This paper deals with test fairness regarding a test consisting of two parts: (1) a "common" section, taken by all students; and (2) a "variable" section, in which some students may answer a different set of questions from other students. For example, a test taken by several thousand students each year contains a common multiple-choice portion and a common essay portion but also a variable essay portion, in which the test-taker may choose to answer any one of five questions. On this test the questions that the test-taker may choose from are intended to be of equal difficulty. When the scoring has been completed and the results tabulated, the data occasionally suggest that two or more essay questions may not have been of equal difficulty. If there had been no reason to believe, a priori, that the questions on the variable portion were of equal difficulty, the scores would need to be adjusted in such a situation. Problems arise with the two adjustments: option A leads farthest away from the assumption of equal difficulty when the evidence against it is weakest; and option B is equivalent to assuming that the questions in the variable portion are, in fact, equally difficult. A compromise is proposed that is closer to option A when the common portion predicts the variable portion accurately and closer to B when it does not. (PN)