This paper presents a new technique for improving coal mine productivity in which a production-function approach was used to model daily coal production at the most fundamental producing unit-- the mine crew. The study shows that mine managers already have easy access to information that can help them explain, control, and predict production at the crew level. Computer-derived models were used to assess the relative effects of labor, technology, and environmental factors on the daily reported coal production of 81 mining crews at 7 underground coal mines in eastern coalfields. Almost all of the data came from the daily production foreman reports that are routinely gathered by most mining companies. Production models were derived from the data using a linear regression technique. The models explained a significant proportion of the day-to-day variation in coal production. Also, the input variables were found to significantly affect production in the expected directions. This technique promises to become an inexpensive and useful management tool for detecting and diagnosing production problems, assessing the effectiveness of a change both before and after implementation, and isolating factors that lead to changes in production. Mine managers can readily implement this technique by using their daily crew reports and simple linear modeling software running on any available computer.