Which one under- or overestimates significance in the test of association?. Often times, I come across both choices in the selection of SD and SEM as a measure of spread, however, little is contributed by the reviewers even when an article is submitted for publication most times regarding statistical analysis. Meanwhile, tests of association in relation with a foretasted hypothesis is a sensitive point to misconstrue types I&II errors. When are these choices best employed, considerably and respectively?.