Friday, November 20, 2009

Design for Validity

Content:

In Dr. Petitti’s last lecture for study design, she introduced the descriptive studies, observational studies and experimental studies. And this week Dr. Petitti talked on the validity for the study design. One example is about the occurrence of breast cancer with being teaching BSE or not. It helped us refresh the knowledge about research design and since Dr. Petitti mentioned the sample size of this research and how researcher conducted it, I think it is also a good example for understanding the validity. The internal validity is about the study itself and the external study is about the application of the achievement of the study in the future. And another example of the Coronary Heart Disease is to explain the risk ratio and odd ratio which are the measurement of the outcome. The RR and OR work for the binary data and from the perspective of time, there is hazard ratio. Dr. Petitti spent much time on type 2 statistical error. It is a common threat to the validity of study. Because we do not know how big the sample is could be suitable for the study. And carefully choosing the subjects and power analysis could avoid the type 2 error. What’ more, the randomization is essential in the Epidemiology study.
Posted by
Xiaoxiao

No comments:

Post a Comment

Gentle Reminder: Sign comments with your name.