This study was designed to gather data on the meaning of imprecise terms from items written by physicians for their students and by test committees for national licensure and certification examinations. A total of 32 members of test committees who write examination items for various medical specialty examinations participated in the study. Each participant was provided with a list of phrases used in multiple-choice items to express some concept related to frequency of occurrence. Participants were asked to indicate what percentage of time was reflected by each phrase and to express this percentage both as a single number (i.e., 75% of the time) and as a band (i.e., 60 to 80% of the time). The list included 15 terms. Responses of two participants were deleted due to excessive aberrance. Results indicate that the phrases used by item writers to express frequency do not have an operational definition that is commonly shared. Both the single values and phrases varied considerably across individuals. Problems created by vague terms are much more severe for true/false type items than for one-best answer items. These findings support the contention that general guidelines on item writing would be beneficial to the medical community. One table and five figures are provided. (TJH)