IPAK-EDU Analytics 200

Applied Biostatistics

Instructors: Matthew Buns, PhD and James Lyons-Weiler, PhD

In this course, you will be introduced to the principles and practice of inferential statistics in action, including key principles of distributional assumptions, random samples, randomization, parametric and non-parametric statistical hypothesis testing, statistical power, measuring association & correlation, comparing means between two and among two or more populations, and introductory regression theory and practice.

All lectures will have in-class/take-home exercise conducted in spreadsheets.Upon completion of this course will be able to know the difference between descriptive statistical analysis perform statistical hypothesis testing in a variety of settings and be able to critique the use of specific approaches to hypothesis testing in published studies.

Recommended book(s): There are many books that make statistical analysis accessible.A few I recommend, in increasing sophistication

-Statistics For Dummies

-Research Methods and Statistics in Psychology (SAGE Foundations of Psychology series)

-A Concise Course in Statistical Inference

Also – How to Lie with Statistics

No book is need for this course; for each lecture, I’ll also send a list of links for readings online that will be useful.

Topic/Concepts

1. Measures of Central Tendency

distributions

mean, median, mode

central limit theorem, i.i.d.

Chebychev's theorem

2. Sampling a Population

Why random sampling is important

Measures of skew

Measures of normality

Random error and bias

Homework 1: Learn how to Randomly select 100 values from 1,000 numbers in Excel

3. Hypothesis testing 1- Means.

Randomization/Permutation testing - difference of means

Homework

3a: Calculate the means of two samples

3b: Randomize the data between groups and re-calculate the mean 10,000 times

concept: alpha (e.g., 0.05)

Example 1: Means ARE different @ a = 0.05

Example 2: Means ARE NOT different @ a = 0.05

4. Parametric hypothesis testing - t-test

Assumptions and testing assumptions

Concepts: p-values, relation to alpha, independent and dependent variables

Homework 4

Testing the hypothesis of different means w/the t-test

Example 1: Means ARE different @ a = 0.05

Example 2: Means ARE NOT different @ a = 0.05

5. Hypothesis testing 2 - Categorical data (counts)

Chi-square

Fisher's exact test

Homework 5

Categorical data example 1

6. Statistical power

Concepts: Effect size, sample size, alpha, intrinsic power of a test

Homework 6: calculate statistical power for the t-test, two populations

7. Measuring correlations 1

Linear models

Homework 7: Calculation the correlation coefficient

8. Non-parametric tests

Spearman rank correlation

Homework 8: Perform a spearman rank correlation calculation

9. Application - Comparison of means

Student challenge - Choose 2 of 5 example data sets and choose a test

Real-time, in class experience.

10. ANOVA

Analysis of Variance – measuring variation among means in multiple populations

Assumptions

Example (what to expect)

No homework

11. Application - Measuring association

Chi-Square test,

Relative Risk

Odds Ratio

Homework

Calculate Chi-Square, RR and OR given some real data

12. Application - Power curves

Empirical demonstration of robust rejection of Ho

Using a given effect size, calculate power curves over sample size range low to high

13. Application - Correlation

Find real-world examples of calculation of correlation w/available data

re-calculate correlation coefficient

14. Application - Regression models

Lecture on Regression

Independent and dependent variables again

Parameter interpretation - slope, intercept, R^2, error

Why curve-fitting is so limited in utility.

15. Summing it up - STUDENTS EVALUATE AND INTERPRET THE ANALYSIS OF A STUDY.

Read and interpret one of a list of studies

Focusing on the data analysis, interpret the results.

Did the study meet the assumptions of the analysis?

Was the study sufficiently powered?

Critique the study using what you’ve learned in this class and in past courses.