Like other respondents to this thread, I also require students to collect data for their own research projects. I would rather see a student risk obtaining null results with a design that asks a new and interesting question than to replicate something that has been done a dozen times and always generates significant differences. Too many students assume that they must get statistically significant differences if they want their project to be "good." One of the first questions students ask about the requirements for the project is about how many participants they must collect data from for an "acceptable" project. I use this question to talk about the purpose of the individual project, which is to demonstrate that they know how to design an empirical study that poses an interesting and meaningful question and generates data that could provide an answer to that question. Evaluation of the quality of the study is based on the logic of the design and procedures, not on the outcome of the statistical test. I make clear that the time constraints for data collection in the course will almost necessarily force them to use sample sizes that are too small to give them enough power to detect all but the largest effects. I define an adequate sample size as one that has enough observations to enable them to conduct the appropriate statistical analysis (e.g., 4 or 5 observations per cell in an ANOVA). I do point out that they will have more fun writing their results and discussion sections if their manipulations work and generate significant differences, but that obtaining statistical significance is not a criterion for quality.
I like to think that removing the pressure for "significance" defuses the motivation to cook the data. Testing that belief is hard. I also include the following deterrents in the structure of the research project: 1. I have running discussions about their projects that, for a few students, begin in the first week of class. By mid-term, I have discussions going with every student in the class. Students really take ownership of their projects and are proud of their work when they finish. I think this sense of engagement undermines the motivation to create fake data. They really want an answer to their question and they don't want an answer that they just made up out of thin air. (Suggestions for how to test whether this really happens or whether I'm just delusional?) 2. I set aside two weeks of lab time just for data collection and data analysis on individual projects, so much of the data collection is quite public. 3. I require my students to document their data collection with copies of data sheets and signed consent forms. These could be faked as well, but it is one more deliberate act of deception. Claudia ________________________________________________________ Claudia J. Stanny, Ph.D. e-mail: [EMAIL PROTECTED] Department of Psychology Phone: (850) 474 - 3163 University of West Florida FAX: (850) 857 - 6060 Pensacola, FL 32514 - 5751 Web: http://www.uwf.edu/psych/stanny.html --- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED]
