A researcher has a series of Likert-scaled 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 assumption of linear relationships. Is this ethical?