| Defines | Factors, units, and comparisons |
|---|---|
| Chosen according to | Objective and constraints |
| Helps limit | Confounding between explanations |
Definition and context
Experimental design determines which factors to vary, which to control, which units to observe, and how to organize comparisons. It must address a defined objective. The choice of analysis needs to be connected to that organization, not added as decoration at the end.
In a fictional trial of two surfaces, testing each under different conditions can confound the effect of the material with the effect of the conditions. Repeating that comparison many times does not, by itself, eliminate the problem. A good design tries to separate alternative explanations and makes its limitations visible. It also distinguishes independent units from repeated measurements on the same unit. The result can be informative even if it answers only a narrowly defined question.
What we know and its limits
Introductory reference sheet. Properties and applications depend on measurement conditions; consult the sources before making comparisons.
Follow the evidence
The original sources for exploring this explanation and its limits.
- Introductory Statistics 2e · Data and Sampling
OpenStax · Rice University
Samples, bias, and sampling variability. - Engineering Statistics Handbook · Selecting an Experimental Design
NIST / SEMATECH
Objectives and decisions in experimental design.
Image credits
Photo by Jahra Tasfia Reza on Pexels · Original source · Pexels License