How constrained a response: A comparison of binary, ordinal and metric answer formats

A question is the main measurement instrument in the social sciences. Yet no conclusive results exist with respect to the suitability of alternative answer formats for typical constructs studied in this field. Furthermore, no prior studies have used real answers from respondents to investigate differences in responses as a consequence of answer formats, typically assuming the way in which individuals translate their responses from one answer format to another. In this study we make a first step towards filling the above two gaps. We investigate answer format effects for two different constructs (attitudes, intentions) and three formats (binary, ordinal, metric) using a repeat measurement design. Results indicate that formats lead to the same managerial interpretations with the same reliability; differ in speed and perceived speed while being perceived as equally simple, pleasant, and useful to express feelings. Binary and metric answer formats are found to represent interesting alternatives to the predominantly used ordinal format, especially when speed of completion or the availability of metric data for analyses is essential. (C) 2006 Elsevier Ltd. All rights reserved.