Direct Inference and Randomization

Abstract

R.A. Fisher began defending the use of randomized assignments of treatments to plots in agricultural field experiments in 1926 (Fisher 1926). His claims on behalf of randomization were challenged by many other statisticians including his friend and colleague W.S. Gosset, the celebrated “Student” (Gosset 1937). After Gosset's death, Fisher's insistence upon the critical importance of randomization to good experimental design became a cornerstone of methodological orthodoxy, a status it has continued to enjoy until recent years. Even so, an undercurrent of resistance to some of the more excessive claims on behalf of randomization has continued to influence thoughtful statisticians and methodologists. Fisher's rationale for. randomization was grounded on the need to control or eliminate systematic error or bias. For example, the average yield of wheat on plots treated with one kind of fertilizer may be compared with the average yield of wheat of the same variety on plots treated with another kind of fertilizer.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.