ABSTRACT
The inferences that can be drawn from confidence intervals and statistical tests depend on the researcher’s definition of probability and on the researcher’s prior knowledge of the parameter that is being estimated. For two definitions of probability and three types of prior knowledge, the relative advantages and disadvantages of confidence intervals and statistical tests are compared. Under conditions equally favorable to both statistical tests and confidence intervals, statistical tests are shown generally to be more informative than confidence intervals when assessing the probability that a parameter (a) equals a prespecified value, (b) falls above or below a prespecified value, or (c) falls inside or outside a prespecified range of values. In contrast, confidence intervals are shown generally to be more informative than statistical tests when assessing the size of a parameter (a) without reference to a prespecified value or range of values or (b) with reference to many prespecified values or ranges of values. In addition to discussing differences in informativeness, we also compare confidence intervals and significance tests in terms of their cognitive demands and their effects on bias in publication decisions. Finally, we discuss when confidence intervals should be used, describe obstacles to their use, and suggest how these obstacles can be overcome.
