In this study the influence of topic involvement on mail-survey response rate and speed was experimentally investigated. The results show that response rates for topics that are generally considered as high involvement were higher than for topics that are generally considered as low involvement. However, the speed of response did not differ between the two levels of topic involvement. Moreover, in terms of response rates an interesting interaction effect was observed: For the high-involvement topic, the response rate showed a further positive and significant correlation with the within-topic level of involvement, but this was not the case for the low-involvement topic. The implications of these findings to nonresponse bias in mail survey research are discussed. © 2002 Wiley Periodicals, Inc.