On Skinner's pendulum: A framework for assessing s-frame hope

Abstract

Unsatisfied with the effects of behavioral economics’ i-frame, “technology of behavior,” Chater & Loewenstein call for a pendulum swing back to the s-frame, suggesting that such an approach offers a more hopeful path toward societal well-being. In this commentary, I offer a framework to think about this pendulum swing, as well as the scope – and limits – of this hope.

Scientists periodically raise alarms that everyone is doing something wrong. Indeed, B. F. Skinner's (1971) classic Beyond Freedom and Dignity begins this way. Although the world celebrated scientific improvements in medical, political, and housing systems, individual well-being is stalled. Skinner blamed this on science's failure to understand human behavior, writing, “Better contraceptives will control population only if people use them… Overcrowding can be corrected only by inducing people not to crowd, and the environment will continue to deteriorate until polluting practices are abandoned” (p. 4). Thus, he called for a “technology of behavior” (p. 5), which would focus on causal relationships between environments and actions, unleashing the potential of system-level advancements. In Chater & Loewenstein's (C&L's) terms, Skinner said: Let the pendulum swing away from s-frames. i-Frame science is our responsibility and our only hope.

This proposal's prescience is perhaps nowhere as evident as in behavioral economics. Rigorous investigation of the environment's effects on behavior (e.g., choice architecture; Thaler & Sunstein, 2009) has produced an i-frame science with remarkable effects across a range of systems. (Happily, this was done without sacrificing the freedom Skinner seemed ready to forego.)

Four decades later, C&L suggest that we are in danger of doing the wrong thing again. Now, the i-frame is a distraction. The s-frame is our best hope. The pendulum should swing back.

Given that we remain pessimistic about many of the same problems that Skinner lamented (Parker, Morin, & Horowitz, 2019), this hope should be welcomed. But to their credit, C&L don't obscure concerns that might trigger objections: First, s-frame efforts are expensive and slow. Will people support them? Second, corporate interests preserving s-frames are powerful. Can we aggregate enough popular support to compete? Finally, i-frame strategies have been widely deployed and celebrated (e.g., Benartzi et al., 2017). Why wouldn't we use the power of i-frames in cases where they may create positive change?

These objections are not reason for despair. Rather, answering each yields a framework that clarifies the nature of the hope the pendulum swing might offer. First, we can predict when individuals will accept s-frame policies: When a behavior creates costs and benefits that extend to others (i.e., creates high externalities), policy interventions tend to be more accepted than when costs and benefits are isolated to the individual (Fitzgerald, Lamberton, & Walsh, 2016). Second, we can predict when support will aggregate to the point that it might counterbalance corporate interests: Homogeneity regarding a behavior's desirability will support aggregation more than would heterogeneity. Analyzing behaviors based on these factors allows us to address the third objection. When low externalities and high heterogeneity make s-frame policies infeasible, we may recommend i-frame approaches with less concern that they crowd out s-frame options.

Using Table 1 to apply this framework, we understand why some s-frame policies highlighted by C&L deliver on s-frame hope. First, consider inadequate retirement savings. The majority of US citizens see retirement accounts as desirable, and the costs of retirement-age insolvency to society are clear (Institute of Medicine, 2012). As such, this behavior falls in the upper-right quadrant; s-frame policies will be seen as acceptable and capture broad support. Thus, it is unsurprising that s-frame pension reform has been successful, as described in C&L (target article, sect. 2.2, para. 19).

Table 1. Framework for analyzing i-frame and s-frame potential

As a contrasting example, consider snack choices. Rarely do we see personal snack choices as societally relevant – few externalities are salient. Further, the desire to restrain consumption is heterogeneous (Polivy, Herman, & Mills, 2020); not everyone regrets the choice of cake over fruit salad (Vosgerau, Scopelliti, & Huh, 2019). Thus, snacking falls in the lower-left quadrant. In this case, the mea culpa regarding i-frame work may be unwarranted. Effects may have been chilled because of the inhospitable system in which i-frame approaches were used, but experimenting with the display of nutritional information (Downs, Loewenstein) on snacks did not likely block a viable path to systemic change.

As a third example, consider exercise. Exercise's desirability has been broadly institutionalized (US Department of Health & Human Services, 2021), and the decision to exercise is driven by individual, not communal, self-determination (Ng et al., 2012). Thus, we may say that exercise falls into the upper-left corner. Here, i-frame interventions should resonate, whereas costly s-frame policies may receive less support. But if people agree that exercise is desirable, there is also hope for a certain type of s-frame policy: Those that can be undertaken without undermining individuals’ personal cost–benefit assessment. For example, local governments may create bike lanes through s-frame policies, offering gains in well-being with only very diffuse personal cost. Doing so may, in turn, maximize the effectiveness of i-frame interventions.

Finally, consider indoor smoking bans, an s-frame “success story.” This framework places indoor smoking in the lower-right quadrant. As C&L note, decades of research and government leadership made the undesirability of second-hand smoke obvious and the shared costs of the behavior clear. Thus, support for s-frame changes aggregated sufficiently to counter corporate interests, and the policy-level intervention was supported. Here, i-frame interventions alone would have nibbled around the edges, keeping policymakers and researchers from the focus on the s-frame that this behavior required.

This framework is only offered as a starting point; more questions may exist than answers. Where do behaviors fall? What shapes perceptions of externalities? How should we evaluate s-frame research's effects? In another four decades, we will know enough to decide again that some pendulum should swing differently. In the meantime, though, we have the responsibility – and the hope – to move beyond our entrenched s-frames. If we fail to do so, we may say more about our beliefs about our work's externalities and desirability than we do about science itself.