Introducing Web in a Telephone Employee Survey: Effects on Nonresponse and Costs.

This article examines the effects of introducing a web starting mode in a national employee survey in Germany, comparing a traditional single-mode telephone design with a sequential mixed-mode design combining web and telephone. The study found that adding the web mode significantly increased response rates and reduced average survey costs per respondent by about 50 percent. However, despite higher participation, the mixed-mode design exhibited greater aggregate nonresponse bias than the single-mode telephone design, though telephone follow-up of web nonrespondents helped reduce this bias. Participation likelihood varied by employee subgroup, with full-time workers and those in business or administrative occupations more likely to respond via web, suggesting mode choice impacts sample composition. These findings highlight trade-offs between cost, response rates, and data quality in employee surveys when incorporating web modes.