My opinion is also fact, remarkably.

I can support this with data from a probit regression of 'truth' on 'my opinion', using 'day of the week' as an Instrumental Variable estimate. I then calculated the following expression for an estimated impact of my opinion on truth:

beta^IV = [Cov(truth, day of the week)]/[Cov(my opinion, day of the week)]

and found this value to be statistically identical to 1. I countered invariance bias due to error terms systematically equal to zero by ignoring it.

FACT.


Now to work on my statistical model of showing off........ :)