Intergenerational mobility is often studied using survey data. In such settings, selective unit or item non-response may bias estimates. Linking Dutch survey data to administrative income data allows us to examine whether selective responses bias the estimated relationship between parental income and children's mathematics and language test scores in grades 6 and 9. We find that the estimates of these relationships are biased downward due to parental unit non-response, while they are biased upwards due to item non-response. In the analyses of both unit and item non-response, the point estimates for language and mathematics test scores point in the same direction but only one of the two relationships is significant. These findings suggest that estimates of intergenerational mobility based on survey data need to be interpreted with caution because they may be biased by selective non-response. The direction of such bias is difficult to predict a priori. Bias due to unit and item non-response may work in opposing directions and may differ across outcomes.