Current Research > Methods Core > CAPS Methods Core Seminars

Methods Core - Seminars

Please note: All seminars take place at CAPS in the McKusick Conference room, unless otherwise noted. Directions to CAPS.

See materials from past seminars.

Quantitative Methods

 

Friday, Mar 19, 2010, 10:00-11:30 AM

  • Power Estimation for Clustered Sample Designs
    Steve Gregorich, PhD, Associate Professor, UCSF Dept. of Medicine / DGIM
When proposing clustered sampling designs, sample size estimates can often be obtained by first assuming a simple random sampling (SRS) design and then adjusting the result by one or more design effects. Some types of design effects can be estimated with information about cluster size and/or intra-cluster correlation of responses. However, selection of an appropriate design effect estimator requires consideration of the intra-cluster correlation of both X and Y variables. Time permitting, we will also consider design effects resulting from unequal cluster sizes and unequal selection probabilities.

 

Friday, Apr. 16, 2010, 10:00-11:30 AM

  • Journal Club: Discussion of an article by Meng in the American Statistician, with Discussion/Comments
    Special Section: Opportunities and Challenges for the Discipline
    “Desired and Feared—What Do We Do Now and Over the Next 50 Years?”

In the August 2009 issue of the American Statistician, there appeared an invitation from the Editor for an article by the name listed in double quotes above. The author of the main article is Xiao-Li Meng, of Statistics at Harvard. In response to that article, the most recent issue of the journal published 7 short articles, as well as a rejoinder by Meng.

We would like to discuss the collection of articles in our April meeting. The collection of articles is available here. If you have your own copy of The American Statistician, you might want to look at the articles directly.

The discussion will be much productive if participants read the articles ahead of time, and volunteer to present a point of view for one of them. If you are willing to present a short paper, please let Estie know via email. Although geared at the teaching of Statistics, the articles have implications for how we, as statisticians, deal with PIs and others consumers of our statistical labor.

Qualitative Methods

 

Past seminars

Qualitative

Quantitative