Practical information for Biostats Course VHM 802 at AVC -
Winter Semester 2021
Course contents and aims
From the Calendar for UPEI/AVC:
"The course covers linear and logistic models, i.e. multiple linear and logistic
regression and analysis of variance procedures for analysis of continuous and dichotomous
outcomes with respect to multiple factors of explanatory variables. In addition,
the course gives an introduction to experimental design and to analysis of data from
complex experimental designs with multiple levels of variation or repeated measurements.
The course is partially taught in conjunction with VHM 812."
- From the course announcement:
- - Course content overview:
- regression analysis (multiple linear and logistic regression models),
- experimental design and multifactorial analysis,
- methods for dealing with clustering,
- multivariate analysis.
- - Detailed course topics (keywords):
- simple linear and polynomial regression,
- multiple linear regression and linear models,
- logistic regression and generalized linear models,
- model building, selection and assessment of confounding as well as
regression diagnostics for linear/logistic models,
- multifactorial analysis of variance,
- experimental designs: blocking versus replication,
complete/incomplete designs, Latin squares, cross-over designs,
- methods for dealing with clustering in continuous and discrete data,
- topics of multivariate analysis: multivariate inference, principal components
and factor analysis, multivariate distance and cluster analysis, classification,
canonical correlations and ordination.
Course aims:
- to provide an understanding of and practical experience with the
statistical models covered in the course, to the level where the student
is able to apply the models and methods and to critically assess the results,
- to guide the student through a case study of a analysing statistically
a real dataset (preferably the student's own data), reporting
and presenting the results (in a seminar).
Course organization (VHM 802 part)
- Lectures:
- presentation of new material (with brief computer demonstrations),
- Labs:
- discussion of lectures and a prepared exercise,
- tutorial of statistical software,
- individual work on additional exercises using computers,
- Project:
- analyse data using methods in course, write report in the
form of a manuscript (plus detailed explanation of statistical
analysis), and present work in a seminar,
- Homework:
- expected preparation of one exercise for most lab sessions,
- project work and home assignments (see below),
- Examination:
- 40% home assignments: 6 assignments (4 without regression component), marked as passed/not passed,
- 30% project work (report and seminar),
- 30% final take-home exam.
Course material
- main textbook for multivariate component is: Bryan Manly and Jorge Alberto: Multivariate Statistical Methods: A Primer, 4th ed, Routledge/CRC Press, ISBN 1498728960; the 3rd edition will do as well.
- additional main textbook is: Gary Oehlert: A first Course in Design and Analysis of Experiments, 2000, W.H. Freeman,
ISBN 0716735105; the text is out of print, but can be downloaded for free.
- additional main text is Chapters 14-16 of:
Dohoo IR, Martin SW, Stryhn H, Veterinary Epidemiologic Research, 2nd ed., 2009. VER-Inc, Charlottetown, Canada.
Regression, (alternatively the corresponding material in: Dohoo IR,
Martin SW, Stryhn H, Methods in Epidemiologic Research, 2012); students
in VHM 802 only will be provided copies of these chapters.
- course material at web page: lecture overheads, solutions to
exercises and assignments, etc.
Course software
Main statistical software packages for the course are Minitab (version 18) and Stata (version 16).
Additional packages: SAS (version 9.4) and R (version 2.12 or later).
Henrik Stryhn
(hstryhn@upei.ca) 2020-12-21