Appendix C
Study Populations in Research on Learning
Often, groups of learners are categorized according to shared attributes, such as age, ethnicity, socioeconomic status, achievement profile, or, in some cases, a learning disability. Generalizing across groups of people is important for understanding trends, such as the effects of poverty on learning and development. Generalizing is also important for building evidence-based theories of learning and teaching. If every individual is treated as completely different from every other person, it would be an intractably complex task to make assertions about the types of teaching interventions that are most effective for the majority of learners or for teaching a class of students, or even for running a school. And yet, overgeneralizing can dangerously blind one to people’s complexities, nuances, and variability.
Although race, culture, and ethnicity are frequently used demographic variables in medical, sociological, psychological, and genetic research (Lillie-Blanton and Laveist, 1996), attempting to fit individuals within a single descriptor of race or culture and then generalizing the results of the research to a broader population is highly problematic for a number of reasons. First, all individuals function within culturally plural societies, so the identification of any singular cultural groups is likely to be inaccurate. Moreover, whereas a particular cultural group may share a set of values, the development of any one person is influenced by particular values as he experiences them within one or more of the microsystems that he inhabits (e.g., the home, school, workplace, or peer group). Thus, researchers should be very cautious about drawing conclusions about individuals based solely on the cultural group with which they are identified or affiliated. Likewise, because culture is not
a single construct, the influence of culture on learning will vary by individual and learning context (formal and informal educational settings).
Second, the means by which race is determined for research purposes is rarely described, but it most often occurs through the limited methods of observation of physical characteristics, self-identification, or a review of medical records (Dirette, 2014; Kaplan and Bennett, 2003; Williams, 1994). Often, the racial description categories in research studies are too broad to capture all of the possible descriptors accurately, or they simply do not fit an individual’s self-identification. For example, in the United States, a child who has one European American parent and one African American parent is often treated as African American (an example is President Barack Obama) and may identify as African American. But a growing trend is for mixed-race/ethnicity individuals to identify as biracial or multiracial. Moreover, some researchers argue that race is a product of the ways that people think about human differences (Appiah, 1992; Goldberg, 1993), while others note that because race plays a prominent role in human social practices, race is a social, not biological, construct (Appiah, 1996; Omi and Winant, 1994; Outlaw, 1995; Root, 1998; Zack, 1993).
Conceiving of race or ethnicity as a “box” to be checked on a form may bias instructors’ perspectives with respect to categories of students. At the same time, cultural practices are resources that every student and teacher brings to a learning situation. There is evidence that learning and identification with school are facilitated when teachers recognize and support the perspectives and practices of their students.
Forcing mixed-race individuals to select a single identity on surveys and questionnaires may also have negative consequences for survey respondents. Townsend and colleagues (2009) asked mixed-race participants to fill out two versions of a demographic questionnaire. In one version, only one racial background could be specified; in the second, respondents could select multiple races. Mixed-race respondents compelled to choose a single race scored lower on subsequent motivation and self-esteem questions than did mixed-race respondents who were allowed to select multiple races on the questionnaire. In short, the denial of multiracial identities had negative consequences for respondents’ self-perception, in the context of the questionnaire, which in turn may influence how they learn (see Chapter 3 for the role of motivation, identity, emotion, and culture on learning).
The use of race as a demographic variable in research is also highly problematic from a scientific perspective. The Human Genome Project has shown that race is neither a genetic nor a biological construct (Collins, 2004). Genetic research scholars have concluded that racial categories do not accurately reflect genetic diversity and that the use of race in genetics research should be phased out (Yudell et al., 2016). An example that demonstrates that genetic differences are not fixed by race is a comparison of the full genomes
of American scientists James Watson and Craig Ventner (both of European ancestry) with that of Korean scientist Seong-Jin Kim. The genetic sequences of Watson and Ventner shared fewer variations than either shared with that of Kim (Levy et al., 2007; Ahn et al., 2009).
The distinction between race and culture is not clear-cut. For instance, in the United States, racial categories reflect historical factors such as racism, inequality of opportunity, and social stratification. Although the United States is rapidly becoming a majority minority population—a population in which the majority of persons identify as being of one or more racial/ethnic minority groups—the status of these groups is highly dependent on the race construct as it has developed in the mainstream culture. Some behaviors and practices that may be conceptualized as “cultural” have developed within a context of adapting to social positioning (e.g., race, social class, ethnicity, and gender) and in response to social stratification mechanisms (e.g., racism, discrimination, and prejudice) as well as segregation. Under these circumstances, what appear to be cultural practices may emerge as a direct response to macro-level societal factors, such as discrimination (García Coll et al., 1996).
Although scholars strive to be unbiased and objective, doing so can lead to blindness to the cultural nature of one’s own constructions of reality. It is important to realize that the dominant Western scientific-cultural model is one perspective on reality and carries with it its own biases and assumptions. The perspectival nature of “scientific” views of reality is illustrated through the history of shifting dominant paradigms in any field of scientific inquiry. But practitioners typically apply this abundant evidence of cultural embeddedness only to their predecessors, not to themselves. The lack of awareness many social and educational scientists have of their own cultural perspectives has a number of counterproductive consequences.
For example, researchers may design studies that reflect the assumption that cognition and learning are universal processes, and they may further assume that therefore any study population will serve as well as any other. In an important review paper challenging this assumption, Henrich and colleagues (2010a) argued that the overwhelming choice of research subjects—namely Western, highly educated, industrialized, rich, and democratic (WEIRD) samples—for studies of perception and cognition has yielded findings that not only fail to generalize to the world at large but also are especially atypical and unrepresentative (see also Sears, 1986; Hartmann et al., 2013).
Second, this narrow research study population base (usually college students in the West) can bias how research questions are framed and limit the resulting conclusions. For example, there is a considerable body of research on effects of diversity in groups where the natural comparison or baseline has been a homogeneous (nondiverse) group. But as Apfelbaum and colleagues (2014) pointed out, this research typically leads to conclusions about the effects of diversity but remains blind to the possibility that homogeneity has
independent effects of its own on the baseline group’s processes of learning and cognition.
A third negative consequence of a researcher’s embeddedness in her own culture is that the proper focus of her research, the study materials developed, and the methods employed will tend to be guided by her own (cultural) intuitions and consequently are likely (inadvertently) to favor people from the same cultural group (Medin et al., 2010). For example, in developmental research it is common for researchers to interview children one at a time and to ask children questions to which the researchers assume they themselves know the answer. This may not be at all unusual in Western, middle class communities, but in many cultures, isolating a child from his peers and using known-answer questions may be very peculiar, to say the least. The methods as presented may be the same in the two cultures, but the methods as received could be dramatically different.
For efficiency and efficacy, this committee has taken a middle ground with respect to the literature we have examined. Although some of the literature cited in this report refers to groups of people or cultures, such as East Asian, Mayan, or Western (referring to North American and European), we recognize that it is useful to analyze and understand the relevant determinants, processes, and outcomes of learning and the relationships among these factors. But we also recognize that the responsible use of research in educational contexts includes taking the time to integrate the findings to understand what they mean for individual learners: learners who are whole, unique persons, who each live in particular contexts.