The structure and dynamics of the education gap in immigration preferences are not well understood. Does the gap increase when the economy contracts? To what extent does the gap reflect labour market conflicts versus value polarization? Does the structure of the gap change with labour market and refugee shocks? I use European Social Survey data to decompose the gap into parts reflecting labour market position, social background, and value orientation, and explore how their importance in accounting for the gap change over time. I find no uniform trends in preferences or in the size of the gap, but the gap varies with the unemployment rate and the strength of trade unions. Moreover, positions in the labour market are more important for the gap in times of high unemployment, at the expense of the importance of value orientations. The results show the enduring importance of labour market conflicts for the gap.