Assessment of educational misclassification in register-based data on Finnish immigrants in Sweden

Abstract

Aim:

In population registers, information on completed schooling is either missing or misclassified for a large proportion of newly arrived immigrants. It is unclear how quickly the information is updated and whether misclassification, i.e., that the wrong level of education is recorded, biases empirical estimates.

Methods:

We use unique linked Swedish and Finnish register data to determine the extent of such mismeasurement. By running logistic regressions on zero earnings, we also illustrate how mismeasurement might influence the estimated effects of education on health or labour market outcomes.

Results:

We find a considerable bias in estimates based on Swedish records of educational attainment during immigrants’ first few years in the country. Misclassification is additionally very common, even when information on educational attainment exists.

Conclusions:

These findings suggest that research and policies using recently arrived immigrants’ completed schooling as a determinant of socioeconomic integration need to be interpreted with care.