Driver analysis in market research typically involves using Correlation or some variant of Regression (e.g., see Driver Analysis) in an attempt to understand the importance of a set of explanatory variables in determining an outcome. For example, the explanatory variables may be measures of satisfaction with difference components of a service (e.g., price, service) and the outcome variable may be overall satisfaction.
Sometimes it is desired to perform driver analysis with repeated measures data. For example, rather than a single outcome variable, there may be multiple outcome variables (e.g., variables measuring satisfaction with multiple different brands), and multiple sets of explanatory variables (e.g., satisfaction with each brand on different components of satisfaction). A practical challenge that this presents is that the various statistical methods developed for driver analysis generally work on the assumption that there is only a single outcome variable and a single set of explanatory variables. The solution to this is to first stack the data and then perform the driver analysis.