It is a well-documented fact that social trust, i.e., the extent to which people trust others, and political trust, i.e., trust in political institutions, are key factors in social capital theory. However, to compare these concepts in cross-national or longitudinal frameworks, it is important to first establish whether the measurements of these concepts are compatible across countries or over time. This paper tests the measurement and cross-national and longitudinal invariance properties of social and political trust. We use multiple-group confirmatory factor analyses (MGCFA) to evaluate the different levels of invariance (configural, metric and scalar) using data from the European Social Survey (ESS) measured at four different time points (2002, 2004, 2006 and 2008) in seven Western European countries. In a second step, the country mean rankings of social and political trust are computed based on the latent scores and compared with those based on traditional sum score measurements. This comparison illustrates the potential inaccuracy of sum scores for country mean comparisons when measurement invariance is not supported by the data.