our testing of the referendum question are available here
In the Autumn of 2014, we will be asked what is, perhaps, the most important question in our democratic history. The wording probably won’t affect the outcome. We all know what the question is about. And most of us already know what our answer will be.
So if the question wording is unlikely to affect the outcome, does it matter? Yes. It matters that arguments about the wording of the question do not detract from the debate about the real issues. And it matters that people have faith in the democratic process and accept the outcome as being a fair reflection of the will of the Scottish people. It matters that the question doesn’t matter.
For a referendum question to work it must pass two tests. Firstly, voters must understand what they are being asked and find it easy to answer the question in a way that matches their views on the issue. Secondly, voters must perceive the question to be neutral and not to encourage a particular response. Thorough testing with voters is the only way to determine whether a question works.
Note that testing can only examine the perceptions of the neutrality of a question. There is no way to test whether a referendum question actually is neutral. Large scale surveys of the population could be used to test different versions of a question and see which wording was more likely to lead to a particular result and how much difference it was likely to make. But there is no way of knowing the ‘true’ result because all we can do is ask the question in different ways – all of which are potentially flawed or biased in some way. It’s true that a version that led to a wildly different result from other versions (or from recent polling on the issue) would seem suspect, but assuming that the wording isn’t blatantly leading and makes a relatively small difference, we would have no way of knowing which of our versions came closest to the ‘real’ level of support/opposition.
The question recommended by the Electoral Commission and accepted by the Scottish Government - “Should Scotland be an independent country?” - has been thoroughly tested with voters. People understand what they are being asked. The question does its job. It can do no more. What’s needed now is more information and debate on the real issues so that Scotland can better decide how it wants to answer.
This article originally appeared in The Times
Further details about our testing of the referendum question are available here
More insights about Public Sector
