Premium
The Use of Covariance in Observational Studies
Author(s) -
Cochran W. G.
Publication year - 1969
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.2307/2346587
Subject(s) - covariance , observational study , statistics , econometrics , mathematics
Summary When two groups of subjects are being compared, one group exposed to some presumed causal force and the other not, adjustment of the difference ( ӯ 1 – ӯ 2 ) in mean responses by means of a regression on one or more x ‐variables is sometimes employed in order to remove possible biases that arise when subjects are not assigned to groups at random. In such applications, Belson (1956) has suggested that the adjustments be made by means of the regression for the unexposed group only, whereas the routine user of the analysis of covariance employs the pooled regression coefficients from both groups. This note tries to clarify the situations in which Belson's proposal is preferable.