Randomized clinical trials adhere more closely to pre-agreed-on protoc
ols than almost any other type of experiment; yet we can tighten up th
eir analysis if we desire. If we convert the analysis into a randomiza
tion analysis-where the one set of data is analyzed many times-once as
though each acceptable assignment has been employed, we can eliminate
any dependence of the analysis on statistical or probabilistic assump
tions. To do this effectively when many assignments could be acceptabl
e, we can go to double randomization, in which a subset, usefully kept
balanced, of acceptable assignments is selected (perhaps randomly) be
fore data acquisition. If we have one covariate, adjustment for which
answers a question that is at least as appropriate, we can easily buil
d on this. Imperfect covariance adjustments can help almost as much as
perfect ones. If it is appropriate to work with many covariate(s), it
is often desirable to first construct a (few) compound covariate(s) a
nd then work with it (them). Often we can base the coefficients in our
compound covariate on the univariate regressions of response on singl
e covariates. Doing this within each arm of the trial and pooling keep
s the fitting of the final adjustment unbiased. Since we can prespecif
y how the compounds are to be calculated and fitted, we can do all thi
s while retaining rigid prespecification. Prespecification, randomizat
ion, and intelligent use of covariates combined to make the resulting
significance analysis of platinum standard quality. (If we want confid
ence statements, as we ordinarily should, it may make sense, for techn
ical reasons, to plan for somewhat less than platinum standard quality
.)