The study of the factors that affect response rates in survey research in the organizational sciences has been fairly extensive (Cook et al., 2000). The main concern associated with mailed surveys driving this research is the high risk of nonresponse error in these studies (Mangione, 1998). Indeed, these concerns are warranted, as nonresponse error can lead to major trepidation about the external validity of research (Roth & BeVier, 1998). Furthermore, most scholars appear to agree with Mangione (1998) that a solution to nonresponse errors is to “do everything in your power to conduct a study that has a very high response rate” (p. 404). Given the increase in the number of studies utilizing coaches in sport management literature (see Gilbert, 2002), inquiry assessing techniques to improve response rates in coaching samples seems warranted. The purpose of this study was to assess the effectiveness of two specific factors cited in the broader literature as useful toward increasing responses: personalization of the correspondence and questionnaire color (Clark & Kaminski, 1990; Greer & Lohtia, 1994). Specifically, we assessed whether response rates differed based on the cover letter being signed or unsigned in ink, or if the questionnaire was printed on white versus yellow paper. As part of a larger study, a total of 512 questionnaires were sent to NCAA Division I (n = 252) and Division III (n=260) assistant softball coaches. The questionnaire was long (126 items) and consisted of a series of Likert type questions. The sample was randomly assigned to four different treatment groups (128 white, unsigned; 128 white, signed; 128 yellow, unsigned; 128 yellow, signed). Responses for one round of data collection achieved an overall response rate of 25% (n= 128). The signed questionnaires achieved and 28.3% response rate (compared to 21.1%) and the yellow versions received a 29.3% response (compared to 20.7%). Binomial logistic regression was used to analyze the data. Results indicated significant effects on the response rate for both color (Wald = 5.04, p =.02, Exp (B) = .63) and personalization (Wald = 4.18, p = .04, Exp (B) = .65). The interaction term for the two variables was not significant (Wald = .21, p = .65, Exp (B) = 1.21), indicating that the combination of the two design characteristics did not impact responses. Future scholars utilizing coaching samples should be cognizant of these findings and attempt to personalize and include the use of colored paper when conducting their research. Keyword(s): coaching, research, sport management