Presented by:
John C. Pezzullo, PhD
Associate Professor, Pharmacology and Biostatistics
Georgetown University
This 14-session course will give you a practical working knowledge of statistics, based on an understanding of the underlying concepts. It is oriented toward the needs of the practicing researcher in the life sciences, with a special emphasis on the analysis of laboratory and clinical data.
These sessions are actually the first part of a two-part course, and are meant to give you the background you need to participate in the second part, which will begin in January, 2002. The second part will consist of interactive sessions where each week you and members of the Pharmacology faculty will examine in depth a specific topic or issue that arises in the design of scientific experiments.
Pre-requisites: You don't need any previous statistical knowledge or training. You should be familiar with simple algebra (know how to read formulas) and you should know how to do basic numerical calculations, either by hand or with a calculator. You should be comfortable using a personal computer to access the the Web, and should know how to use Excel to do simple calculations.
Textbooks and other resources: You will not have to purchase any textbooks or statistical software for this course; everything you will need is available on the Internet. You should have access to a computer capable of connecting to the Internet, either through a network connection at work or through a dial-up (modem) connection at home. If you do not have an Internet account (such as Erols, ATT WorldNet, America Online, or Georgetown's dial-up service), free Internet access software will be provided. You will also be given a CD-ROM with a large selection of statistical analysis software for use during this course and afterwards.
Structure of the classes: Most classes will be divided into several parts:
Lectures will be informal, and you should feel free to speak up with any questions or comments you might have.
Examples: This course will use examples from the medical and scientific literature. In addition, your are encouraged to bring your own data to class, and we'll analyze it during class.
Schedule of Classes: Classes will be held once a week, on Wednesdays, from 3:00PM to 5:00PM, in the Pharmacology Library, located on the fourth floor of the Med-Dent building, in room NE-401, unless otherwise specified in the schedule below..
The following is a tentative outline of the material to be presented in each session. This syllabus is subject to change, based on the background of the students, and the progress of the classes.
| Session 1: The Basics (click for material) |
Wed, Sep 4 |
In this session, we will review some basic math techniques - elementary arithmetic calculations and reading mathematical formulas. We'll go over some fundamental statistical concepts - random variability, probability, sampling, and frequency distributions. You'll learn how to construct frequency distribution diagrams (histograms). We'll generate random values by several processes, and observe their behavior. You'll also learn how to locate and use statistical software, textbooks, and other resources on the Internet. |
| Session 2: Describing Data (click for material) |
Wed, Sep 11 |
This session will cover "descriptive statistics" - the craft of summarizing and presenting data. You'll learn how to reduce large collections of numbers down into a few numbers that concisely represent the collection - measures of central tendency, dispersion, skewness, and other characteristics. |
| Session 3: Statistical Inference (click for material) |
Wed, Sep 18 |
This session will cover the fundamental concepts and terminology underlying almost all the statistical tests and analyses you're likely to encounter -- populations, samples, hypotheses, types of inference errors (Type-I, Type-II), significance, power, precision, and confidence levels. |
| Session 4: The Statistical Probability Distribution Functions (click for material) |
Wed, Sep 25 |
This session will give you an intuitive understanding of six of the most useful probability distribution functions: the Binomial, Poisson, Normal, Student t, Chi Square, and Fisher F distributions. You will learn what they are, how they arise, how they behave, and why they are useful. At the end of this class you will understand what the the statistical probability tables for these distributions really mean, and how they are generated. This will set the stage for the classes that follow, which apply this knowledge to the real-world statistical analysis of data. |
| Session 5: Analyzing Categorical Data (click for material) |
Wed, Oct 2 |
In this session you'll learn how to construct contingency tables (cross-tabs) to summarize the relationships between categorical variables, and you'll apply your understanding of the Chi Square distribution to analyze these tables.. |
| Session 6: Comparing Groups of Numbers (click for material) |
Wed, Oct 9 3:00-5:00PM |
In this session you'll apply your understanding of the Student and Fisher distributions to analyze two or more groups of numbers to see if they signify real differences, or merely represent random fluctuations. This class will cover the Student t and other related tests, and the simplest of the Analysis of Variance (ANOVA) tests, with "post-hoc" tests to determine where the "real" effects are. |
| Session 7: More Analysis of Variance |
Wed, Oct 16 3:00-5:00PM |
This session continues the discussion of the ANOVA, as extended to more complex designs that involve multiple factors, repeated measures, and interactions. |
| Session 8: Measurement Errors (click for material) |
Wed, Oct 23 3:00-5:00PM |
This class examines the avoidable and unavoidable errors that affect every measurement you make. You'll learn about the different types of measurement errors, how they arise, how they can be minimized. You'll learn how to estimate the magnitude of these errors, and how to construct confidence intervals around your measured values. You'll also learn how measurement errors propagate through subsequent calculations you perform on your data. |
| Session 9: Simple Regression, Correlation and Curve Fitting (click for material) |
Wed, Oct 30 |
This class will provide an introduction to the concepts, terminology, and techniques of regression and correlation analysis. You'll learn how to prepare informative graphs, how to test for significant relationships between continuous variables, and how to determine the functional form (the equation) that describes that relationship. |
| Session 10: More Regression (click for material) |
Wed, Nov 6 3:00-5:00PM |
This class will concentrate on the special considerations that come into play when you have more than one independent variable (predictor) -- multivariate regression, model-building, stepwise regression, collinearity, confounders, and interactions. |
| Session 11: Still More Regression |
Wed, Nov 13 3:00-5:00PM |
Continuing and extending the material from the previous classes, this session covers non-linear curve-fitting. You'll learn how to interpret the results of these analyses, and how to construct confidence intervals around your estimated parameters and confidence bands around your fitted curves. |
| Session 12: Survival Analysis (click for material) |
Wed, Nov 20 3:00-5:00PM |
In this class you'll learn how to analyze survival data using the life-table and Kaplan Meier methods. You'll learn how to prepare survival curves, and to compare two curves two determine whether two or more groups of subjects have significantly different survival characteristics. You'll also learn how to use Cox Proportional Hazards regression to examine the effects of one or more variables on survival. |
| No Session | Wed, Nov 27 | This is the day before Thanksgiving, and many of us will be traveling, so class will not be held today. |
| Session 13: The Design of Experiments and Studies -- Power and Sample Size (click for material) |
Wed, Dec 4 3:00-5:00PM Pharm. Library (NE-401) |
This class will provide an introduction to the concepts and techniques involved in planning lab experiments and clinical studies. You'll learn the different types of study designs, how to minimize various sources of bias in sampling and measuring, and how to ensure that your sample size is large enough to detect the effects you're looking for, or to attain the precision you expect. |
| Session 14: Review and Miscellaneous Topics (click for material) |
Wed, Dec 11 3:00-5:00PM Pharm. Library (NE-401) |
In this final session we'll review everything we've presented in the course, and cover other topics that might be of interest to the class. |
| Session 15: Final Examination |
Wed, Dec 18 3:00-5:00PM Pharm. Library (NE-401) |
The final exam for the course will be administered during this session. |
StatPages.net -- Interactive Statistical Calculating Pages
(to be provided)
(to be provided)