The cost effectiveness of using certified mail as a device to raise response rates in mail surveys is investigated. Of the three optional services offered by the Post Office when sending certified mail, the base certified rate was recommended, since the Post Office often failed to perform the more expensive services. A sample of 120 students from 30 schools was used, with two males and two females participating from each school. Continuing for 19 days after the mail-out date, the experiment randomly assigned people from the same sex and school to either a certified mail group or a first-class mail group. The overall difference in response rates appeared to be appreciable. The cost effectiveness of using certified mail for a 10 item and a 150-item questionnaire was then formulated. The results indicated that the most cost effective plan was the use of certified mail with the long questionnaire and first-class mail with the short one. It was recommended that the cost effectiveness of mail survey plans be calculated prior to their implementation.