140.614.13 Data Analysis Workshop II
- Winter Inst. term
- 2 credits
- Academic Year:
- 2012 - 2013
- East Baltimore
- Mon 01/14/2013 - Fri 01/18/2013
- Class Times:
- M Tu W Th F, 8:00am - 12:00pm
140.613; students must have a notebook computer with Stata 8 software installed.
Intended for students with a broad understanding of biostatistical concepts used in public health sciences who seek to develop additional data analysis skills. Emphasizes concepts and illustration of concepts applying a variety of analytic techniques to public health datasets in a computer laboratory using Stata statistical software. In the second workshop (140.614), students will master advanced methods of data analysis including analysis of variance, analysis of covariance, nonparametric methods for comparing groups, multiple linear regression, logistic regression, log-linear regression, and survival analysis. Enrollment limited: students must have a laptop computer with Stata 11.0 installed.
- Learning Objectives:
- Use STATA to visualize relationships between two continuous measures
- Use STATA to fit simple linear regression models, and interpret relevant estimates from the results
- Use STATA to fit multiple linear regression models to relate a continous outcome to multiple predictors in one model and to help assess confounding, interaction, and goodness-of-fit
- Interpret the relevant estimates from multiple linear regression
- Use STATA to graph lowess smoothing functions to relate the probability of a dichotomous outcome to a continous predictor
- Use STATA to fit multiple logistic regression models to relate a dichotomous outcome to multiple predictors in one model and to help assess confounding, interaction, and goodness-of-fit
- Set up cohort study data into STATA survival analysis format
- Use STATA to graph Kaplan-Meier curves and perform log-rank tests
- use STATA to fit Cox regression models to relate time-to-event data to multiple predictors in one model and to help assess confounding, interaction, and goodness-of-fit
- interpret the confounding estimates from Cox regression
- Methods of Assessment:
Student evaluation based on laboratory exercises, an exam, and completion of an independent data analysis project.
- Instructor Consent:
Consent required for some students
- Consent Note:
Instructor consent required for registrants not enrolled in a degree program currently with enrollment in the Winter Institute.
- For consent, contact:
- Special Comments:
Instructor consent required for registrants not concurrently enrolled in a JHSPH part-time degree program