Courtney Snow Presentation Handout The Critiques of Hypothesis Testing 15 April 2003 The information in this presentation is from an article written by Joel Levin and Daniel Robinson entitled Further Reflections on Hypothesis Testing and Editorial Policy for Primary Research Journals in The Educational Psychology Review Vol 11 No 2 1999 Levin and Robinson answered the following questions in this article in response to an earlier article written by Bruce Thompson 1 What sort of language should be used in published articles when discussing the results of hypothesis tests T replace the word significant with statistically significant L R use statistical instead of other two or create a magazine of many different journals for lay readers 2 Is there a purpose of even using hypothesis testing T no because it does not come up with very useful information L R yes because it is provides evidence that one condition is better than another 3 Should the process of hypothesis testing be altered in any way to make it more useful and or practical T gives no comment or possible alternative yet he rejects its use L R want hypothesis testing to be reconstituted into a form in which replication distributional characteristics and selectivity play critical roles 4 How should the results of a hypothesis test be reported T wants universal reporting and interpretation of effect sizes L R want to instate a process that hypothesis testing results are used when giving the fact that an experimental outcome differs from that expected by chance and along with it effect size information is reported 5 Should effect sizes be published T yes they should be visible to the reader in each article L R afraid of possible mistakes when reporting figures 6 Should internal or external replication be used in an attempt to recreate the same results T external wants researchers to do true external replications L R internal reanalysis constrained by specific operational situational and data collector features of the particular study Do you believe that hypothesis testing is a completely valid and understandable concept that can be used in medical journal articles circle one Yes No Maybe