When given the chance to choose between two tasks, one will more likely choose the easier, less demanding task. This effect has been shown in various domains and referred to as the law of minimum effort or demand avoidance. Kool and colleagues (2010, 2013) designed the demand selection task (DST) and showed that most of their participants exhibited clear demand avoidance. We attempted to replicate and extend their results in a series of three studies. Here we argue that DST confounds demand detection and demand selection, which weakens its ability to reliably measure demand avoidance in different populations. In our first study, most participants did not show reliable demand avoidance and those who showed it had higher working memory capacity. The following two studies aimed to de-confound the two processes. We define a new measure of demand avoidance that affords a more robust estimation of demand avoidance in different populations.