| 释义 |
experimental design (in statistics) Observational studies can suggest things for which the explanation lies with some hidden variable. For example, the performance of pupils taught in large classes is better than that of pupils in small classes, not because the large class is a more effective learning environment, but because schools tend to operate larger classes for the more successful. Experimental design methods seek to control the conditions under which observations are made so that any differences in outcome are genuinely attributable to the experimental conditions and not to other confounding variables. Some of the most common methods are the use of paired or matched samples, randomization, and blind trials.
|