In the target article, Bentley et al. have created a multiscale “map” in an attempt to capture the essence of decision making along two axes: social influence and transparency of payoffs. Grounded in discrete-choice approaches and with a few simplifying assumptions (sect. 2, para. 7), this concept tool sheds light on understanding different domains of human decision making. In what follows, we argue that this map fails to capture the relationship between social influence and payoff transparency.
Bentley et al. have interpreted their model as a “principal-components representation” (target article Abstract), which assumes that two axes are orthogonal to each other. However, it is well documented in social psychological literature that social influence and payoff transparency influence each other to a great extent. When two parameters are correlated with each other, one axis should not be perpendicular to the other axis (DeVellis Reference DeVellis2012, p. 142). Thus, the map as well as the formula at the start of section 2 should be revised to accommodate the hypothesized angle that represents the covariance of the two key parameters. Even though Bentley et al. did acknowledge this interaction in some places (e.g., “Intuitively, what is happening here is a pile-up of correlated behaviors caused by the interaction of social influences coupled with strong enough intensity of choice”; target article, sect. 2.2.1, para. 2), when they compare their map to a Google map (sect. 1, para. 6), they run the risk of misleading the readers into believing that the two axes are perpendicular to each other.
Decision-making is a dynamic process in which transparency and social influence invariably interact with each other. Under social influence, even the most transparent decision-making task becomes ambiguous. For example, in a classic conformity experiment (Asch Reference Asch and Guetzkow1951), participants were instructed to complete a perceptual task in which they had to match the length of a given line with one of three comparison lines. Although the correct judgments were easy to make, 75% of the participants made an incorrect judgment in at least one trial when all the confederates gave the same wrong answer. Thus, social influence can mask even the most transparent decision. Under social influence, even simple tasks like line comparisons are not as transparent as when there is no social influence (Asch Reference Asch and Guetzkow1951; Reference Asch1952; Reference Asch1956). Moreover, other studies have shown that social influence could change one's cost–benefit estimation of a decision (Louis et al. Reference Louis, Taylor and Douglas2005).
On the other hand, people are more likely to seek social influence when facing opaque decisions than when facing transparent decisions (Deutsch & Gerard Reference Deutsch and Gerard1955; Stasser & Davis Reference Stasser and Davis1981). Deutsch and Gerard (Reference Deutsch and Gerard1955) also asked participants to compare lines just like Ash did in the aforementioned study. In half of the trials (visual condition), the lines were physically present when participants and confederates made their judgments. Thus, in this condition, the decisions are fairly transparent. However, in the other condition, the lines disappeared before the participants had the opportunity to make judgments; hence, the decisions depended on memories and were less transparent than decisions in the first condition. Results showed that participants in the second condition were more likely to conform to social influence than those in the first condition.
When people have little knowledge about what to base their decisions on, it is helpful to imitate the successful judgments or to average the judgments of others to exploit the “wisdom of crowds” (Gigerenzer & Gaissmaier Reference Gigerenzer and Gaissmaier2011). Findings from neural imaging researches support this claim by showing that social information processing and decision making have shared neural substrates (Tomlin et al. Reference Tomlin, Nedic, Prentice, Holmes and Cohen2013). Opaque decisions drive people to seek social influence.
We recently conducted a survey (Zhou Reference Zhou2013) to examine whether transparency and social influence correlate with each other in daily decision making. We asked 55 participants to think about one upcoming decision they have to make in real life. Participants wrote down keywords best describing this decision. Next, participants rated how transparent the decision payoff is to them and how much social influence they are under in making the decision on an 11-point scale (0 = not at all; 10 = very much). The order of these two questions was counterbalanced so that half of the participants rated the transparency first and half of them rated the social influence first. Regardless of which question was asked first, social influence and payoff transparency turned out to be negatively correlated, (r(53) = −.71, p < .0005). These results support our argument that social influence and transparency interact with each other to a great degree and cannot be considered as independent dimensions.
In conclusion, we agree that Bentley et al. have provided a comprehensive map to evaluate collective behaviors in the big-data era. We also agree that this map will lead to new promising hypotheses on human decision making. However, Bentley et al. have not represented the interaction of the two axes on their map in a precise manner that would reflect the factual nature of the interaction. The accurate depiction of the covariance of social influence and payoff transparency is critical because it exerts direct impact on decision making in the digital age. The map proposed in the target article would function more quantitatively and accurately if it were revised taking the interaction between the two dimensions into account.
You have
Access
In the target article, Bentley et al. have created a multiscale “map” in an attempt to capture the essence of decision making along two axes: social influence and transparency of payoffs. Grounded in discrete-choice approaches and with a few simplifying assumptions (sect. 2, para. 7), this concept tool sheds light on understanding different domains of human decision making. In what follows, we argue that this map fails to capture the relationship between social influence and payoff transparency.
Bentley et al. have interpreted their model as a “principal-components representation” (target article Abstract), which assumes that two axes are orthogonal to each other. However, it is well documented in social psychological literature that social influence and payoff transparency influence each other to a great extent. When two parameters are correlated with each other, one axis should not be perpendicular to the other axis (DeVellis Reference DeVellis2012, p. 142). Thus, the map as well as the formula at the start of section 2 should be revised to accommodate the hypothesized angle that represents the covariance of the two key parameters. Even though Bentley et al. did acknowledge this interaction in some places (e.g., “Intuitively, what is happening here is a pile-up of correlated behaviors caused by the interaction of social influences coupled with strong enough intensity of choice”; target article, sect. 2.2.1, para. 2), when they compare their map to a Google map (sect. 1, para. 6), they run the risk of misleading the readers into believing that the two axes are perpendicular to each other.
Decision-making is a dynamic process in which transparency and social influence invariably interact with each other. Under social influence, even the most transparent decision-making task becomes ambiguous. For example, in a classic conformity experiment (Asch Reference Asch and Guetzkow1951), participants were instructed to complete a perceptual task in which they had to match the length of a given line with one of three comparison lines. Although the correct judgments were easy to make, 75% of the participants made an incorrect judgment in at least one trial when all the confederates gave the same wrong answer. Thus, social influence can mask even the most transparent decision. Under social influence, even simple tasks like line comparisons are not as transparent as when there is no social influence (Asch Reference Asch and Guetzkow1951; Reference Asch1952; Reference Asch1956). Moreover, other studies have shown that social influence could change one's cost–benefit estimation of a decision (Louis et al. Reference Louis, Taylor and Douglas2005).
On the other hand, people are more likely to seek social influence when facing opaque decisions than when facing transparent decisions (Deutsch & Gerard Reference Deutsch and Gerard1955; Stasser & Davis Reference Stasser and Davis1981). Deutsch and Gerard (Reference Deutsch and Gerard1955) also asked participants to compare lines just like Ash did in the aforementioned study. In half of the trials (visual condition), the lines were physically present when participants and confederates made their judgments. Thus, in this condition, the decisions are fairly transparent. However, in the other condition, the lines disappeared before the participants had the opportunity to make judgments; hence, the decisions depended on memories and were less transparent than decisions in the first condition. Results showed that participants in the second condition were more likely to conform to social influence than those in the first condition.
When people have little knowledge about what to base their decisions on, it is helpful to imitate the successful judgments or to average the judgments of others to exploit the “wisdom of crowds” (Gigerenzer & Gaissmaier Reference Gigerenzer and Gaissmaier2011). Findings from neural imaging researches support this claim by showing that social information processing and decision making have shared neural substrates (Tomlin et al. Reference Tomlin, Nedic, Prentice, Holmes and Cohen2013). Opaque decisions drive people to seek social influence.
We recently conducted a survey (Zhou Reference Zhou2013) to examine whether transparency and social influence correlate with each other in daily decision making. We asked 55 participants to think about one upcoming decision they have to make in real life. Participants wrote down keywords best describing this decision. Next, participants rated how transparent the decision payoff is to them and how much social influence they are under in making the decision on an 11-point scale (0 = not at all; 10 = very much). The order of these two questions was counterbalanced so that half of the participants rated the transparency first and half of them rated the social influence first. Regardless of which question was asked first, social influence and payoff transparency turned out to be negatively correlated, (r(53) = −.71, p < .0005). These results support our argument that social influence and transparency interact with each other to a great degree and cannot be considered as independent dimensions.
In conclusion, we agree that Bentley et al. have provided a comprehensive map to evaluate collective behaviors in the big-data era. We also agree that this map will lead to new promising hypotheses on human decision making. However, Bentley et al. have not represented the interaction of the two axes on their map in a precise manner that would reflect the factual nature of the interaction. The accurate depiction of the covariance of social influence and payoff transparency is critical because it exerts direct impact on decision making in the digital age. The map proposed in the target article would function more quantitatively and accurately if it were revised taking the interaction between the two dimensions into account.
ACKNOWLEDGMENTS
The authors wish to thank Boya Yu and Kornel Tomczyk for their helpful comments on an earlier draft of this commentary. This work was supported by the Key Program and General Program of National Natural Science Foundation of China (Grants 91124004 and 31171002) and Distinguished Young Scholars Program (No. 13wkpy87).