Given that this is a psychological study with little parameters other than yes or no and a self-selected bias, I imagine this has more to do with the relationships between the responses and/or the quality of the responses rather than the actual choice of response in of itself. Logically, the yes or the no is an answer of anything and everything that a person could think of; so little data can be gathered. For example, you would not be able to feasibly measure openness or optimism (as it was no doubt the thinking of some here) because the nature of the question that the person is answering in their minds could be framed either positively or negatively and you have no way of knowing which. The way I see it, the answer is both yes and no; whichever I choose is completely arbitrary.
Although you might be able to get some insights into how often and how reasonably we explain our thinking...
That's my guess for this little puzzle.