STAT 6020

Introduction to Biostatistics

Introduction to Biostatistics is a course on contemporary and medical research methodology for biostatistics. We will discuss descriptive and inferential statistics including parametric and nonparametric hypothesis testing methods, sample size, statistical significance and power, survival curve analysis, relative risk, odds ratios, chi square modeling, and analysis of variance. Data will be analyzed using statistical software.

Course Syllabus

Textbook:  Practical Statistics for Medical Research, Altman.  Chapman and Hall, 1999.   ISBN 0412276305

Date Assigned	Date Due	Homework
1/15/09	1/22/09	Download the R statistical language.  http://www.r-project.org/ .  Click on the “CRAN” link under "Download" on the left of the screen, choose any mirror site, then follow directions to download and install R.  (Choose the base package and then download and run the .exe file.) Load the babyboom data. Email me with the answer to the following questions: What is the current version number of R? How many minutes past midnight was the 30th child born, and what gender was that child? p103 # 1, 2
1/22/09	1/29/09	p. 71 # 2, 4, 5, 6, 8
1/29/09	2/5/09	Answer the questions on this worksheet.  You can either print the worksheet and bring the answers to class, or you can copy-paste into email.  p 177 # 1, 2
2/5/09	2/12/09	1. A researcher has tested a new treatment for poison ivy against an old treatment. He concludes that “the new treatment is more effective than the old, with a p-value of 0.047.” What does the p-value mean in this context? 2. a. In a previous homework problem, you found that, if 15% of all pregnancies result in miscarriages, the probability that more than half a group of ten pregnant women would have a miscarriage was 0.0013. If, during a hypothesis test to see whether the true percentage of miscarriage really is 15%, we observed more than 5 miscarriages in a group of 10 women, what would be our conclusion? b. In the same problem, you found that if 20000 clusters of 10 pregnant women are examined, we would expect to observe 26 clusters in which more than 5 had a miscarriage, if the true proportion is 15%. Relate these numbers to the probability of Type I error for the above hypothesis test. 3. p 225 # 9.4
2/12/09	2/19/09	p 223 #9.1, 9.6, 9.7
3/19/09	3/26/09	Download the temperature data below. 1.  Use a one-way ANOVA to test whether average temperatures for men are the same as for women.  Comment on whether another test could have been used. 2.  Create a scatterplot for body temperature compared to heartrate and calculate the correlation between the two values.  Perform a linear regression to predict temperature given heartrate.  Give the model and describe how the slope of the model can be interpreted.  Then comment on statistical significance vs. actual significance in light of this example.
4/2/09	4/9/09	p 474  # 15.2, 15.3, 15.4, 15.5abc
4/9/09	4/16/09	p 438 # 14.4
4/23/09	5/7/09	Files necessary for final exam: Tryptone.dat.txt tryptone.txt lifespan.txt survival.txt

R code for distributions

R code for histograms

temperature example
temperature data

tryptone data and description

ANOVA in R example

Correlation applet
another one

wheezing.txt
cheese.txt
twins.txt
fev.dat.txt

Interesting Links:

BBC Radio:  More or Less
R Statistical Software