Past Meetings Spring 2000
March 29, 2000
Title: Inflation of the Type I Error Rate when both Additive and
Multiplicative Interactions are Evaluated in Case-Control Studies
Speaker: Aimee Loarr
Summary:
March 29, 2000
Title: The Effects of Attrition on Conclusions Drawn from Longitudinal
Data
Speaker: Jackie Starr
Summary:
April 5, 2000
Title: Mixture Models for Genetic Changes in Cancer Cells
Speaker: Manisha Desai
Summary:
April 12, 2000
Title: Self-Designing Optimal Group Sequential Clinical Trials
Speaker: Chau Thach
Summary:
April 26, 2000
Title: Insights on Bias and Information in Group-Level Studies
Speaker: Lianne Sheppard
Summary:
May 10, 2000
Title: The Analysis of Cytogenetic Data to Detect Chromosomal Abnormalities
Associated With Cancer
Speaker: Scott Emerson
Summary:
May 17, 2000
Title: The Lymphoma Prognostic Index and Alternatives
Speaker: Michael LeBlanc
Summary:
May 24, 2000
Title: Accounting for Selection Bias in Observational Studies
Speaker: Rebecca Koperski
Summary:
Depression is a common mental illness and it is unclear what effect
a mental health specialist has on improving depression symptoms. A main
problem for this confusion is the fact that observational studies experience
selection bias; those with more mental health specialist contact are sicker
to begin with. The ideal situation would be to carry out a randomized clinical
trial. If randomization is not possible, then it is important that correct
results be obtained by these observational studies. The purpose of my thesis
is to attempt to account for selection bias using various statistical methods
in order to show the true effect of specialist referral on depression outcomes.
In addition to analysis on actual data from the Physician Referral Study,
a computer simulation was carried out in order to assess the performance
of these methods. This talk will discuss the statistical methods used and
present my findings.
Last Modification: 08 December 2000