Sex ratios and temperature

[Tph]

What, not one of you is using good old spss? being the most expensive it must be the best ...

 

[rr16]

I've used SPSS but only to enter data into! Had never been trained in it or got to the stage where I had to do any dreaded statistical analysis!

 

[wethumbs]

SPSS was originally designed for social science where there are large data sets available. In most science and engineering, typically it is expensive and time consuming to run experiments and data sets are well controlled and limited therefore SPSS would not be applicable. The whole idea is to run the least amount of experiments to come up with the conclusion, that's what DOE is for.

 

[rr16]

Do people not just design their own experiments with logic and reasoning these days or have the computers really taken over?

 

[dw1305]

Hi all,
From my experience most university Biologists are now using "R", and this is true of an ever growing number of Psychologists etc. as well. Last week there were over 6000 contributed packages for R, as well as a huge amount of on-line and print literature.

cheers Darrel

 

[gerald]

Experiments MUST be designed with good old logic, reasoning, and clearly stated assumptions. Computer programs just make the statistical data interpretation easier. They certainly dont ensure that the experimental design and assumptions are valid. It's frighteningly easy for someone with a weak understanding of experimental design and statistics to misapply a computer program and make all sorts of unfounded proclamations about what is or isn't "statistically significant". It's a dangerous tool in the wrong hands, and MANY scientists will admit that statistics was never their best subject in school.

Do people not just design their own experiments with logic and reasoning these days or have the computers really taken over?

 

[rr16]

I hated statistics. We did a module at university and I just couldn't get my head around it. The only reason I used SPSS was because somebody had created a patient database in it and I had to extract the information from it to go alongside my experimental results. Unfortunately that work never got finished and that was quite some time ago. I've been out of the experimental field for quite some time now. I've always looked at statistics as a tool I would learn to use upon the advice of people with more understanding of it.
Chi squared and Spearman's Rank are about my limits!

 

[dw1305]

Hi all,
I'm definitely not a statistician, but I know enough to tell that a lot of published literature was carried out, and peer reviewed, by Scientists who didn't understand the implications of their results.

If any-one wants a non-statistics statistics book read I'd recommend David Salsburg's 2002 book <"The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century"> and another favourite is Stephen J. Gould's 1996 <"The Mismeasure of man">.

cheers Darrel

 

[Mike Wise]

I always called my statistics courses 'sadistics'. They were for me anyway.