Exploring Mode Effects in a Panel Survey of New Businesses

Establishment surveys among the business population are often hampered by low response rates and high costs. In an effort to improve response rates and reduce costs, survey researchers have employed multi-mode data collection methods, often mixing a self-administered visual mode (such as a mail or web survey) with an aural mode such as telephone. This method provides both the benefits of the self-administered questionnaire (cost and convenience) and those of the more persuasive interviewer-administered mode for nonresponse follow-up. A better understanding of the potential effects of conducting multi-mode establishment surveys can provide information about the characteristics of businesses and respondents who choose the different modes and also guide future data collection efforts among this population. We used data from the baseline and three follow-up waves of the Kauffman Firm Survey (KFS), a longitudinal survey of new businesses offering a web survey to the entire sample followed by an opportunity for web nonresponders to complete their interviews by telephone, to shed light on the differences between businesses that complete the survey using the different modes and determine whether any mode effects can be detected. Specifically, we present a longitudinal analysis of three groups of KFS businesses: those that opted to complete all four rounds of the survey on the web; those that opted to complete all four rounds by telephone; and a group of “switchers” that completed two rounds by web and two rounds by telephone. For this descriptive analysis, we examine differences in these three groups across industry, size, and the complexity of the business, as well as differences in respondent demographics. We also analyze indicators of business growth, such as employment, assets, revenue, and profit or loss