Measurement Error When Surveying Democratic Support: A MultiTrait MultiError Approach

Recent studies have suggested that support for democracy may be in decline even in some established democracies. To better understand these processes, it is essential to measure such complex concepts as support for democracy correctly, since attitudinal survey questions are affected by measurement error. This study investigates the effect of different types of measurement error on survey items measuring democratic support. The items were specifically designed to achieve high construct validity. The study is based on two waves of the Finnish Citizens’ Opinion panel using a MultiTrait MultiError (MTME) approach. We estimate concurrently correlated errors (social desirability bias, acquiescence bias, method effects) and random error. The design manipulates statement framing, response scale direction, and scale polarity across two measurement points while employing latent variable modeling to estimate the errors. We find that correlated errors make up a relatively small proportion of the error, suggesting less pronounced measurement validity issues. We find a much higher proportion of random error, suggesting higher levels of attitude instability. Still, there is significant variance both between survey items and different forms. Moreover, differences in random error structure suggest that attitude instability is greatly reduced when respondents can anchor democratic norms to recent ‘real-world’ events.