The purpose of this study was to investigate the reliability, subscale intercorrelations, and validity of the Administrator Role Perception Inventory (ARPI), an instrument developed by the investigator and a co-author in the year preceding the study. The ARPI is a 50 item, Likert-type scale which reflects variables thought to be associated with burn-out: expectation, motivation, accomplishment, psycho-physical state, and relationships. Another variable, previously unresearched, was also included; it is the variable of "time" and reflects a negative time orientation, a longing for the "good old days." The inventory was mailed to all of the 2,113 active members of the Confederation of Oregon School Administrators during the middle of August, 1982. The return rate for the ARPI was 62 percent. Analysis of the data showed that the ARPI has an internal consistency of .91. The coefficients for the subscales fell between .70 and .85. As expected, there were substantial intercorrelations among the subscales, ranging from a low of .31 to a high of .63. Total ARPI scores and the subscale scores were correlated with measures of self-perceived burnout, desire for early retirement, and felt job stress. All of these correlations were in the expected direction, were significant at the .001 level, and ranged from -.41 to -.59. It was concluded that the Administrator Role Perception Inventory is comparable to measurement instruments similar in purpose and format.