In Question 14, the census did not invite anyone to tick a box labelled “Scottish only”.
It invited us instead to answer the question: “What do you feel is your national identity?” by marking at least one of the six boxes on offer, one of which read “Scottish” (top of the list) and another (placed fifth in the list) “British”.
I wonder just how many of those who filled in this questionnaire took any notice of the instruction to “Tick ALL that apply”, and instead, having marked “Scottish” in the first box on the top line , moved on to answer Question 15.
I wonder too what would have been the outcome if the “British” box had been placed first on the list instead of fifth, and the “Scottish” box fifth instead of first.
Would Question 14 have produced the same outcome?
I for one doubt it very much, judging by the way that many people set about reading and filling in forms.
What a pity the question asked was not in a form such as, for example: “Which of the descriptors below most closely describes how you feel about your national identity?” followed by boxes labelled, in this order, “Scottish only”, “Scottish and also British”, “Welsh only” “Welsh and also British” etc, with the penultimate box labelled “British only”, followed by the last box “Other, please write in” (as appeared on the form).
Questionnaires are tricky things to put together: compilers do not always manage to ask the question they really intended to ask, and may fail to recognise that fact, which then runs the risk of making their interpretation of the answers subject to error.
It appears to me that there is at least a possibility that this is what has happened on this occasion.