This paper explores the comparability of item calibrations for three types of items: (1) text only; (2) text with photographs; and (3) text plus graphics when items are presented on written tests and computerized adaptive tests. Data are from five different medical technology certification examinations administered nationwide in 1993. The Rasch model was used to calibrate items for the two test formats. Item calibrations obtained from each administrative mode were then compared. No significant differences were found between text only item calibrations obtained from the written tests and the computerized adaptive test. While some items with photographic or figure accompaniment showed slightly different item calibrations between the administrative modes, nonstatistical explanations explain most of the minor differences discovered. The results of this investigation confirm that Rasch item calibrations from written tests are appropriate for use on computerized adaptive tests. Included are four tables and six figures. (Contains 7 references.) (Author/SLD)