After these brilliant words about an important topic, a less luminous opinion about something that may be related somewhat. I mean statistical tests. I don't want to be a rebel. I believe this has been talked before and repeated many times. But there I go. Why to make those complicated and time consuming statistical tests when you can simply draw a plot and your conclusion comes? Ok, computers make everything fast. Do they? I need to learn that populations must be normal, they must be homoscedastic, there are at least 3 models for ANOVA, there is something out there with the name of ANCOVA, and I have no single idea if this is useful for me or not. I admit that in some cases statistical tests do help to understand the obtained results, but the path to dominate and understood what is behind is long, and not easy. Therefore, I follow with my faith. I use my software and it gives me the indexes that will allow me or not to do my parametric tests, and then I apply the tests, only to confirm something that I knew weeks ago. Or I learn that my observation is not good because I could not achieve enough power with my test. And then I have the alternative of doing a similar test, but I don't like the idea of learning another test, and then I discover that I need to do other kinds of preliminary tests... wow, maybe you get the point. I know this will lead to nothing, but I would like to say: isn't much better only do the right plots and look at the data? Of course this do opens the doors for ill-done plots, but for me personally it is much better. I believe that in the online statistical manual from NIST the author claims something similar about many plots, that they substitute ANOVA with advantages, for example. A pit that such ideas are not more widespread. They would save some of my time.
Regards, Matheus --- Wayne Tyson <[EMAIL PROTECTED]> escreveu: > Dear Cheryl Heinz and Forum: > > The subject subhead is intended to be only a bit > humorous. > > I respect mathematics, but I don't overrate it. > What I am waiting > for is an equation or a computer program that can > stand up to proof > and predict--describe the phenomenon in terms of > principles--Laws, if > you will. I have to have that in order to respect > the > mathematicians' claims of omnipotence. > > But bean-counters are, of course, in the driver's > seat. They (a > fraction of self-proclaimed mathematicians) have > taken over > ecology. This fraction likes to bully so-called > non-mathematicians > with disdainful sneers about the "non-mathematical" > approaches to > what they have gotten away with calling, without > proof, > "non-science," including ecology. These > number-bullies don't like > inference, chaos, and the like. To be "science," > they say, it must > be reduced to numbers, to (endless, irrelevant) > decimal > points. "Bionumerologists," one old-fashioned > botanist once called > them. One smells a bit of put-down as a means of > feeling bigger. > > Biology and ecology do require disciplined thinking, > and certainly > math is a necessary and useful tool in making sense > out of > observations, but the reductive nature of > mathematics is impotent > when it comes to getting a handle on such a squishy > subject as > ecology. The fact that it is so frustrating to > study ecology, so > endless and without firm conclusions, does not mean > that the human > need to conquer all will necessarily be satisfied. > As my wife says, > "Nature bats last." Unraveling ecology, if that is > ever "done," will > require a kind of "metamathematics," an infinitely > complex array of > integrated principles that simply IS--not a > construction of any > single person, even any team or IT (ironic, eh?). > > Good luck with your calculus--I hope it will prove > me wrong, add more > light than heat. But don't be intimidated. > Everything really is > connected to everything else, and while we should > pursue a better and > better understanding of ecological phenomena, > including by using > mathematics, my forbidden intuition suggests that we > will have to go > beyond math as we now claim to understand it (and > certainly far > beyond reductive statistics) if we want to get > beyond cutting ecology > up into little decimal-pieces and making mere > dissertations out of > them. But Homo doubly-wise has always preferred > self-validated > fantasy to reality, no? Except, maybe, those who > find sufficient > satisfaction in the Quest, who demand no ego-salving > "certainty," > those for whom a significant dose of uncertainty is > no vice, and for > whom outliers can be seen as just possibly where the > cutting edge may > lie. Of course, since burning at the stake is no > longer cool, > certain banishment shall be (has been) their fate. > > WT > > PS: "Fuzzy logic," gets closer to recognizing the > trends and degrees > that make up ecological phenomena than anything else > I've seen in the > region of math, but even that is limited by the fact > that variables > are infinite--or, well, too numberous to count any > way. Ask your > colleagues to "solve" for that. Again, I jest--a > little. > > > At 11:38 AM 7/16/2007, you wrote: > >I'm involved as an ecologist in a project to > develop a two-semester > >biocalculus course and textbook. As a biologist, my > role is in helping to > >write snippets of the biology and help track down > some data sets. (I keep > >trying to explain to my math colleague just how > long it's been since I took > >a math course...) > > > >So, I'm asking the community if you have any > datasets that you would be > >willing to lend that could be modeled using > calculus -- potential topics (on > >the ecology side) include exponential and/or > logistic population growth, > >succession, predation -- and much more. On the math > side, the topics are > >ordinary differential equations, difference > equations, matrix models, > >differential calculus, and more. > > > >Data are only to be used as examples for the > course/ text -- we're thinking > >it would be nice to provide some real-world data > (along with all the > >faults!) instead of simply generating data sets. > (My colleagues in the math > >department -- and the PI for the project -- will > likely be the ones to track > >down permissions as needed.) > > > >Tim Comar (in our math dept) is the PI and lead > author of the text. > > > >I'd be happy to answer questions to the best of my > abilities, and I thank > >anyone who has data to share in advance! > > > >Thanks! > >Cheryl > >-- > >Dr. Cheryl A. Heinz > >[EMAIL PROTECTED] > >Assistant Professor, Biology > >Benedictine University > >(630) 829-6581 phone > >(630) 829-6547 FAX > >http://www.ben.edu/faculty/cheinz/index.htm > Matheus C. Carvalho PhD student Kitasato University - School of Fishery Sciences Japan ____________________________________________________________________________________ Novo Yahoo! Cadê? - Experimente uma nova busca. http://yahoo.com.br/oqueeuganhocomisso
