Despite the political prominence of intergenerational social mobility over the past twenty ears, there is much that we do not know about its patterns and trends. Problems of data availability and measurement quality abound, with different measures and datasets producing conflicting results. In this talk I will present results from a set of papers that use data from the Office for National Statistics Longitudinal Study (LS) which is a linked 1% sample of census records running from 1971 to 2021. Compared to other data sets, the LS offers a number of advantages, notably its very large sample size and high rates of response/record linkage. I will present descriptive evidence on trends over time for cohorts born in the late 1950s to the early 1980s, separately for men and women and for different regions in England and Wales. I will also show how temporal and geographic variation in social mobility is related the local level of selective schooling and (time permitting) voting in the 2016 EU Referendum.