The consequences of incorrect analysis of blocked experiments

[摘要] Errors are always present in experimental measurements so, it is important to identify them and understand how they affect the results of experiments. Statistics suggest that the execution of experiments should follow random order, but unfortunately the complete randomization of experiments is not always viable for practical reasons. One possible simplification is blocked experiments within which the levels of certain factors are maintained fixed while the levels of others are randomized. However this has a cost. Although the experimental part is simplified, the statistical analysis becomes more complex.

[学科分类] 化学（综合）

[关键词] optimization;split-plot design;factorial design.