The Impact of Survey Methodology and Context on Central Tendency, Nonresponse and Associations of Subjective Indicators of Government Performance

摘要

A comparison is given of (ordinal) measures of trust in government institutions that were collected with the aim of facilitating comparisons between different modes and contexts of data collection. Statistical analyses reveal that the mode effect could amount upto a 10% change of item nonresponse and a 9% change of the level of trust, whereas the context effect could lead upto a 19% change of the bivariate associations of trust. The observed changes of both measures of central tendency and of associations affect conclusions of social scientific research in isolated studies, which implies that comparative studies are seriously hampered by bias related to mode and context, two sources of obscurity adding to the two other threats which are differences of sample source and differences of question phrasing.