Yes, our techniques for studying social disparities have some methodological weaknesses. But Cesario says something much stronger than this. Regarding two decades' research across many dozens of scholarly projects, he says in his abstract that “the current research tradition should be abandoned” (abstract). Now that's a conclusion! But iconoclasm is merited only if the argument is smashing. I will show two central flaws in Cesario's reasoning, either of which neutralizes his ambitious conclusion. Importantly, I will grant (for the sake of argument) Cesario's interpretive claims about the empirical literature. My two objections are instead about the logic of Cesario's argument; I will show that even if he is right about how to read these experiments, it is premature to recommend abandoning, or even drastically revising, the research tradition.

First, consider Cesario's claim that laboratory studies of bias strip away context from real-world decisions (the flaws of “missing information” and “missing contingencies,” in Cesario's terms). Let's grant this for the sake of argument. How does it support the conclusion that the current research tradition should be abandoned? Because, Cesario says, such artificially barren laboratory decisions cannot predict real-world decisions.

But Cesario is wrong to assume that decision-prediction is the only socially relevant use of this research. He does concede that the research bears on the apparently anodyne question of “the function and process of storing and using categorical information” (sect. 5, para. 1). But there is something else which directly touches on the social questions that researchers take themselves to be addressing. It is this: These laboratory studies show that (at minimum) people tend to treat social categories like race and gender as arbitrary-decision resolvers. And that is an important fact to study.

To see the point, recall the fable of Buridan's ass. Faced with two piles of hay, each equally tempting, the donkey starves to death for lack of reason to resolve an arbitrary choice. Human agents have ways of avoiding this fate – we might flip a coin, or perhaps favor whatever is closest to our dominant hand. That's perfectly fine, so far. But our choice of arbitrary-decision resolver can have ethical implications.

Imagine you have two children and you have just won a sweepstakes that entitles them to random items from an expensive toy catalog. You are given a list of lot numbers and told to divide them up between the kids. You have no idea which toy is represented by which lot number (that's the “fun” part, according to the toy manufacturer). If you had this information, of course, you might decide which child would like which toy more. But you don't. How should you resolve this arbitrary decision?

Here's a wrong answer: give 80% of the toys to one child, and 20% of the toys to the other child. This is the wrong answer because it displays inappropriate favoritism among your children. And it's no defense to insist that, because the choice was arbitrary, your decision resolution practice can be anything you want. Sometimes, how we choose to resolve an arbitrary decision reveals a great deal about the respect and care we have for the people affected. (Seriously, ask your kids.)

So, even if Cesario is right that these laboratory studies are arbitrary-decision contexts, that doesn't eliminate their social and political importance. It is important to know whether certain groups are implicitly treated as “less-than” even in arbitrary contexts. Ethicists have recently demonstrated how systemic derogation of a group constitutes disrespect even if it leads to no further consequences (Basu, Reference Basu2019). Further, even when stereotypes are ostensibly supported by statistical group regularities (as Cesario suggests at times), this doesn't prevent individual decisions from being morally risky (Bolinger, Reference Bolinger2020; Moss, Reference Moss2018). All of which means this research tradition is valuable even if Cesario's criticisms are right.

My second objection to Cesario's logic concerns his apparent theory of how social scientists should synthesize reasoning across micro- and macroscopic causal phenomena. Here's what I mean. We have strong evidence of a micro-scale phenomenon: bias in lab conditions. We also have strong evidence of a macro-scale phenomenon: systemic outcome inequities in housing, employment, and policing. What we don't (yet) have is conclusive evidence of the micro-to-macro causal linkages between these two phenomena – though researchers are working on it (Mallon, Reference Mallon2021). Finding those linkages will take a long time, given that the causal system is enormously complicated. While that work is ongoing, the approach is especially vulnerable to skeptical challenges from alternative causal explanations. Cesario presents one: group differences. He suggest (in his “missing forces” argument) that social scientists must take more seriously the possibilities that Black citizens simply are more connected to violent crime, or that women simply are less qualified in science, technology, engineering, and mathematics (STEM) fields.

The problem is that Cesario overemphasizes the significance of these alternative theories. To see this point, consider the parallel to climate change skepticism. There we have another micro-scale phenomenon (thermal properties of carbon) and another macro-scale phenomenon (historical change in average global temperature), with complex and still not-fully-understood causal linkages between them. The skeptic presents an alternative explanation: natural epochal temperature cycles. The skeptic insists we cannot focus attention on carbon emissions until we have nailed down our micro-to-macro causal linkages and ruled out their alternative explanation.

Cesario isn't doing exactly this, but his argument is not too far off. If his point were simply that a complete science of social disparity will ultimately need to rule out group differences (just as complete climate science needs to rule out temperature cycles), then fair enough. But he goes far beyond this when he suggests the need to radically restructure, or even “abandon,” the existing research paradigm. Research on social decision biases is only two or three decades old. Demanding airtight causal demonstration from it this early is comparable to judging the theory of anthropogenic climate change by the state of science in 1990.