Civic Non-Participation in Scotland: A Missing Data Perspective

Democratic government relies on a great deal of voluntary participation from the population. This participation is not guaranteed, and when citizens are absent from the data collecting processes of the state, a missing data problem occurs. The co-occurrence of the Scottish local elections and the Scottish Census in 2022 provides an unusual opportunity to test hypotheses about common and context-specific drivers of civic non-participation. Using a geospatial approach, we show that ward-level census nonresponse is highly correlated with electoral abstention despite abstention being about five times more common than initial census nonresponse. Statistical modelling demonstrates strong common roles for deprivation and English-language speaking in driving the participation rates of a ward. Meanwhile relatively specific effects are shown for cohabitation in driving census response, and the proportion of young people in driving voting behaviour. We conclude that the similarity between these nonresponse processes represents an opportunity for improving practice in increasing response rates. For example, census “hard to count” indices can incorporate data from elections as “interim” predictions for low-response areas, and framing voting as a household duty, rather than an individual one, may improve turnout.