Thanks to all of the responses and comments from the list on this query
In short it appears that ZAR 5th ed is the most favored textbook and "R" was
the most universally confirmed software
--a comment from my colleague suggested that R would require some training on
writing scripts, and may become problematic if students don't have a
programming background
I had 11 responses from the list--so take it for what it's worth
1) What text book(s) would you recommend?
**Zar 5th ed Biostatistical Analysis-- 4 votes a couple of comments
about how ZAR has really outlined the needs of an intro course in the 5th
edition
Whitlock and Schluter's The analysis of Biological data
Glover, T. and K. Mitchell (2008). An Introduction to Biostatistics,
2nd Ed., Waveland Press
Baldi and Moore, the Practice of Statistics in the Life Sciences
Quinn and Keough 2002--Experimental Design and Data Analysis for
Biologist"
Heath's Experimental Design and Statistics for Biology
Ben Bolker's book "Ecological Models and Data in R
Supplemental Texts
Ramsey and Shafer's "Statistical Sleuth"
Scheiner, S.M. and J. Gurevitch. Design and Analysis of
Ecological Experiments. 2nd Edition
Burnham K.P. and D. R. Anderson. Model Selection and
Multi-model Inference.
Peter Dalgaard -"Introductory Statistics with R"
2) What are the core subset of skills/tests do you believe need to be delivered
in a course of this nature? Here are 2 samples I think there is pretty decent
overlap
a. Descriptive Statistics to get started (mean, median, mode, etc.),
variance, SS, standard errors, coefficience of variation.
b. Hypothesis testing (one and two sample, one and two way, parametric
and non-parametric t-tests), Z tests also for populations.
c. Simple Experimental Design and ANOVA (one and two-way factorials,
blocking, nesting, repeated measures, non-parametric ANOVA). I only emphasize
one-way and two-way, blocking and nesting. But I expose them to
other designs, including split-plot, but they don't analyze such designs. That
would be for a second semester follow-up: Experimental Design.
d. Regression (linear, logistic (exposure only), multiple (exposure
only), correlation analysis (parametric and non-parametric).
e. Goodness of Fit, Chi-Square, Contingency Table Analysis.
1. Experimental design****** (this can make or break
adissertation!)
2. Hypothesis Testing**** (you should really ask your
students if they know the difference between a hypothesis and a prediction or
if they know what the null hypothesis of their experiments
are)
3. Statistical Inference (maximum likelihood and bayesian
only)
4. Power (and effect sizes)****
5. Non-parametric stats (Mann-Whitney, randomization tests,
permutation tests, boot strapping)
6. Multivariate: Mult Var Normal distribution, PCA
(especially explain when to use covariance matrix vs when to use correlation
matrix and that SAS by default uses corr while R by default
uses covariace), CDA, DFA, MANOVA, MDS, Decision Trees
And the use of General Linear models (3 votes)
3) what common use statistical software should the students be using?
**R-- 4 votes with a few comments about this being the future
(http://www.r-project.org/ )
SAS
SPSS- 2 votes
Matlab
Vassar stats (http://faculty.vassar.edu/lowry/VassarStats.htm )
Excel- a couple of comments about this not being optimal
Example of Syllabus
http://www.public.asu.edu/~jlsabo/courses.html#biometry
******************************************
Charles R. Bomar PhD
Applied Science Program Director
Executive Director, Orthopterists' Society
Professor of Biology
University of Wisconsin-Stout
Menomonie, WI 54751
[email protected]
office 715-232-2562
fax 715-232-2192
