From PsychWiki - A Collaborative Psychology Wiki
Revision as of 18:03, 9 July 2009 by Doug
- Helpful to run data in multiple ways to look for converging evidence of your results. When inconsistencies arise, track down the source.
- Its appropriate to test for simple effects even if no omnibus interaction.
- Data analysis should be planned BEFORE the study is conducted so as to ensure that the methodological design of the study permits the type of data analysis of interest to the investigator.
◄ Back to Research Tools mainpage