On cognitive preferences and the plausibility of rule-based models

It is conventional wisdom in machine learning and data mining that logical models such as rule sets are more interpretable than other models, and that among such rule-based models, simpler models are more interpretable than more complex ones. In this position paper, we question this latter assumption by focusing on one particular aspect of interpretability, namely the plausibility of models. Roughly speaking, we equate the plausibility of a model with the likeliness that a user accepts it as an explanation for a prediction. In particular, we argue that—all other things being equal—longer explanations may be more convincing than shorter ones, and that the predominant bias for shorter models, which is typically necessary for learning powerful discriminative models, may not be suitable when it comes to user acceptance of the learned models. To that end, we first recapitulate evidence for and against this postulate, and then report the results of an evaluation in a crowdsourcing study based on about 3000 judgments. The results do not reveal a strong preference for simple rules, whereas we can observe a weak preference for longer rules in some domains. We then relate these results to well-known cognitive biases such as the conjunction fallacy, the representative heuristic, or the recognition heuristic, and investigate their relation to rule length and plausibility.

Extended Knowledge, the Recognition Heuristic, and Epistemic Injustice

This chapter argues that the interaction of biased media coverage and widespread employment of the recognition heuristic can produce epistemic injustices. It explains the recognition heuristic as studied by Gigerenzer and colleagues, highlighting how some of its components are largely external to the cognitive agent. Having connected the recognition heuristic with recent work on the hypotheses of embedded, extended, and scaffolded cognition, it argues that the recognition heuristic is best understood as an instance of scaffolded cognition. It considers the double-edged sword of cognitive scaffolding before using Fricker’s (2007) concept of epistemic injustice to characterize the nature and harm of these false inferences, emphasizing the Darfur Inference. Finally, it uses data-mining and an empirical study to show how Gigerenzer’s population estimation task is liable to produce Darfur Inferences. It ends with some speculative remarks on more important Darfur Inferences, and how to avoid them by scaffolding better.

The use of the recognition heuristic as an investment strategy in European stock markets

Purpose People often face constraints such as a lack of time or information in taking decisions, which leads them to use heuristics. In these situations, fast and frugal rules may be useful for making adaptive decisions with fewer resources, even if it leads to suboptimal choices. When applied to financial markets, the recognition heuristic predicts that investors acquire the stocks that they are aware of, thereby inflating the price of the most recognized stocks. This paper aims to study the profitability against the market of the most recognized stocks in Europe. Design/methodology/approach In this paper, the authors perform a survey and use Google Trends to study the profitability against the market of the most recognized stocks in Europe. Findings The authors conclude that a recognition heuristic portfolio yields poorer returns than a market portfolio. In contrast, from the data collected on Google Trends, weak evidence was found that strong increases in companies monthly search volumes may lead to abnormal returns in the following month. Research limitations/implications The applied investment strategy does not account for transaction costs, which may jeopardize its profitability given the fact that it is necessary to revise the portfolio on a monthly basis. Despite the results obtained, they are useful to understanding the performance of recognition heuristic strategies over a comprehensive time horizon, and it would be interesting to depict its viability during different market conditions. This analysis could provide additional information about a preferable scenario for employing our strategies and, ultimately, enhance the profitability of recognition heuristic strategies. Practical implications Through the exhaustive analysis performed here on the recognition heuristic in the European stock market, it is possible to conclude that no evidence was found for the viability of exploring this type of strategy. In fact, the investors would always gain better returns when adopting a passive investment strategy. Therefore, it would be wise to assume that the European market presents at least a degree of efficiency where no investment would yield abnormal returns following the recognition heuristic. Originality/value The main objective of this paper is to study the performance of the recognition heuristic in the financial markets and to contribute to the knowledge in this field. Although many authors have already studied this heuristic when applied to financial markets, there is a lack of consensus in the literature.

Measuring Age-Related Differences in Using a Simple Decision Strategy

According to the recognition heuristic, decision makers base their inferences on recognition alone, assuming that recognized objects have larger criterion values than unrecognized ones. Knowing that recognition is a valid cue and thus using the recognition heuristic should increase with age. This was tested in two experiments with preadolescents (N = 140), adolescents (N = 186), and adults (N = 78). The results show, as expected, a monotonic age-related trend in the improvement of domain-specific knowledge but, unexpectedly, a non-monotonic one for using the recognition heuristic. More specifically, use of the recognition heuristic increased from preadolescents to adolescents, but then dropped for adults.