Directional and Non-directional Hypothesis Testing: A Survey of SIG Members, Journals, and Textbooks.

The use of directional and nondirectional hypothesis testing was examined from the perspectives of textbooks, journal articles, and members of editorial boards. Three widely used statistical texts were reviewed in terms of how directional and nondirectional tests of significance were presented. Texts reviewed were written by: (1) D. E. Hinkle, W. Wiersma, and S. G. Jurs (1994); (2) G. V. Glass and K. D. Hopkins (1996); and (3) R. C. Sprinthall (1990). All three focused on nondirectional tests of significance. While all three texts introduced one-tail (directional) tests of significant, two downplayed its significance. Research methods texts reviewed were by: (1) J. H. McMillan (1992); (2) R. E. Slavin (1992); and (3) R. C. Sprinthall, G. T. Schmutte, and L. Sirois (1991). Directional hypothesis testing was not given as much respect as nondirectional in any of the three texts. A review of 11 issues of "School Science and Mathematics" and 2 issues of the "Journal for Research in Mathematics Education" found nondirectional tests usually used, although the authors made directional conclusions. A survey completed by 10 members of the editorial board of "Multiple Linear Regression," members of the Special Interest Group (SIG), showed strong feelings by 3 board members that directional hypothesis testing should be thoroughly understood and used appropriately. Attachments present three text examples. (Contains three tables and nine references.) (SLD)