Getting Started in Formulating your Research Question
PAIR Project data can be set up for analysis fairly easily; however, deciding what questions to ask, what constructs to examine, and what analyses to do is not simple. The accessibility of the data may tempt the unwary to carry out a series of exploratory analyses, hoping some fascinating results will emerge from the effort. Such an approach is apt to produce frustration, intensified perhaps by glimmerings of excitement from "statistically significant," though ultimately theoretically trivial, findings.
The problems with "fishing" for results can be readily appreciated once it is recognized that the PAIR Project database includes variables measuring about 500 distinct constructs. The possibilities for analyses are legion - if all variables were correlated with each other, an intercorrelation matrix of more than a quarter of a million correlations would be created. Approximately 2,500 would be significantly associated at the .01 level by chance alone! How then can a person come up with a provocative yet tractable question using PAIR Project data? Obviously, there is no foolproof formula, and scientific creativity is not something that can be pre-packaged and distributed to anyone who wants to acquire it. In this series of pages, however, we will attempt to outline a series of steps that hopeful researchers may follow to formulate a well-grounded research question.
