The purpose of this study was to develop a stressor inventory for the Taiwanese elderly that allows the respondent to rate the degree of stressfulness of each stressor experienced. The research was done in three phases: identifying stressors, generating instrument items, and establishing psychometric properties. In phase one, 33 subjects were interviewed to determine stressors they experienced since they turned 65. The interviews were tape recorded and transcribed. Content analysis was done to analyze the interviews and determine the stressors. Interviews were done until no new stressors were described by subjects. Seventy-three stressors were identified and listed in the instrument. In phase two, seven experts were asked to judge if they believed these items to be of relevance as a stressor. The index of content validity was used for quantification of content validity. In phase three, 351 subjects were used in determining the psychometric properties of the instrument. Generalizability theory was used to test stability and internal consistency of the TESI. State-anxiety and the Positive and Negative Affective Schedule were used to determine the construct validity of the TESI. The content validity for three experts were.97 and for two experts was.90. In determining the stability, the generalizability coefficients for the relative decisions was.899 and generalizability coefficients for the absolute decisions was.897. As to stability of stressor frequency, the generalizability coefficients for the relative decisions was.719, whereas it was.717 for the absolute decisions. For the internal consistency for stressfulness, the generalizability coefficients was.523. As to construct validity of the TESI, a correlation of.586 with state-anxiety, of.558 with negative affect, and $-$.283 with positive affect were obtained. Continuing work to revise the instrument is to remove six months time limitation and combine similar items.