Haha, BK Marcus posted this the other day and I couldn’t help but laugh because I am so headstrong when it comes to apriori versus aposteriori.
For instance, during my oral exam for my masters, the committee chairman asked me what I thought of social sciences in general — if I had changed my mind about Likert scales. I told him that despite all of the statistics classes and various case-studies we had dissected, interval methodology was invalid for anything but anecdotes in social sciences.
For example, I thought the most recent episode of It’s Always Sunny In Philadelphia was 4.5 out of 5. You think it is also 4.5 out of 5. However, the problem is, our values are entirely subjective and cannot be averaged or quantified as if it were one concrete opinion. Sure sites like Amazon or NewEgg usually create useful generalizations with user ratings, but you can’t use the interval scale as an ironclad law because subjective minded humans do not have the same internal preference scales.
It’s one of the reasons that Quant funds on Wall Street were decimated this past year because they attempted to quantify human action (or in their case, artificial incentives) using inherently incomplete equations — using false assumptions to fill in for a continually changing series of individual preferences (or in their case, risks/demand).
See also: Chapter 1 of Human Action