| Attitude questions differ in format, e.g. differences in numbering and labelling of response categories. It has been argued that the validity and reliability of attitudes is affected by the choice of question format. At the same time, it is acknowledged that response style behaviour can bias the measurement of attitudes as well as bias the estimates of the effect of covariates. This research project links these two issues by focusing on the impact of question format on the likelihood of response bias, i.e. acquiescence and extreme response style, in attitude research. |
