Workers at a Canadian industrial site read a vignette asking them to send a message to a co-worker and then rated their preferences for available message channels. We explored the respondents' preferences for either a word-processed or a handwritten message. The results indicate that (a) main effects and interactions involving hierarchical level, message length, message complexity, anticipated reaction, communication task, need for documentation, and communication across work shifts affect preferences for wordprocessed versus handwritten messages; (b) the cost control perspective can explain preferences for word-processed versus handwritten messages; and (c) scholars should distinguish between various types of written messages rather than grouping all written messages together in a single category.