In the current set of experiments our goal was to test the hypothesis that individuals avoid courses of action based on a kind of metacognitive evaluation of demand in a Demand Selection Task (DST). Individuals in Experiment 1 completed a DST utilizing visual stimuli known to yield a dissociation between performance and perceived demand. Patterns of demand avoidance followed that of perceived demand. Experiment 2 provided a replication of the aforementioned results, in addition to demonstrating a second dissociation between a peripheral physiological measure of demand (i.e., blink rates) and demand avoidance. Experiment 3 directly tested the assumption that individuals make use of a general metacognitive evaluation of task demand during selections. A DST was utilized in a forced-choice paradigm that required individuals to either select the most effortful, time demanding, or least accurate of 2 choices. Patterns of selections were similar across all rating dimensions, lending credit to this notion. Findings are discussed within a metacognitive framework of demand avoidance and contrasted to current theories.