| The cognition of randomness consists of perceptual and conceptual components. One might be able to discriminate random from non-random stimuli, yet be unable to identify which is which. We show that the ability to distinguish random from non-random stimuli is superior to the ability to correctly identify which given stimuli are "random." We further evaluate the encoding hypothesis according to which the tendency of a stimulus to be labeled "random" varies with the cognitive difficulty of encoding it. For some stimuli, the difficulty of encoding fails to predict the probability of being labeled "random," providing evidence against one version of the encoding hypothesis (Chapter 2). Using different methods, we next replicate these findings. Moreover, we observe that the ability to distinguish repeating stimuli from random is reliably better than the ability to distinguish alternating stimuli from random, presumably due to the ease of encoding of repeating stimuli compared to alternating (Chapter 3). In addition, we examine the ability to produce stimuli as random, and find that the produced sequences are more biased towards alternations than sequences labeled as random (Chapter 4). Finally, we provide a set of experimental paradigms measuring the sensitivity to randomness without invoking the notion of randomness or related concepts (Chapter 5). Overall, we present a comprehensive investigation of randomness cognition and findings suggesting a gap between the perception and the conception of randomness. |
