RE: Teaching Bayesian inference in 2 hours

From: Lee Corbin (lcorbin@tsoft.com)
Date: Thu Jan 30 2003 - 22:54:36 MST


Rafal confides

> Damien Sullivan wrote:
> > Longish, but can be summarized as saying people understand
> > frequencies better than probabilities. Higher immediate training
> > effect, and much higher stability (i.e. they can still solve problems
> > 15 weeks later.)
> >
> http://www.apa.org/journals/xge/press_releases/september_2001/xge1303380.html
>
> ### This is very interesting - I have always translated Bayesian problems
> into natural frequencies before trying to plug the data into Bayes' rule.

So you say that you use natural frequencies (e.g. "eight in ten") rather
than using probabilities ".8" or "80%", and you believe that this affects
your thinking?

I have been very interested in probability and probability problems for
more decades than is pleasant to contemplate. I would have supposed
that for me there could be no possible difference. But if it's really
true that there is a difference, I naturally would like to change my
mental habits.

Do you think that this holds also for people used to probability
problems?

Thanks,
Lee



This archive was generated by hypermail 2.1.5 : Sun Feb 02 2003 - 21:26:04 MST