Effect of Questionnaire, Cover Letter, and Follow-Up on Non-Response Rate in an E-Mail Assisted Internet Survey in Pakistan

Survey non response has always been a teasing research concern due to its implications for reliability, validity, generalizability, and replicability of the findings and Internet surveys are no exceptions (see e.g. Best et al. 2000; Braithwaite et al. 2003; Cook and Thompson 2003; Kypri and Gallagher 2003; Schaefer and Dillman 1998). Internet surveys are fairly popular in developed countries for a verity of research purposes, and in variety of fields, ranging from healthcare, to social behaviour to tourism (Bayliss et al. 2003; Best et al. 2001; Hwang and Fesenmaier 2004; Im and Chee 2005; Schleyer and Forrest 2000). This is particularly so if the survey universe has access to e-mail (Berrens et al. 2003). Even when the entire universe does not have access to e-mail, multimode approach is fairly successful in gaining high response rates (Schaefer and Dillman 1998). With increasing access to Internet and e-mail by various groups in Pakistan, research questions concerning Internet-based surveys are of paramount relevance. In Pakistan, the research culture, however, is close to being absent in social science disciplines (Shah, Qureshi, and Abdul-Ghaffar 2005). Consequently, common people do not understand the importance of research. Besides, response rate in Internet surveys tend to be generally lower than equivalent mail surveys, unless rigorous follow ups are conducted (Cooper, Blair and Triplett 1999; Sheehan 2001). This may be due to lack of understanding in research community on how to achieve high response rates using the Internet surveys. Given this scenario, understanding factors associated with non response in e-mail assisted Internet surveys of Pakistan is important. Our study investigates the effect of nature of the subject matter being studied, font color of questionnaire, follow up letter, and presence of cover letter on the non-response rate in e-mail assisted Internet based survey of a closed sample of respondents.  Although non-response rate in internet surveys has been a subject of investigation elsewhere (Trouteaud 2004 ), to our knowledge no published study addressing Internet survey non-response rate variation or its correlates exists in Pakistan. This signifies a strong justification for and strength of our study.