A researcher has a series of Likert-scale items and a measure of frequency of absenteeism. He uses each of the Likert-scaled items in a separate bivariate correlation with absenteeism without testing for the assumptions of linear relationships. Is this ethical?