The OSCTR Biostatistics, Epidemiology and Research Design Training Unit is offering a series of seminars aimed at orienting investigators and research personnel to statistical thinking.

 

The first series of seminars will take place on Mondays, from 4-5 PM in the Oklahoma Clinical and Translational Science Institute conference room in the Harold Hamm Diabetes Center, Suite 2100.

Each one hour session will include a 40 minute lecture demonstration, followed by 20 minutes of time devoted to questions and discussions of topics on which participants want further information.

While the presentations seek to introduce researchers to statistical topics and ways of thinking about analysis, they presume no special knowledge of the field of statistics or research design.

October 6 - Statistical thinking: Quantifying certainty and uncertainty

Research results, because they are founded on samples observed in particular places and times, point somewhat provisionally to the true state of nature.  However, statistical tools can quantify any uncertainty -- including the oft referenced margin of error -- and reinforce the interpretation of results.

October 20 - Statistical significance: Alternatives to the p value

While p values are well known products of statistical analysis, the calculations that produce them will generally also generate confidence intervals, which can be more informative and easier to interpret.

November 3 - Statistical power 1: Brute force approaches

Researchers' first encounters with statistical software, and with statisticians, often involve planning of sample size.  The last two lectures in the series will explore, guided in part by input from participants, key issues related to calculations of power and sample size.

November 17 - Statistical power 2: Artful approaches

 

The OSCTR BERD Training Unit is funded by award U54GM104938 from the National Institute of General Medical Sciences, National Institutes of Health.