For Students

Using OSL to Learn Statistics

Psychological scientists are faced with numerous decisions when analyzing data. We must not only determine the proper analysis to conduct (e.g., chi-square or ANOVA; moderation or mediation), for which there is a usually a “correct” answer, but we must also make decisions for which there are ambiguous answers (e.g., post hoc tests vs. planned comparisons; include covariates or not).

These ambiguities often makes it difficult for researchers (like me and you) to learn how to perform data analyses. OSL is a great resource for helping us make these decisions. By putting ourselves in the shoes of other researchers, we get the opportunity to look at their data files, consider the various options in front of us, and then compare our decisions to those of the authors. In this way, OSL helps us conduct and interpret statistical analyses because we can always "check" our answers against those in published research. Below, I have some advice on how to get the most out of OSL labs.

Using OSL Labs

Each lab contains three elements: A published article, a data set, and an activity. I suggest that you first read the introduction and method sections of the published article. Focus your reading on paper's hypotheses, and how the authors operationalized their variables. Now, open the data file and look at the variables. Find which ones are raw variables and which ones are computed variables. Think back to the hypotheses and try to figure out which variables and which analyses would be most appropriate, given the data set in front of you. Only after you have gone through these steps should you open 'Activity' file and follow the steps provided.

Once you have completed each step, compare your results to those reported in the original article. If your results match, great! If they don't, try to figure out why. Did you analyze the wrong variables? Did you perform the wrong test? Were there follow-up tests reported? Was there a better analysis possible than the one the authors performed? Keep in mind that you are using the same data sets as the authors and that it is possible to reproduce all of the results.

After you have performed all of the analyses, and matched the results reported in the article, close the article file (so that you are not tempted to refer back to it) and write up your results. Only after writing your complete results section should you open the article file again and compare your write-up to the original paper. Were you complete in your description?