If you can design an elegant experiment that only requires a t-test for its analysis, that is admirable. But the simple truth of the matter, in my experience, is that many folks don't take the time to design a good experiment, often collect data with disregard to any theory, and simply collect what is easy or is the data that everyone else collects, hoping in the end that somehow through mathemagic, they can make something out of their efforts. To paraphrase Burnham and Anderson, 90% of our time should be spent thinking and only 10% doing. I'd suggest folks be aware of theory and design experiments with regard to it, such that the design and analysis are set BEFORE the data are collected. Often, but not always, if that is done, an overly complex analysis may not be necessary...but some complicated hypotheses do require complex analyses. This is the nature of good science.
Mike Sears Assistant Professor Department of Zoology Southern Illinois University Carbondale, IL 62901 email: [EMAIL PROTECTED] web: http://equinox.unr.edu/homepage/msears Quoting stan moore <[EMAIL PROTECTED]>: > Folks -- > > In view of recent discussions of modeling and the correct choice and use of > > them, the correct analysis of assumptions, etc., I am reminded of a > contrarian method of science related to Frederick and Fran Hamerstrom, two > > graduate students of Aldo Leopold, now deceased. Frederick and Fran were > expert ecologists, but of the "old school" in that they practiced a form of > > science that is in danger of extinction. > > The September, 1992 issue of the Journal of Raptor Research (Vol. 26, No. 3) > > was dedicated to the work of the Hamerstroms and edited by two former > "gabboons" of theirs, Drs. Josef Scmutz and Keith Bildstein. An interesting > > letter was published in that issue of JRR, written by Dr. Schmutz, entitled > > "Hamerstrom Science From a Gabboon's Point of View", pp. 206-210. > > Here is a quote from that paper: "To think that only those who employ > up-to-date statistical procedure carry out "good science" is flawed. The > difficulties encountered in the study of complex natural events are so > enormous that even approaches which are considered to be state-of-the-art by > > peers are often insufficient. S.H. Hurlbert concluded that of 176 > experimental studies published between 1960 and 1983, 27% were designed > inappropriately... Perhaps, the message from the Hamerstroms is not to use > > the term "chain" when the strength is equivalent to that provided by a > "string". Much of what is considered "good science" is done not because the > > method warrants it or because a paradigm dictates it, but because it is > consideredc the approach of choice by peers within one's "invisible > college". " > > A couple of years ago, Dr. Steve Herman in an invited paper to the Wildlife > > Society Bulletin, lamented the "lust for statistics" and the distortion > between research and management now so prevalent in the wildlife > profession. > > I admire and enjoy the philosophy of such contrarians! > > > Stan Moore San Geronimo, CA [EMAIL PROTECTED] > ------------------------------------------------- This mail sent through https://webmail.unr.edu
