Dichotomizing continuous predictors in multiple regression: a bad idea
Distribution of the number of citations over years.