Should We Trust the Polls on Whether Scotland Should Leave the U.K.?

On Thursday (Sept. 18th) residents of Scotland will vote on whether they should break away from the United Kingdom.  Today polls show the vote is too close to call. Moreover, the polls are much less accurate than usual.  Why are these particular polls so untrustworthy?

The pro-independence forces are to be congratulated on running an effective campaign.  One of their most effective and earliest victories happened in June 2013 when they passed in Parliament a bill called the “Scottish Independence Referendum (Franchise) Act.”

This act which expires at the end of this year gives 16 and 17 year olds the power to vote for or against independence.  Until now a person in Scotland needed to be 18 years or older to vote, but for this special election the pro-independence forces were able to allow more teenagers to cast a ballot.

Teenagers, especially those around 16, are rebellious.  They rebel against their parents, teachers and any sign of authority.  Adding these teens to the voting rosters is a brilliant move by the pro-independence party since many of the the extra voters will eagerly embrace the cause of throwing off 300 years of British rules.

Ohio State helps run the National Longitudinal Surveys (NLS) which tracks over time the lives of tens of thousands of people as they transform from teenagers to adults.  One of the most important lessons I have learned from working on these surveys is that tracking and polling teenagers is even more difficult than tracking and polling adults.

The Franchise Act added about 125,000 16 and 17 year olds to the voting polls.  It is difficult to contact these teens.  They keep very different hours, use different methods of communicating and are even more susceptible to sudden changes in opinion than older teenagers and adults.

Political polls in general are not very good because polling organizations use the smallest possible sample size to maximize the speed with which results can be released. Small sample sizes mean the possibility of the survey producing an erroneous result is magnified.

Long time readers of this blog know that I am terrible at predicting the future, so I will not guess at the referendum’s result.  However, my belief is that while polls show a toss-up today, the final result on Thursday will be much more decisive.  The reason will be in large part because 16 and 17 year olds were added to the actual election rosters, but not added effectively to the polls tracking the feelings and expected actions of likely voters.

Leave a Reply

Your email address will not be published. Required fields are marked *