Cesario identifies three broad concerns regarding how social psychologists design their experimental research examining individuals' displays of social biases. These include critiques regarding: (1) limited information about the targets, such as presenting targets that are similar on many social dimensions except for one attribute (missing information flaw); (2) oversight of factors other than social bias that might contribute to the outcome of interest (missing forces flaw); and (3) whether results translate to real-world scenarios and decision-making (missing contingencies flaw). These critiques are not limited to social psychology and are akin to Dahl's (Reference Dahl2017) critiques of developmental research. When designing developmental studies, researchers control what participants see or experience and often what the possible responses are. Consequently, researchers make assumptions (what Dahl referred to as ecological commitments) about how children think, behave, and emotionally respond. They also postulate what experiences children have outside of the lab setting that contribute to developmental changes. Without actually observing individuals in their natural environment, researchers need to be particularly careful about whether their findings have ecological validity.

Cesario concludes the article by calling for a different approach to investigating social biases. The proposed methodology includes: (1) learning how decision-making ensues in real-world settings and what training/modeling occurs to support this process; (2) assessing inequities that members of particular social groups experience and what disparities beyond categorical membership perceivers consider in the decision-making process; and (3) using this information to create experimental studies. These suggestions echo Dahl's (Reference Dahl2017) call to use experimental and naturalistic methods in a complementary manner for understanding children's development. For example, Rennels and Langlois (Reference Rennels and Langlois2014) compared 3- to 11-year-olds' explicit biases based on facial attractiveness, gender, and race and found that biases based on girls' facial attractiveness were the most robust. In this study, participants saw faces of two children who differed in attractiveness, gender, or race but had similar attributes otherwise. Their task was to assign positive and negative attributes to the children depicted. In the forced choice condition, participants had to choose one of the two children when assigning attributes. In the non-forced condition, participants could choose one of the two children, or both or neither child. Although the non-forced choice condition permitted more flexibility in how participants responded, the study provided no information regarding the targets other than appearance. This missing information flaw could be rectified by observing children in their natural environments where classmates' faces vary on more than one attribute and children have developed knowledge regarding the behavior of other children in the classroom. Bias could be assessed by documenting approach/avoidance behaviors and the positivity/negativity of interactions between children. If Rennels and Langlois’ (Reference Rennels and Langlois2014) results generalize to natural environments, then compared to other classmates, children should be most likely to avoid and negatively interact with low attractive girls, and most likely to approach and positively interact with high attractive girls. It would also be important to assess whether teachers/staff model such differential behavior when interacting with children in the classroom who differ in attractiveness.

In terms of the missing forces flaw, a developmental example is the interpretation of children's gender biases and preferences for the same-gendered peers. It is well established that preschool and elementary children spend most of their time with the same-gendered peers who have similar interests in gender-typed activities (Martin et al., Reference Martin, Kornienko, Schaefer, Hanish, Fabes and Goble2013). Yet this gender bias varies based on children's reinforcement of gender stereotypes – boys typically adhere more strictly to gender roles than girls (Katz & Walsh, Reference Katz and Walsh1991), potentially because socialization teaches individuals to value masculine activities more than feminine activities. For example, school-aged children preferred a girl who engaged in masculine activities as a potential classmate more than a boy who engaged in feminine activities (Braun & Davidson, Reference Braun and Davidson2017). Children's friendship and activity preferences, therefore, not only reflect gender similarity but also their endorsement of gender stereotypes and evaluation of masculine and feminine activities. Naturalistic observations could complement these experimental findings and reveal additional missing forces by documenting factors contributing to the quality and length of interactions between the same and mixed gender peers.

Applying the critique of the missing contingencies flaw to developmental research could provide insight regarding why children's displays of explicit biased attitudes do not consistently translate across situations into discriminatory behavior (Dunham & Degner, Reference Dunham, Degner, Banaji and Gelman2013). One experimental contingency often overlooked when examining children's biases is the extent to which their usual real-world scenarios might incentivize them to express or control their biases. For instance, when a researcher told 6- to 10-year-olds that other adults and children would see their responses to an explicit racial attitude questionnaire, those children showed less explicit bias than children who were told their responses would not be shared (Rutland, Cameron, Milne, & McGeorge, Reference Rutland, Cameron, Milne and McGeorge2005). Thus, children can be externally motivated to inhibit displays of explicit bias. With development, children's internal motivation to inhibit displays of bias becomes contingent upon their understanding of others' attitudes and emotions (i.e., theory of social mind [ToSM]; Abrams, Rutland, Pelletier, and Ferrell, Reference Abrams, Rutland, Pelletier and Ferrell2009). High ToSM enables children to internalize their ingroup's social norms and inhibit displays of bias, whereas low ToSM requires external motivation to inhibit such displays (Fitzroy & Rutland, Reference Fitzroy and Rutland2010). Conducting naturalistic observations of how students, teachers, and administrators respond to bias in conjunction with reviews of school policies could provide relevant data regarding what incentivizes individuals to display or inhibit bias.

Our recommendation is to use naturalistic observations to complement, validate, or negate experimental developmental findings and is not limited to studying individuals. As per the missing contingencies example, these recommendations should extend beyond assessments of individual level bias to consider participants' real-world settings. For example, institutional factors, such as a school's diversity, equity, and inclusion policies and actions, are typically not included in developmental explanatory models of biased behavior. Often, only the school's racial makeup is provided (e.g., McGlothlin and Killen, Reference McGlothlin and Killen2010). We encourage developmental researchers to use naturalistic settings to enhance our understanding of when, why, and how children display bias in real-world scenarios. Theories derived from such observations could then be experimentally tested (Dahl, Reference Dahl2017).