On Mar 21, 8:40=A0am, Art Kendall <Arthur.Kend...@[EMAIL PROTECTED]
> wrote:
> There clearly is a difference about what is desirable for people who
> will be full time statisticians/methodologists and people who
> concentrate on other disciplines.....
>
>
> It true that many people have magical thinking about computers and rely
> on them too much. =A0Then there are people who use them too little. One
> example. =A0Today it is very easy to do little simulations and draw 1000
> samples of say 10 from a 100,000 case population with a pop percent of
> 25 and graph the distribution of results in a few minutes. =A0These
kinds
> of things can greatly reinforce the theoretic and formulaic aspects of a
> lesson.
>
Computerization has made it much easier to graph the data.
Computerization has made it much easier to find outliers and to easily
test out the implications of those outliers being in/out.
As Art notes above, it's also much easier to do similations /
resampling exercises -- these have the added virtue of emphasizing the
variability of the estimates which are inherent to statistical
estimates.
By and large, computers have been a good tradeoff for statistics.
