Answer format effects revisited

The effect of answer formats presented to respondents in written surveys are investigated for two constructs (attitudes and behavioral intentions) and three response scales (binary, ordinal and metric). Results indicate that (1) formats differ in their susceptibility to response styles but lead to the same results with respect to average values and underlying dimensions; (2) binary format is quicker to complete and perceived as quicker while all formats are perceived as equally simple, pleasant, and useful to express feelings; (3) an interaction between the construct measured and the answer format clearly exists which should be investigated more systematically in future research.