Intergenerational assimilation of UK immigrants in the labour market: A minor assumption with enormous implications for inference

Studies on the intergenerational assimilation of UK immigrants and their UK-born children have mainly relied on ethnicity and birthplace to measure nativity status because of data limitations. This would inevitably lead to classification errors in the sample. We present analytical results showing biases resulting from classification errors can go in any direction even when the sole regressor is a binary variable. The empirical analysis confirms such unpredictable implications for inference. A more accurate measure of nativity status based on parent's birthplace indicates the integration of immigrants might be different to what we would get from a measure prone to wrongly classifying individuals.