Fall 2012 (Wed. 12:30-13:55 at Purvis 24)

September 12, 2012

Dr. Erica Moodie

Models for Longitudinal Data: A Generalized Estimating Equation Approach
Zeger1988.pdf
Adobe Acrobat Document 308.6 KB

Background Reading:

Liang1986.pdf
Adobe Acrobat Document 1.1 MB

September 19, 2012

Dr. James Hanley

Hanley1982.pdf
Adobe Acrobat Document 1.0 MB

Background Reading:

Swets1973.pdf
Adobe Acrobat Document 2.3 MB
Metz1978.pdf
Adobe Acrobat Document 1.8 MB

Background Stories:

"This week is the 32nd anniversary of my introduction to ROC analysis. At home in Natick, Mass, on the evening of Aug 27, 1980, while the intervals between AnnMarie's contractions were getting shorter, I was reading the manual for the FORTRAN program by Dorfman and Alf, as Barbara McNeil has contacted me to ask how to deal with the fact that 1 of the 4 radiology readers in the 'is a radiologist more accurate if (s)he reads CT scan with patient's history in hand vs. if not?' study had not used all 5 rating categories. At 4 a.m. on the 28th, we went to Framingham Union Hospital, where Stephen (already 9 days late, with McGill calling every day to know when was I going to show up to teach EPIB607) arrived that p.m. We arrived in Montreal on Sept. 10th, 1980, so 607 started a bit late.

Barbara gave me two very helpful papers for orietation, the (now very heavily cited) one by Charles Metz,
http://news.uchicago.edu/article/2012/07/10/charles-metz-pioneer-imaging-science-1942-2012
http://radiology.uchicago.edu/page/metz-roc-software
who died this summer, and one by John Swets, whom I consider the founder of ROC analysis, and whose book
http://metz-roc.uchicago.edu/MetzROC/publications/review-of-john-swets-new-book
I must get and read.

I attach both of these papers as background reading, along with (as the primary reading) my own 1st roc paper, which I developed in 1980-1981 while I was still relatively unknown and not yet involved in that many activities at McGill. For the computations, carried out in FORTRAN, I did not have any grant money to pay the mainframe computing bills, so I used some of the small amount of computing money I was allocated for teaching 607 that Fall, and 621 (data analysis for epi students, Winter '81) while the usual teacher (Stan Shapiro) was on sabbatical in Boston in 80-81. Programming steps and data were entered on punched cards.

Some reasons that 1st roc paper of mine became a citation classic: it was self contained; it had a worked example; it was distribution-free and was easy to explain to, and be used by radiologists; it came at a time when VisiCalc (first spreadsheet, developed at MIT for repetitive 'what if' calculations in case studies in business plan projections; forerunner to Quattro, and ultimately, when Bill Gates saw how much $$ he could make form it, Excel) was becoming popular on Apple II personal computers and liberated users from the mainframe and FORTRAN.."

 

Jim, 2012-08-31

September 26, 2012

Please note that at the beginning of this meeting, Professor Andrew Piper, from the Department of Languages, Literatures, and Cultures, has been invited to briefly talk about his work (http://literarytopologies.org/) and how statistics can be useful in his area of research.

Luc Villandré

Episodic Sexual Transmission of HIV Revealed by Molecular Phylodynamics
Lewis2008.pdf
Adobe Acrobat Document 532.4 KB
Presentation
Villandre2012.pdf
Adobe Acrobat Document 271.1 KB

October 3, 2012

Dr. Robert Platt

Quantifying discrimination of Framingham risk functions with different survival C statistics
Pencina2012.pdf
Adobe Acrobat Document 647.0 KB
Pencina2010.pdf
Adobe Acrobat Document 1'003.1 KB

Background Reading:

Pencina2008.pdf
Adobe Acrobat Document 201.7 KB

October 10, 2012

Dr. David Stephens

Maximum Likelihood from Incomplete Data via the EM Algorithm
DLR1977-JRSSB.pdf
Adobe Acrobat Document 4.3 MB
The EM Algorithm--An Old Folk-Song Sung to a Fast New Tune
MengVanDyk1997-JRSSB.pdf
Adobe Acrobat Document 5.8 MB
Presentation
Stephens2012.pdf
Adobe Acrobat Document 112.6 KB

October 17, 2012

Dr. William Astle

Common SNPs explain a large proportion of the heritability for human height
Yang2010.pdf
Adobe Acrobat Document 640.8 KB
Presentation
Astle2012.pdf
Adobe Acrobat Document 7.8 MB

Background Reading:

Manolio2009.pdf
Adobe Acrobat Document 241.6 KB
Pritchard2002.pdf
Adobe Acrobat Document 256.8 KB

October 24, 2012

Dr. Russell Steele

Multiple imputation: a primer
Schafer1999.pdf
Adobe Acrobat Document 280.9 KB
Review of inverse probability weighting for dealing with missing data
Seaman2001.pdf
Adobe Acrobat Document 242.2 KB
A comparison of multiple imputation and doubly robust estimation for analyses with missing data
Carpenter2006.pdf
Adobe Acrobat Document 127.6 KB

October 31, 2012

Dr. Olli Saarela

Spearman and the origin and development of factor analysis
Bartholomew1995.pdf
Adobe Acrobat Document 700.0 KB
Presentation
Saarela2012.pdf
Adobe Acrobat Document 955.9 KB

November 7, 2012

Dr. Benjamin Rich

Estimating treatment effects for individual patients based on the results of randomised clinical trials
Dorresteijn2011.pdf
Adobe Acrobat Document 597.0 KB
Presentation
Rich2012.pdf
Adobe Acrobat Document 727.5 KB
Appendix1.pdf
Adobe Acrobat Document 22.3 KB
Appendix2.pdf
Adobe Acrobat Document 20.3 KB
Appendix3.pdf
Adobe Acrobat Document 29.7 KB

Background Reading:

Julien2008.pdf
Adobe Acrobat Document 162.3 KB

November 14, 2012

Dr. Andrea Benedetti

Using generalized additive models to reduce residual confounding
Benedetti2004.pdf
Adobe Acrobat Document 245.1 KB
Tutorial in biostatistics: spline smoothing with linear mixed models
Gurrin2005.pdf
Adobe Acrobat Document 303.9 KB

November 21, 2012

Dr. Aurélie Labbe

 

Q&A: Promise and pitfalls of genome-wide association studies
Brookfield2010.pdf
Adobe Acrobat Document 370.3 KB
Learning from our GWAS mistakes: from experimental design to scientific method
Lambert2012.pdf
Adobe Acrobat Document 66.6 KB
Presentation
Labbe2012.pdf
Adobe Acrobat Document 1.4 MB

November 28, 2012

Dr. Celia Greenwood

Controlling the false discovery rate: a practical and powerful approach to multiple testing
Benjamini1995.pdf
Adobe Acrobat Document 244.8 KB
Presentation
Greenwood2012.pdf
Adobe Acrobat Document 884.7 KB