Enneagram Main Board Archive

I suppose that it's rant time, again.

I suppose that it's rant time, again.

[ Follow Ups ] [ Post Followup ] [ Enneagram Message Board ] [ FAQ ]

Posted by Peter on January 20, 1999 at 12:32:40:

In Reply to: Re: 8w5 INTJ ?.....HERE....LEMME ASK YOU SOMETHING posted by Walter on January 20, 1999 at 10:22:42:

> These are some possible explanations. They show how important it is to have a quantitative tool such as RHETI to settle the argument by testing a hypothesis with numerical data. This allows the scientific method to be used to support or refute an explantation.

>
> Walter
>

Quantitative tools, eh? Problem is that quantitative tools aren't perfect. The maths may be, but to get the maths to apply to reality, then you need to make assumptions - and if those assumptions are off, then so are all of the bets...

Your RHETI, for instance. It's scores vary quite a bit when I take it, depending on what mood I'm in - some of the questions are very hard to call. And, it does mistype people every now and again. But, lets be fair, call that genuine random error. Do large enough statistical anlyses (and considering the the level of complexity you want to go into, large means LARGE), and they should all drop out.

But what of inbuilt biases? I've heard from others on the board that the RHETI regularly mistypes 5s as 4s. I can certainly belive it myself - whenever I do the RHETI my 4 score is almost as high as my 5 score, and I'm definitely much more of a 4 than a 5. Whatever this is, just by looking at the way the tests are constructed, its easy to see how all sorts of biases could occur.

Now statistical analysis won't neutralise those systematic errors - in fact it will intensify them. Do a large enough experiment, and you don't find a subtle truth about human nature, you find a subtle truth about your quantitative tool. Especially with the human psyche - there are just so many different interacting factors that _any_ quantitative study is going to be dripping with artefacts...

Oh, and two words. 'Predictive value'. If your theories can make any set of results explain any other set of results, then you can throw any hope of finding empirical justification for your theory in the bin.

Don't get me wrong - quantitative methods _are_ very useful - I use them in lab classes every week. Still, you have to be careful with them, or you can come to all sorts of unjustified conclusions.

Peter, 5w4, INTP

Follow Ups:

I agree with Peter's 'rant' Dafyd 09:31:17 1/24/99 (0)

Post a Followup

Name:
E-Mail:

Subject:

Comments:
> > > These are some possible explanations. They show how important it is to have a quantitative tool such as RHETI to settle the argument by testing a hypothesis with numerical data. This allows the scientific method to be used to support or refute an explantation. > > > > Walter > > > Quantitative tools, eh? Problem is that quantitative tools aren't perfect. The maths may be, but to get the maths to apply to reality, then you need to make assumptions - and if those assumptions are off, then so are all of the bets... > Your RHETI, for instance. It's scores vary quite a bit when I take it, depending on what mood I'm in - some of the questions are very hard to call. And, it does mistype people every now and again. But, lets be fair, call that genuine random error. Do large enough statistical anlyses (and considering the the level of complexity you want to go into, large means LARGE), and they should all drop out. > But what of inbuilt biases? I've heard from others on the board that the RHETI regularly mistypes 5s as 4s. I can certainly belive it myself - whenever I do the RHETI my 4 score is almost as high as my 5 score, and I'm definitely much more of a 4 than a 5. Whatever this is, just by looking at the way the tests are constructed, its easy to see how all sorts of biases could occur. > Now statistical analysis won't neutralise those systematic errors - in fact it will intensify them. Do a large enough experiment, and you don't find a subtle truth about human nature, you find a subtle truth about your quantitative tool. Especially with the human psyche - there are just so many different interacting factors that _any_ quantitative study is going to be dripping with artefacts... > Oh, and two words. 'Predictive value'. If your theories can make any set of results explain any other set of results, then you can throw any hope of finding empirical justification for your theory in the bin. > Don't get me wrong - quantitative methods _are_ very useful - I use them in lab classes every week. Still, you have to be careful with them, or you can come to all sorts of unjustified conclusions. > Peter, 5w4, INTP

Optional Link URL:
Link Title:
Optional Image URL:

[ Follow Ups ] [ Post Followup ] [ Enneagram Message Board ] [ FAQ ] wwwboard/messages/4011.html

Your Ad Here