Abstract
Conclusions about human behavior are primarily based upon observations from western, educated, industrialized, rich, and democratic (WEIRD) samples, especially from the United States. One consequence may be the promotion of assumptions that research findings from these populations are more generalizable to humankind than findings from non-WEIRD populations. We tested this with an archival study comparing the extent to which titles of over 5,000 published psychology articles specify samples’ racial/ethnic/national/cultural characteristics—a practice that implies constraints to generalizability. We observed that samples from the United States were less frequently specified in titles compared to both other WEIRD and non-WEIRD regions. Yet, samples from the United States (compared to other regions) were more frequently specified in titles if they referred to racial/ethnic/cultural minorities who may be perceived as exceptions to assumed generalizability of the White American population. These findings suggest that one consequence of a USA-centric sampling bias in psychology may be biased assumptions of (White) people from the United States as especially reflective of humankind.
As the scientific study of human nature, psychology seeks to produce conclusions about a broader population, often humanity itself, based on observations from a specific study sample. Given the diversity of humankind, understanding human nature should thus involve recruiting and studying appropriately diverse samples. Yet psychological science has been criticized for the persistent practice of overrepresenting and overgeneralizing findings from an extremely narrow subset of the human population comprising young adults drawn from western, educated, industrialized, rich, and democratic (WEIRD) populations (Henrich et al., 2010), particularly North Americans of European ancestry (Arnett, 2008; Rad et al., 2018; Sue, 1999; Van de Vijver, 2013). Here, we present an archival study that suggests another implication of this lack of diversity in sampling: a potential bias in perceptions of how generalizable psychological and behavioral findings may be to humanity based on the origin of the study sample.
We argue that whether titles of research articles qualify the racial, ethnic, or cultural backgrounds of their human subjects may be an important indicator of implicit assumptions of generalizability of findings to humankind. A scientific article’s title is the most basic unit for describing the objectives or findings of a project. Titles usually communicate the purpose or aim of the research along with its scope. Given their primacy and conciseness, titles also convey information about the study deemed as a priority. Importantly, titles may also contain information that highlight or constrain the generalizability of the findings by qualifying characteristics or origins of the sample (e.g., “The effect of X on Y among preschool-aged children”). Empirical studies containing observations and inferences assumed to be widely generalizable to humankind may be much less likely to qualify demographic characteristics or origins of their samples in titles. Conversely, studies involving conclusions that are considered to be culture- or population-specific may be much more likely to qualify sample characteristics such as race, ethnicity, nationality, or cultural background in the title (e.g., “The effect of X on Y among Chinese consumers,” as opposed to simply, “The effect of X on Y”). Qualifying sample characteristics in such cases may function to caution readers from inappropriately extending or applying the study’s conclusions to other populations.
The concern that empirical research publications in leading psychology journals do not adequately report or specify sample characteristics such as race, ethnicity, nationality, gender, or socioeconomic status is not new (Rad et al., 2018; Rozin, 2001; Simons et al., 2017). Recently, Rad and colleagues (2018) revealed that over 72% of abstracts published in Psychological Science (one of the leading journals in psychology) contained no information regarding the population being sampled. Simons and colleagues (2017) also cautioned about the widespread lack of published papers in psychology clearly identifying their target populations or constraints on the generality of their findings. A complementary concern is a sampling bias in psychology involving restricted levels of diversity among samples recruited for study, who are predominantly recruited from the United States (Arnett, 2008; Cheek, 2017; Nielsen et al., 2017). Along with the recognition that most psychology researchers are based in western institutions, this persistent sampling bias has contributed to recent calls for greater inclusiveness and internationalization of psychology as a discipline (Christopher et al., 2014; Hruschka et al., 2018; Van de Vijver, 2013). These practices and trends suggest that psychological research will involve samples from WEIRD populations (especially the United States) by default, unless otherwise stated by qualifying the origins of the sample in the study’s title or abstract.
Although these practices comply with effective communication according to Gricean Maxims (Grice, 1989), whereby redundancy (such as “default” information) should be avoided, we argue that lower frequencies of mentioning WEIRD samples in the title relative to a non-WEIRD sample could produce a bias—a potential assumption that findings of a study may be more readily generalizable to humankind if derived from participants from WEIRD societies (especially the United States) because its title does not contain qualification of sample origin or background. By contrast, if titles of published research studies involving participants from non-WEIRD populations (e.g., from Asia, Africa, Latin America) exhibit substantially greater frequencies in specifying or qualifying demographic characteristics of their sample (e.g., race, nationality, cultural background), the findings from those studies may be assumed as having more constrained generality and less readily reflecting human psychological nature. That is, articles relying on non-WEIRD samples may be unduly perceived as reflecting something “cultural” or society-specific, despite the aims and designs of the research seeking to investigate processes generally assumed to apply to other populations. 1
Given its potential impact on generality inferences, it is important to examine the potential bias of WEIRD (particularly the United States) samples being assumed as the “default” raised here. As an initial step to tackle this issue, the current study tests whether there are indeed wide disparities in frequencies of qualification of samples in journal article titles as a function of sample characteristics, such as whether WEIRD versus non-WEIRD samples were involved. Among the WEIRD samples, we also differentiate USA samples from other non-USA WEIRD samples (e.g., from Australia, Canada, the United Kingdom). Moreover, the United States has a diverse ethnic population. What would be the practice if a study involved ethnic minority groups within the United States (e.g., African Americans, Asian Americans)? Given that Americans are prototypically assumed as being “White” (Devos & Banaji, 2005), it is possible that studies involving racial/ethnic/cultural minority samples in the United States would qualify sample characteristics in the titles much more frequently compared to studies involving predominantly White American participants.
Current Study
If this “default” bias indeed exists, we should observe higher frequencies of non-WEIRD samples than WEIRD (U.S. or non-USA) samples being specified in the article titles (Hypothesis 1). In addition, given the “American = White” phenomenon (Devos & Banaji, 2005), we should also observe higher frequencies of specifying ethnic minorities in the article titles for studies conducted in the United States than the two other regions (WEIRD and non-USA WEIRD; Hypothesis 2). Finally, it would be important to examine the trends in these practices over years and to identify whether they are becoming more prevalent. If articles from non-WEIRD regions indeed violate the perception of reflecting the “default” sample and more often qualify sample origins in the title, then this practice could become normative and self-reinforcing—ultimately increasing in frequency over a period of many years (Hypothesis 3). To test these hypotheses, we conducted an archival study on the titles of articles published from a large and diverse number of psychology journals (almost 50 journals) across a period of 14 years.
A straightforward way to test these hypotheses would be to compare the probabilities of sample specifications (e.g., country origins, racial, ethnic minority) in the titles given the origins of samples (from non-WEIRD, non-USA WEIRD countries, and the United States). That is, we can calculate a probability by dividing the frequencies of sample specifications in article titles over the total number of articles involving participants from each of the three types of countries/regions. If the probabilities across the three types are not statistically different from each other, our hypotheses would be rejected. That said, given the massive number of journal articles that we would need to include in the content analysis in order to achieve representativeness, coding the actual characteristics of samples studied in every single article would be unrealistically demanding. We have therefore derived a less than perfect but arguably “practical” way to test our hypotheses. Instead of calculating probabilities, we focus on comparing the frequencies of titles that specify samples (or minority samples) from USA, non-USA WEIRD, and non-WEIRD countries.
This may be both a valid and practical approach for the following reasons. First, it has been established that most publications in psychology involved a vast overrepresentation of samples from WEIRD regions—mostly the United States (Arnett, 2008; Hruschka et al., 2018; Nielsen et al., 2017; Rad et al., 2018). Given this, if we find that the frequency of the sample qualifications for non-WEIRD samples is higher than those of non-USA WEIRD and USA samples, we can be reasonably certain that the probability of sample qualifications for non-WEIRD samples is much higher than that of the other two types. This method is “practical” for our purpose, although we cannot make the reverse argument given the low base rate of non-WEIRD studies (i.e., a lower frequency for non-WEIRD samples indicating that the probability of sample qualifications is lower for non-WEIRD samples than the other two types of samples).
Method
We conducted an archival content analysis of published research articles across journals in psychology to compare the frequency of sample origins being qualified in titles of publications based on their WEIRD status (from a non-WEIRD country, from a non-USA WEIRD country, or from the USA). Importantly, sampling is based exclusively on whether an article mentions sample origins in the title rather than all published articles from the targeted journals and time span. We separated the United States from other WEIRD countries (non-USA WEIRD), given prior indications that historical sampling biases favoring USA may be even more pronounced than other WEIRD regions (Adair & Huynh, 2012; Arnett, 2008; Rad et al., 2018).
First, we identified journals from diverse subfields of psychology based on 2012’s 5-year journal impact factor generated by the Journal Citation Reports database. The database categorized the journals into the following subfields of psychology: applied psychology, biological psychology, clinical psychology, developmental psychology, educational psychology, experimental psychology, multidisciplinary psychology, and social psychology. At least seven journals with the highest 5-year impact factors were selected from each subfield. However, journals that focused primarily on review/theory articles, a specific psychological condition or disorder (e.g., journals on autism or schizophrenia), animal research, or did not have publications going back at least 10 years were subsequently excluded. We added culturally oriented psychology journals into the selection and oversampled journals in social/personality psychology given that cultural and social psychology may be especially likely to include research with themes related to race, ethnicity, and culture that may be more relevant to questions of generalizability. In addition, we included some interdisciplinary journals with focus on psychological and behavioral processes (e.g., Appetite, Physiology & Behavior). This process resulted in a final total of 49 selected journals (see Table 1 for list of journals).
List of Journals (by Category) That Titles of Research Articles Were Extracted From for Coding and Analysis.
Note. “Number of extracted titles” refers to total types extracted from the respective journal from 2004 to 2017 that made any mention of race/ethnicity/nationality/culture. “Number of analyzed titles” refers to the titles that were identified to refer to race/ethnicity/nationality/culture to specify sample origin or characteristics and included in analyses for the current hypotheses.
Next, all article titles, except commentaries and corrections, published between 2004 and 2017 that contained general mentions of or named specific race, ethnicity, nationality, culture, or gender in any context in any context were extracted for each selected journal. This produced a total of 5,609 titles (titles referencing gender were excluded from the present analysis). From this pool of extracted titles that made any mention of race/ethnicity/nationality/culture, two independent raters coded each article title based on the sample origins and the nature of the research following a coding rubric (see Table 2 for the full coding rubric and examples of corresponding titles). Extracted and coded titles are available on Open Science Framework (https://osf.io/epbfg/; Cheon, 2020). The objective of this coding was to identify titles that mentioned race, ethnicity, nationality, or culture specifically to describe the origin or characteristic of the sample being studied. Thus, titles that mention sample characteristics not for the purpose of qualifying the study sample were assigned a separate set of codes and were excluded from analysis. These irrelevant mentions of race/ethnicity/nationality/culture were coded based on, but not limited to, general use of the words “race” or “culture,” references to languages (e.g., referring to French language rather than French samples), names of organizations, policies, or products that incidentally contained the name of a country/culture (e.g., reference to the American flag), and so on (see Table 2 for further examples). Furthermore, we excluded from analyses titles that mentioned sample origins in the context of a cross-cultural comparison (e.g., “X influences Y among Japanese but not Americans”), since such studies inherently assume observations from one sample may not generalize to the other, thereby necessitating qualification of sample origins in the title.
Descriptions and Examples of Codes Applied to Article Titles.
Note. These example titles do not represent actual titles of published articles. WEIRD = western, educated, industrialized, rich, and democratic.
Each title was assigned one code by each rater; therefore, no recurring or redundant titles were present in the analyses (Table 2). The two independent raters demonstrated an acceptable consistency (Cohen’s κ = .72) and resolved coding inconsistencies on specific titles through a discussion. Specifically, to test Hypothesis 1, research titles were coded based on whether the sample characterized a population from the United States (n = 307), other non-USA WEIRD regions (i.e., Canadian, European, Australian, etc.; n = 726), or non-WEIRD regions (Asian, African, Latin American, Middle Eastern, etc.; n = 1,055). Notably, our coding scheme for classifying WEIRD and non-WEIRD societies does not exactly align with the conventional definition of WEIRD. 2 For instance, we classified Australia as a WEIRD region although it is not “western,” while Latin America was classified as non-WEIRD although it is in the west. Furthermore, there may be some countries in Europe (considered WEIRD in our coding) that may not typically be considered rich or democratic, just as there may be countries in Africa and Asia (considered non-WEIRD in our coding) that could be described as educated, industrialized, and rich. We used the current classification of WEIRD and non-WEIRD for data reduction and also to be consistent with the structure that the WEIRD regions typically consist of majority groups of White/European ancestry while the non-WEIRD regions typically consist of majority groups of non-White/European ancestry. Based on these criteria, the “W” in WEIRD for the current study could more accurately be defined as “White” rather than “western.”
Importantly, if the title mentioned an ethnic/racial/cultural minority sample from these regions, it was assigned a different set of codes instead. To test Hypothesis 2, titles of published research that specified samples of minority groups were retained and coded based on the regions the minority sample originated from USA (n = 807), non-USA WEIRD (n = 74), and non-WEIRD (n = 64). This process formed a final sample of 3,033 article titles, on which the analyses were performed. Titles receiving other codes, such as those denoting cross-cultural comparisons or linguistics studies described above, were excluded in the present analyses.
Results
A χ2 test comparing the number of titles qualifying samples from the United States, non-USA WEIRD, and non-WEIRD regions revealed a significant difference in the frequency of published research titles qualifying sample characteristics from these regions, χ2(2, N = 2,088) = 403.88, p < .001. Pairwise comparisons further revealed that there were significantly fewer incidents of titles qualifying samples with origins from the United States (M = 14.7%) than from non-WEIRD regions (M = 50.5%), χ2(1, N = 1,362) = 410.80, p < .001, as well as from non-USA WEIRD regions (M = 34.8%), χ2(1, N = 1,033) = 169.95, p < .001. Furthermore, there were also significantly fewer titles qualifying samples from non-USA WEIRD than non-WEIRD origins, χ2(1, N = 1,781) = 60.78, p < .001. Supporting Hypothesis 1, there was a significantly lower frequency of titles qualifying samples with origins from WEIRD regions, with titles qualifying participants from the United States occurring the least frequently relative to other regions (Figure 1A).
The percentage of published research articles that qualified sample origins (A) and minority group samples (B) in the titles. There were significantly fewer published articles that qualified samples recruited from the United States as compared to non-WEIRD and non-USA WEIRD regions. However, the pattern was reversed when the study involved minority group samples, in which (non-White) minority groups from the United States were qualified at a significantly higher frequency than minority groups from non-WEIRD and non-USA WEIRD regions.
To test Hypothesis 2, a χ2 test was also performed to compare the number of titles qualifying minority group samples from the United States, non-USA WEIRD, and non-WEIRD regions. The analysis revealed a significant difference in the distribution of the number of titles among the three regions, χ2(2, N = 945) = 1,152.84, p < .001. Supporting Hypothesis 2, there were significantly greater incidents of titles qualifying the non-White minority group samples in the United States (M = 85.4%) compared to minority groups in non-WEIRD (M = 6.8%), χ2(1, N = 871) = 633.81, p < .001, and non-USA WEIRD (M = 7.8%) regions (Figure 1B), χ2(1, N = 881) = 609.86, p < .001. The number of titles qualifying samples of minority groups from non-WEIRD and non-USA WEIRD regions was not significantly different, χ2(1, N = 138) = .73, p = .40.
Analyses were performed to identify broader trends in the number of titles qualifying sample characteristics from each of the three regions (USA, non-USA WEIRD, and non-WEIRD) from 2004 to 2017. Results of the analyses are illustrated in Figure 2A and 2B. Offering initial support for Hypothesis 3, a notable upward trend in the frequency of titles qualifying the samples with origins from non-WEIRD and non-USA WEIRD regions were observed, with a trend for steeper frequency for non-WEIRD regions, whereas the frequencies of titles qualifying samples from the United States remained largely flat and unchanged across the examined period (2004–2017). When examining trends for titles that specify samples consisting of minority groups from these regions, the frequency of titles qualifying these characteristics remains largely stable across all three regions over time (2004–2017), with minority samples from the United States being qualified the most.
The trend illustrating the number of published research articles that qualified sample origins (A) and minority group samples (B) by each region from 2004 to 2017.
Discussion
The present research was motivated by two concerns raised about psychological science: (1) that researchers focus too narrowly on observations and conclusions drawn from some samples (USA and other WEIRD societies) while leaving samples from most of the human population largely underrepresented (Arnett, 2008; Henrich et al., 2010; Hruschka et al., 2018; Nielsen et al., 2017; Van de Vijver, 2013) and (2) the need for greater recognition and mindfulness about potential constraints to the generalizability of findings from one sample of subjects to humanity more broadly, especially for studies conducted on participants in the United States (Christopher et al., 2014; Rad et al., 2018; Rozin, 2001; Simons et al., 2017). Here, we examined whether one manifestation of these trends in psychology may be a bias or asymmetry in the assumed generalizability of study findings based on the racial/ethnic/cultural characteristics of a study’s sample. Examining the qualification of these characteristics in titles of published articles in psychology journals provided one potentially naturalistic test of this question.
Overall, we found support for all three of our hypotheses. We found that titles of articles from psychology specifying characteristics of samples from the United States were encountered much less frequently compared to samples from other WEIRD regions and non-WEIRD regions. Additionally, titles specifying samples from non-WEIRD regions were encountered the most frequently in psychology journals (Hypothesis 1). These observations are consistent with prior criticisms of psychology being not only western-centric but too exclusively representative of the United States (Arnett, 2008; Christopher et al., 2014; Henrich et al., 2010; Rad et al., 2018). One implication for this pattern could be the presence of an implicit or lay assumption that findings derived from the United States and other WEIRD populations may be more representative of or more readily generalizable to humankind. According to this perspective, conclusions about human psychological and behavioral processes emerging from studies on non-WEIRD populations may be unduly dismissed as being culturally constrained (Buchtel, 2014) or reflecting a “niche” pattern of thought and behavior. If this is the case, the bias in the tendency to qualify sample characteristics in titles may reflect and/or reinforce a subtle form of infrahumanization (Leyens et al., 2000), in which the nature, experiences, and behaviors of some populations are assumed to be a better reflection of humanity than others.
This potential bias in assumptions of varying degrees of different populations as adequately representing human nature may be one consequence of historical overrepresentation of Western researchers and perspectives in psychology (especially from the United States). Discrepancies in rates of qualifying origins of samples in titles based on their race, nationality, or ethnicity may have initially emerged due to reasons other than assumptions of generalizability, such as the disproportionate amount of editors, contributors, and readers of psychology journals assuming that most studies will involve Western samples by default. Eventually, these trends may have unintentionally promoted assumptions about generalizability of different racial/ethnic groups to human nature and become reinforced over time. Supporting this idea, we also observed the gap in relative rates of titles qualifying samples from the United States and other regions (WEIRD and non-WEIRD) to be increasing across a span of approximately 14 years (Hypothesis 3), with this difference mainly being exacerbated by growing frequencies for samples from non-WEIRD regions being qualified in article titles. This trend may suggest not only an increase of research recruiting non-WEIRD samples but also a perpetrating norm, such that contributors of research studying non-WEIRD samples (presumably because they are based in institutions in non-WEIRD societies) are inclined to qualify the origin or characteristics of their samples in titles, thereby constraining the generalizability of their findings. This perpetrating norm could also operate by reviewers and editors of leading journals of psychology (which are based in the United States or WEIRD societies) to increasingly expect submitted manuscripts involving non-WEIRD samples to communicate this information in article titles.
Importantly, we also observed that when racial/ethnic minorities within a given society are being qualified within a title, this most often occurs for minority groups (i.e., non-European Americans) in the United States (Hypothesis 2). Further supporting this point are observations from a supplementary code assigned to titles that specified White Americans as the target sample (Table 2). Only 32 titles specified a White American sample compared to 807 titles that specified non-White racial/ethnic/cultural minorities in the United States as samples. Given an implicit expectation that people in the United States are prototypically White (Devos & Banaji, 2005), having a sample in a study that is from the United States, but is systematically non-White (i.e., Asian Americans) may produce the expectation that the study’s findings cannot be generalized as if it were a typical study involving a sample largely comprised of White Americans. This may motivate the tendency to specify in the title that the study involves an ethnic minority sample. This finding suggests that the patterns observed in the present study may indeed be a manifestation of assumptions about some populations (namely White Americans) being more generalizable to humanity than others (e.g., Asian Americans). One may argue that this result is a by-product of the high levels of racial and ethnic diversity of the United States compared to other non-WEIRD societies that may be more racially or ethnically homogenous (e.g., Japan). Yet other societies we classified as WEIRD (i.e., Canada, Western Europe, Australia, etc.) and non-WEIRD (i.e., Latin America, Southeast Asia) are also distinguished by their high levels of racial/ethnic diversity and multiculturalism.
There are some notable limitations in the present study. The first is that our content analysis is only limited to the titles of psychology articles rather than abstracts or the main text. 3 There are a couple of reasons for this decision. First is that we sought to sample many titles from not only a comprehensive range of journals across subfields of psychology (49 journals) but also across the span of nearly 15 years. Due to the sheer quantity of articles and content this process would produce, our analyses only focused on titles. Second, given their function of concisely communicating the most important point or aim of an article, the inclusion of sample characteristics in a title may be an especially sensitive reflection of how much importance authors may have placed on setting expectations on the generalizability of their findings.
Another limitation is that we did not confirm the actual samples recruited in these published articles. There may be many articles that sampled from non-WEIRD populations but did not specify this in their titles. Such articles would have been overlooked in our title sampling and coding procedures. Yet it has already been well-established that target populations that are actually sampled in published papers in top psychology journals predominantly represent the United States or other WEIRD regions (Hruschka et al., 2018; Nielsen et al., 2017). For instance, for articles published in Psychological Science (one of the leading psychology journals) in 2014, only 3.9%, 1.6%, and 0.6% of reported samples were from Asia, Africa, and Latin America, respectively (Rad et al., 2018). Looking further back to 2003–2007, only 3%, less than 0.5%, and 1% of reported samples were from Asia, Africa, and Latin America, respectively, across a diverse range of journals published by the American Psychological Association (Arnett, 2008). These observations suggest that our current study may involve a relatively conservative test of our hypotheses and that the majority of titles qualifying sample origins should actually be from the United States (and other WEIRD regions) if sample origins were being qualified in titles at rates consistent with the samples being studied.
Future research could test for similar disparities in specification of sample origins in other fields (e.g., public health, biomedical sciences). Bias in qualification of sample characteristics in titles may be even more pronounced in fields where sampling bias across nations has been even more extreme. For instance, overrepresentation of samples from WEIRD populations may be even greater in the neurosciences given prohibitive costs required for neuroimaging research (Chiao & Cheon, 2010). Consequently, conclusions drawn from WEIRD samples may be asymmetrically assumed as more representative of humanity in such fields.
Disparities in the perceived generalizability of psychological research based on sample origins could be addressed through the following practices in publishing. The first is to standardize how constraints to generalizability are applied and reported. For instance, Simons and colleagues (2017) proposed that the discussion section of all empirical research articles should include a statement of the Constraints on Generality (a “COG” statement) that clearly describes the target populations—a policy that editors and reviewers could standardize to limit asymmetric demands for such statements from studies on non-WEIRD samples. In regard to stating sample characteristics in article titles, our findings (Figure 2) suggest widening disparities in the frequency of this type of constraint on generality, with studies from the United States showing no changes in recent years despite studies based on samples from all other regions (especially non-WEIRD) increasing in this practice.
Second is that editors and reviewers of psychology journals, particularly high-impact journals, should be mindful of recognizing that studies conducted on non-USA or non-WEIRD samples are no more “cultural” or limited in generalizability or applicability than studies on samples that are more commonly reported in their journals (e.g., from the United States). One potential rebuttal to this suggestion from researchers in the United States is that the United States is much larger, vaster, and more diverse than many other countries. For instance, according to this argument, a study based on samples from much smaller countries such as Singapore or Portugal should shoulder greater responsibility to constrain the generality of their samples compared to a study conducted on a sample from the United States. But psychological studies from the United States rarely sample from a representative population and instead focus on a very narrow subset of the population (e.g., a convenience sample drawn from a single university or city in Midwestern United States). Therefore, such assumptions and rebuttals are dubious.
Ultimately, guidelines to promote greater inclusion in psychological research must also be recognized as issues impacting credibility and replicability rather than diversity alone. The broader discussion in psychological science about the WEIRDness of research findings (Henrich et al., 2010) surfaced around a similar period as mainstream concerns about the “replicability crisis” in psychology (Pashler & Wagenmakers, 2012). Yet, while the replicability crisis has prompted widespread changes to expectations and practices about how research should be conducted (e.g., open science approaches, preregistration), concerns about the WEIRDness of psychological science have prompted limited actions. This may be because there is much greater urgency to alleviate obstacles to the validity and credibility of one’s research compared to conducting research that is more inclusive and generalizable. Ensuring research is valid, credible, and replicable may be experienced as a nonnegotiable “ought” that must be vigilantly defended among researchers, whereas promoting research that is inclusive and sensitive to diversity may instead be experienced as an “ideal” (see Higgins et al., 1994). Thus, motivating deeper changes to promote inclusion and diversity in psychology may require reframing of the WEIRD (particularly USA-centric) nature of psychology as primarily a challenge to credibility and validity of research done largely on WEIRD (especially the United States) samples rather than portraying it solely as an issue of inclusion, diversity, or equity. Given that perceived generalizability of findings also contributes to perceived impact of empirical research (Simons et al., 2017), without wider and more consistent application of guidelines for communicating generalizability, research that seeks to contribute to the need for widening understanding of the diversity of human nature may ironically be devalued instead.
Footnotes
Authors’ Note
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of the A*STAR.
Acknowledgments
We thank Johan Ng and Sheeling Neo for assistance in coding and organizing data. This study was supported by Nanyang Technological University HSS-Startup Grant (M4081569.SS0) and Nanyang Assistant Professorship Award Grant (M4081643.SS0) awarded to Bobby K. Cheon and by A*STAR under its IAF-PP Food Structure Engineering for Nutrition and Health Programme (Grant ID No: H17/01/a0/A11 and H18/01/a0/B11).
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research and preparation of this manuscript was supported by Nanyang Technological University HSS-Startup Grant (M4081569.SS0) and Nanyang Assistant Professorship Award Grant (M4081643.SS0) awarded to Bobby K. Cheon, and by A*STAR under its IAF-PP Food Structure Engineering for Nutrition and Health Programme (Grant ID No: H17/01/a0/A11 and H18/01/a0/B11).
