Erola and Moisio (2007) argue that in Finland the class positions of grandparents and grandchildren are almost independent of each other, once parents ’ social class has been taken into account. We show that this conclusion of ‘almost conditional independence ’ is actually not supported by the results reported in their paper. We further show that the strong evidence against conditional independence is not due to the large N of the Finnish mobility table alone, as the same critique applies to much smaller sub-samples drawn randomly from the data. We then demonstrate with some illustrative outflow mobility rates that the grandparents effect in social mobility in Finland is not only statistically significant, but is also of substantive importance. Finally, we discuss the two ‘lagged ’ effects reported in Erola and Moisio (2007), and show that they fail to capture much of the net GC association.

We examine social mobility patterns over three generations of family members, using data collected in three British birth cohort studies. We show that for both men and women, and for all three birth cohorts, absolute mobility rates (i.e. total, upward and downward mobility rates, and inflow and outflow mobility patterns) in the partial parents–children mobility tables vary substantially by grandparents’ social class position. As regards relative mobility pattern, we show, again for all cases that we consider, that there is a statistically significant net association between grandparents ’ and grandchildren’s class positions, after parents ’ social class has been taken into account. Unlike the more complex association between the class positions of grandparents and parents, or between parents and children, the net grandparents–grandchildren association can be summarised by a single uniform association parameter. Net of parents ’ class position, the odds of grandchildren entering professional–managerial class rather than unskilled manual class is between two and three times better if their grandparents were in professional–managerial rather than unskilled manual class position. We thank David Firth and Louis-André Vallet for helpful advice and suggestion. An

We examine social mobility patterns over three generations of family members, using data collected in three British birth cohort studies. We show that for both men and women, and for all three birth cohorts, absolute mobility rates (i.e. total, upward and downward mobility rates, and inflow and outflow mobility patterns) in the partial parents–children mobility tables vary substantially by grandparents’ social class position. As regards relative mobility pattern, we show, again for all cases that we consider, that there is a statistically significant net association between grandparents ’ and grandchildren’s class positions, after parents ’ social class has been taken into account. Unlike the more complex association between the class positions of grandparents and parents, or between parents and children, the net grandparents–grandchildren association can be summarised by a single uniform association parameter. Net of parents ’ class position, the odds of grandchildren entering professional–managerial class rather than unskilled manual class is between two and three times better if their grandparents were in professional–managerial rather than unskilled manual class position. We thank David Firth and Louis-André Vallet for helpful advice and suggestion. An