ABSTRACT
Dunnett (1955) pioneered the concept that, when a control is present, the comparisons of primary interest may be the comparison of each new treatment with the control. For example, the control may be a placebo, or it may be a 'standard' treatment, or any other 'specified' treatment (such as a new drug). We call such comparisons multiple comparisons with a control (MCC). Suppose Hi,..., IJL^-I are the means of the new treatments and /^ is the mean of the control; then in MCC the parameters of primary interest are fa — ^ for i = 1,. . . , k — 1, the difference between each new treatment mean fa and the control mean ^ .
