EPIB-694   Principles of Inferential Statistics II     June 2004

updated: May 12, 2004

 Instructor

 Dr James Hanley

personal
web-page
.

www.epi.mcgill.ca/hanley/

E-mail

.

James.Hanley@McGill.CA

Course
web page

www.epi.mcgill.ca/hanley/c607/

This course web page will be the distribution medium for all class notes, assignments, resource material, etc . It is the same webpage J Hanley used when he taught the 4-credit c607 in the Fall of 2001, and in earlier summers.

For those who will not have easy access to the web during the month of June, J Hanley plans to provide them with a CD containing all the material on this (and his other) course website(s) as of June 1.


You will need a username and password to access the course website (and other course websites maintained by J Hanley).

The username is c607.

If J. Hanley has not already sent it to you, you can obtain the password by emailing
James.Hanley@McGill.CA
 

Telephone

398-6270

Objectives

To provide consumers/producers of biomedical research with basic principles of statistical inference applicable to clinical and epidemiologic research so that they can (i) understand how statistical methods are used by others (ii) apply them in their own research and (iii) use them as a base for more advanced biostatistics courses.

Content

Introduction to basic principles of statistical inference used in clinical and epidemiologic research. Topics to be covered:

non-parametric methods;
inferences regarding proportions and rates;
correlation and regression
.

Moore and McCabe's
Introduction to the Practice of Statistics, 4th Edition, chapters 8-10 and 14; with supplements.

Dates/Time/
Place

7 days: Tuesday and Thursdays, June 1,3 ; 8,10; 15,17; 22 [no class on 24th].

8-:30-11:45 (i.e., 2 lectures of 90 min each, with 15 min break)

First Class: Tuesday June 1

LOCATION: Room 25, Purvis Hall

Tutorial times: to be determined

Assessment

5 assignments [40%], project [20%], and final (take-home) examination[40%].

Prerequisites

Differential and integral calculus.

No.of Credits

2

Required Text




Other texts

4th edition of Moore and McCabe's Introduction to the Practice of Statistics


This book, along with the others below can be purchased at the McGill Bookstore

>>Statistics in Medicine, Colton, T, Little, Brown and Co., Boston, 1974. 372 pp. Used for 10 years in 607; well-liked; doesn't do ANOVA or statistics for epidemiology. Better on sample size.

>>Statistical Methods in Medical Research Armitage, P, Berry G, and Matthews JNS (4th Ed) Blackwell, Oxford, to be published August 2001. New edition of a standard text on medical statistics. Several sections relevant to epidemiology; includes treatment of 'mid-p' p-values.More advanced; Appreciated more as a comprehensive reference once one has already taken this course.

>>Primer of Biostatistics, Glantz, SA, (4th Ed) McGraw Hill, 1997. 473 pp. Slightly different approach; good diagrams;

>>Principles of Biostatistics, Pagano, M, Gauvreau K (2nd Ed) Duxbury, 2000. 568 pp. used in Harvard SPH. ISBN
: 0534229026

>>Biostatistics in Clinical Medicine, Ingelfinger, JA, Mosteller, F, et al., (3rd Ed) , McGraw Hill, ,1993. 418 pages. Uses management of individual patient to explain concepts. Well presented.

>>Basic and clinical biostatistics, 3rd Edition Dawson-Saunders, B., Trapp, RG. McGraw Hill, 2000. 399 p.
ISBN: 0838505104 Quite good.

>>Biostatistics, The Bare Essentials. Norman, GR and Streiner, DL
,( both at McMaster) ISBN: 1550091239 BC Decker, 2nd Edition 2000. 324 pp. Their first book PDQ Statistics was written "to satisfy consumers of statistics" and went from Intro to multivariate in 160 pages . This new text says "it takes somewhat more knowledge and skill to do something - plumbing, wiring, or statistics - than it does to recognize when others are doing it well or poorly. That then is the intent of this book." Like PDQ Statistics, an irreverent, highly readable text.

Computing

The emphasis will be on hand calculation, in order to understand the principles involved.

However, students may wish to carry some of their calculations using the Excel spreadsheet software or -- if they already know it --a statistical package such as SAS, Stata or SPSS.

An introduction to spreadsheets is available under 'Resources' on the instructor's web page for
course 323. [username c323]


There are somesimple calculators (and very interesting applets) available from the website for the Moore & McCabe textbook
http://bcs.whfreeman.com/ips4e/

Or you may wish to use the slightly more extensive web-based package package
http://www.statcrunch.com/
I do not recommend this for serious research, (and I would always want to check it on a worked example from a textbook before using it) but in the interests of time, it can be useful for purposesof the course.


Here are links to learning resources for SAS and Stata [both in course c681 webpage]

CONTACT

Marielle Olivier (tel. 398 6257) about how to use SAS. SAS and Excel are available in the department's Computer Laboratory in the basement of Purvis Hall, once the student has set up an account on the departmental server with Marielle Olivier.

MORE DETAILS

to come