|
The GraphPad Guide to
Comparing Dose-response or Kinetic Curves
By Dr. Harvey Motulsky
President, GraphPad Software
hmotulsky@graphpad.com
All contents are copyright © 1995-2002 by GraphPad Software, Inc. All rights reserved.
Statistical tests are often used to test whether the effect of an experimental intervention is statistically significant. Different tests are used for different kinds of data – t tests to compare measurements, chi-square test to compare proportions, the logrank test to compare survival curves. But with many kinds of experiments, each result is a curve, often a dose-response curve or a kinetic curve (time course). This article explains how to compare two curves to determine if an experimental intervention altered the curve significantly.
Download the .pdf (printer-ready file) file now (397kB).
Approach 1. Pool several experiments using a best-fit parameter from nonlinear regression
Approach 2. Pool several experiments without nonlinear regression
Approach 3. Analyze one experiment with nonlinear regression. Compare best-fit values of one variable.
Approach 4. Analyze one experiment with nonlinear regression. Compare entire curves.
Approach 5. Compare linear regression lines
Approach 6. Comparing curves with ANOVA
Comparing more than two curves
Summary
References
More information on nonlinear regression:
GraphPad Guide to Analyzing Radioligand Binding Data (40 pages).
Our web site, curvefit.com , which has far more detailed information that this Guide. It includes general information on nonlinear regression, as well as its application to analyzing radioligand binding, enzyme kinetic, and dose-response data.
Learn more about our program, GraphPad Prism; the best general-purpose nonlinear regression program for Windows and Mac.