If the polls are to be believed, the yes-no gap in the impending Scottish referendum appears to be narrowing. Come a close-run thing in September, every millimetre of advantage will count. From a marketing perspective, therefore, I’m surprised we’ve not heard more from the MR fraternity.

Most marketers get briefed at some point in their careers on the subject of bias in questionnaires (how to avoid it, and – cynically – how to use it to your advantage).

The referendum question, despite the best efforts of Westminster, seems well-placed to take advantage of the phenomenon known as acquiescence bias – that is, the human tendency to agree with questions. Should Scotland be an independent country?

Of course, a more common type of response bias can occur when the choices to a question are not randomised. All other things being equal, the first option tends to benefit. It seems we naturally associate the first choice with the correct choice.

If you wanted the most reliable outcome in brand research, you’d randomise the options – binary ‘ballot’ papers would alternate yes-no with no-yes.

A possible solution to this would be to provide a single space for the answer. However, the Yes Campaign might argue that No is easier to write (unless in the vernacular, in which case Aye has the same number of letters as Naw).

Clearly, ticks and crosses can look remarkably similar, and must be discounted – but this notion raises a real issue for a public that has been trained to tick online boxes on a daily basis. If I want to vote Yes do I put a tick (= ‘agree’) in the Yes box and reserve my cross (= ‘disagree’) for the No box?

I know what I meant but I’ve just spoiled my paper.