Bentley et al. propose a two-dimensional map that aims to describe decision making in the big-data era. We believe that by analyzing how the big-data era – together with the omnipresence of the Internet and the increasing interconnectedness of people via social networks – may shape decision making, Bentley et al. have chosen a highly relevant and timely topic. However, we think that the east–west axis, which contrasts independent and socially influenced decision making, is not a suitable dimension to use in this context because it is not well grounded in the field of psychological decision research. We identify two shortcomings: First, the type of behavior that they are trying to classify is ambiguously defined. Second, the notion of independent decision making is not empirically tenable.

First, much of what Bentley et al. refer to as “decision making” is not decision making proper as the term is used in the pertinent literature on psychological decision-making research (Baron Reference Baron2007; Hardman Reference Hardman2009; Kahneman & Tversky Reference Kahneman and Tversky2000; Koehler & Harvey Reference Koehler and Harvey2004) or behavioral economics (Camerer et al. Reference Camerer, Loewenstein and Rabin2004; Kahneman Reference Kahneman2003; Loewenstein Reference Loewenstein2007). Most importantly, Bentley et al. confound decision making and learning, which is, surprisingly, a fact that they acknowledge themselves: “we blur the distinction between learning and decision making” (target article, sect. 2, para. 7). Psychologically, however, decision making and learning involve fundamentally different processes. A decision-making process anticipates the future, whereas learning generates our memories. The acts of anticipating the future (Gilbert & Wilson Reference Gilbert and Wilson2007) and remembering the past (Loftus & Pickrell Reference Loftus and Pickrell1995) are prone to systematic biases. Furthermore, human decision making is notoriously resistant to learning from experience (Brehmer Reference Brehmer1980; Kahneman Reference Kahneman2011), leading to a multitude of non-optimal choices. Many of Bentley et al.’s signature patterns are patterns of optimal behavior, such as microeconomic utility maximization or the ideal free distribution. Such patterns may emerge under very specific conditions (environmental regularity, constant feedback) but are usually not present in situations of individual choice. In fact, we think that a more appropriate mapping would include learning and decision making as orthogonal dimensions, with the learning axis contrasting basic operant conditioning at one end with socially mediated learning at the other end.

Second, the contrasting notions of independent versus socially influenced decisions seem ill-founded from a psychological perspective. Assuming an independent decision maker in Bentley et al.’s sense invokes the traditional model of a selfish utility-maximizing homo oeconomicus who is equipped with a set of fixed and immutable preferences. This model has been rendered untenable by accumulating empirical evidence (Akerlof & Shiller Reference Akerlof and Shiller2009; Ariely Reference Ariely2008; Gilovich et al. Reference Gilovich, Griffin and Kahneman2002; Kahneman Reference Kahneman2003, Reference Kahneman2011; Loewenstein Reference Loewenstein2007). Human preferences, which are the primitive elements of decision making, are anything but fixed and stable. A major finding from behavioral decision research is that preferences are constructed rather than given and are highly dependent on the context. Studies on constructive processes (Lichtenstein & Slovic Reference Lichtenstein and Slovic2006), as well as studies on preference reversals and framing effects (Kahneman & Tversky Reference Kahneman and Tversky2000; Tversky & Simonson Reference Tversky and Simonson1993; Tversky & Thaler Reference Tversky and Thaler1990) have demonstrated that preferences are subject to a large variety of contextual factors. An example is the ubiquitous role of anchors and reference points in determining utilities (Kahneman & Tversky Reference Kahneman and Tversky1979; Ordónez et al. Reference Ordónez, Connolly and Coughlan2000; Thaler & Sunstein Reference Thaler and Sunstein2008); for example, by simply changing the reference point (the default) from opt-in to opt-out, the preference concerning organ donation can be changed by 40% in favor of donation (Johnson & Goldstein Reference Johnson and Goldstein2003). In sum, preferences are always shaped by a multitude of contextual influences, many of which are social in origin.

If the western end of Bentley et al.’s east–west dimension cannot be adequately characterized as independent decision making, then we must ask what defines its opposite. Bentley et al. label the eastern pole as socially influenced decision making where people thoughtlessly copy what others do. But is this the opposite of independent decision making? We view this as just a special, though common, case of a constructed preference: In many situations, the modal behavior of others may serve as a natural reference point; and relying on what others do is just an instance of a fast and frugal heuristic (Gigerenzer et al. Reference Gigerenzer, Czerlinski, Martignon, Gilovich, Griffin and Kahneman2002). We believe that it is not so much social influence that characterizes the decision-making processes at the eastern pole – because social influence is also present at the western pole – but their more intuitive character.

What then could serve as a psychologically founded decision-making dimension? We suggest that any realistic model of human decision making should be based on the roles of affect and emotions (Damasio Reference Damasio1994; Loewenstein & Lerner Reference Loewenstein, Lerner, Davidson, Goldsmith and Scherer2003; Pfister & Böhm Reference Pfister and Böhm2008; Zeelenberg & Pieters Reference Zeelenberg, Pieters, DeCremer, Zeelenberg and Murnighan2006). Emotions shape human preferences, and they provide the crucial link between deciding and acting (Böhm & Pfister Reference Böhm and Pfister2000; Pfister & Böhm Reference Pfister and Böhm2008; Zeelenberg & Pieters Reference Zeelenberg, Pieters, DeCremer, Zeelenberg and Murnighan2006). Furthermore, a look at the functional roles of emotions suggests that individual as well as social decisions can be mapped onto specific emotions (Pfister & Böhm Reference Pfister and Böhm2008; Reference Pfister and Böhm2012). Some emotions are simple immediate affective reactions, such as joy or disgust, whereas others are cognitively saturated, such as guilt or envy.

We suggest that decision making be represented by a dimension that runs from deliberate/emotionally complex to intuitive/emotionally simple, a distinction emphasized in current dual-system approaches (Evans Reference Evans2008; Kahneman Reference Kahneman2011). A decision maker may rely either on an effortful process of deliberation and reasoning or on intuitions. Both modes of decision making can be advantageous or misleading, depending on the circumstances (Gigerenzer Reference Gigerenzer2007; Hogarth Reference Hogarth2010). We speculate that in a big-data era, it will become a critical issue for decision makers to select the appropriate mode, as the two modes often conflict. An Amazon purchase recommendation may superficially conform to a person's intuitive preferences, but may be rejected after some deliberate reasoning. On the other hand, when facing an unmanageable number of options, a deliberate decision might not be feasible, as emphasized by Bentley et al., thus raising the issue of how to educate our intuitions to survive in times of limitless choices.