In this paper we examine the possibility of using the standard Kruskal-Wallis (KW) rank test in order to evaluate whether the distribution of efficiency scores resulting from Data Envelopment Analysis (DEA) is independent of the input (or output) mix of the observations. Since the DEA frontier is estimated, many standard assumptions for evaluating the KW test statistic are violated. Therefore, we propose to explore its statistical properties by the use of simulation studies. The simulations are performed conditional on the observed input mixes. The method, unlike existing approaches in the literature, is also applicable when comparing distributions of efficiency scores in more than two groups and does not rely on bootstrapping of, or questionable distributional assumptions about, the efficiency scores. The approach is illustrated using an empirical case of demolition projects. Since the assumption of mix independence is rejected the implication is that it, for example, is impossible to determine whether machine intensive project are more or less efficient than labor intensive projects.