This article discusses how aggregate exchange-level demographic data can be used to adjust for non-resolvability in telephone surveys. Traditionally, surveys have been challenged by non-working or disconnected phone numbers, but this method allows researchers to adjust for these issues. The article presents a method using telephone exchange data to impute values for non-working numbers based on characteristics of the local telephone exchange area. This helps reduce non-response bias without requiring costly efforts to re-contact numbers or obtain additional individual-level data.