As an alternative to other techniques, such as ordinary regression analysis, Kruskals Driver Analysis for relative importance may produce misleading results if there is no data or when variable data is strongly connected to each other (which is typical of research data).