Robert George Douglas Steel

I just noticed this "Steel and Torrie" biometrical statistics book in my "to-read" pile and thought..."where the heck did that come from?" I remembered that I had lost my original copy ...probably purchased in the mid 60's .....when rain damaged my personal effects on the dock in Acapulco. And I remember searching high and low for another copy years afterwards. Obviously I found another copy though this is a paperback and more compact and I preferred the larger original version. However, this is a really nice stats book and I used it extensively in my later work as an Agronomist. I remember being particularly struck by Figure 10.6 which shows that a single standard deviation does not apply to all regressed values of Y but must depend on an X value that determines the Y population. All sounds very complex but effectively you can plot confidence lines for a regression equation (and these are curved). I used this in one of my published papers. I noticed that there is a chapter on non-parametric statistics and I confess that I know nothing about this. ...must find out more. Just going back and re-reading some of this text makes me realise how deficient I am now in basic statistics. I would certainly benefit from re-reading this book. Will I do it? Hmm.....life is short and opportunity fleeting and of making many books there is no end. ....So probably not. But It's good to know I have it there on my shelves for when I just need to do a two way analysis of valence. And I think stats has moved on since I studied this book...especially with computer aided analysis. Seems to me that these days you can just enter your data and run programs which will analyse and torture the data to give whatever is desired. (Not like we had to do with hand-cranked calculators!!....it's true!) Four stars from me.… (más)