A call for considering ethnic minority status in adolescent research

Recent initiatives urge researchers to consider the ethnicity of adolescent research and to evaluate how well the field responds to ethnic minority youth's needs. For example, the Society for Research on Adolescence (SRA) recently released its important report on the place of cultural and ethnic diversity in the study of adolescence (Larson, Cauce, and Umaña-Taylor, 2004). The analysis featured a survey of its members and a close examination of the society's organization and activities. The report found encouraging efforts to address issues of diversity and considerable success in achieving inclusiveness. The survey, for example, revealed a general sense that the association fairly treated ethnic/minority researchers and that over 3/4 of survey participants found the organization either “satisfactory” or “very satisfactory” on issues of ethnic and cultural diversity. An informal analysis of the association's journal, the Journal of Research on Adolescence, also revealed impressive results. That analysis indicated that 53% of the studies conducted in North America had samples constituted of a European American majority (50% or greater); 19% of the studies had a mix of ethnic groups, 19% had an African American majority, and the remaining 8% had an Hispanic majority. Although the report's authors highlighted well the limitations of their membership and journal surveys and cautioned against over-interpreting their results, their findings were nonetheless quite informative. Efforts like these, including the commissioning of the report itself, reveal an energizing interest in taking ethnicity seriously.

A close look at the SRA report reveals, however, mixed messages. The membership survey, for example, uncovered a wide variety of views regarding the extent to which adolescent research appropriately addresses ethnic issues. Striking differences emerged along minority/ethnic and European American group lines. Indeed, a look at the findings reveals disagreement for the vast majority of questions asked about the association and the journal's coverage of ethnicity. Arguably the most intriguing set of disagreements deals with the findings related to the journal's content. For example, minority members (n = 97) reported considerably more disagreement than European American members (n = 268) in their evaluation of whether they agreed with the view that SRA's journal “provides coverage on issues of cultural and ethnic diversity” (59.8% vs. 20.9% disagreed). As the authors of the report reveal and looking at the actual number of articles published in the journal, the journal's publishing around half of its articles on samples constituted of multiethnic or ethnic groups of adolescents does appear to support the conclusion that the journal has done a relatively “good job” in the way it has been inclusive of ethnicity (Id., p. 3). Yet, it is unclear whether minority members of SRA would agree with that characterization. The journal study's evidence may well influence the perceptions of its members and sway their opinion toward a more favorable view of the manner the journal addresses minority youth. But, before seeing the study results and simply relying on their perceptions, European-American researchers agreed much more than minority researchers that the journal did cover issues of cultural and ethnic diversity. Findings like these raise important concerns about the extent to which the study of adolescence does indeed embrace research that could provide useful information about minority adolescents.

Existing research does not provide enough evidence to reconcile the disparity in members’ perceptions and the actual findings suggesting that the journal does seem to devote considerable space to studies that include ethnic minorities. The membership survey may have been limited. The response rate for the membership survey was 34%, which may skew the findings. Yet, many studies do publish useful results with lower response rates. And, the reality remains that a significant proportion of those who care enough to respond to a membership survey, and who receive the journal as part of their membership to the organization, find ethnicity insufficiently addressed. The journal analysis also had its limitations. The study's informality may well mean that its findings were unrepresentative or that the study failed to index relevant concerns. Unfortunately, it is the only study of its kind. Existing analyses focusing on adolescent research, then, raise important issues but do not provide enough information to resolve them.

A look at studies that have examined how related fields have addressed ethnicity actually would support ethnic minority researchers’ impressions. Published studies that examine leading psychology journals, for example, reveal that those journals devote much less space to issues of ethnicity than SRA's: those that have reported proportions reveal that the percentage of articles with an ethnic-minority focus are, on average, in the low, single digits. For example, Iwamasa and Smith (1996) found that, of articles published in three behavioral psychology journals during a range of 16–23 years, only 1.31% had content specific to people of color; Iwama, Sorocco, and Koonce (2002) summarized a comprehensive content analysis of five of the leading clinical psychology journals over a 17-year period and found that 2% of that literature focused on ethnic minority groups; Imada and Schiavo (2005) reported 4.7 representation of ethnic minority focused research in 6 APA journals; Flores et al. (2006) found that, out of 4,181 articles in professional career development journals, only 281 or 6.7% related to race/ethnic minorities; Nilsson, et al. (2003) found that 3.0% of a decade of the leading professional psychological practice journal focused on race/ethnicity; and Behl and her colleagues (Behl, Crouch, May, Valente, and Conyngham, 2001) found that 6.7% of 1,133 articles dealing with child maltreatment focused on ethnicity. Studies focusing on the most studied ethnic group, African Americans, report similarly low percentages: Graham (1992) found that only 3.6% of the articles in 6 journals published by the American Psychological Association had content specific to African Americans. Perhaps more strikingly, a comprehensive study of all psychology-related journals (published from 1993 to 1999) concluded as follows: “Publications involving cross-cultural issues represented 1% and ethnic minority issues represented 3% of all the publications in the English language in the PsychINFO database” (Hall and Marmaba, 2001, p.15). Although these estimates are low, they probably are overestimates given the large amount of missing data; e.g., Case and Smith (2000) analyzed 14 counseling journals and found it encouraging that 7.2% of the studies had content specific to an ethnic group, but the authors also noted that the percentage actually could be an overestimate given that nearly 40% of the reviewed studies had to be excluded because they had failed to report the ethnicity of participants. Available evidence, then, suggests that specialized publishing outlets for the study of adolescence either are performing much better than other fields or that the informal SRA study missed the mark in the way it approached ethnicity. It also may well be that studies from other fields simply are not good comparisons, as they, for example, may not have captured recent trends or may have taken a very different approach to operationalizing the use of ethnicity. We also, of course, do not know the extent to which other journals dedicated to the study of adolescence publish research that considers ethnicity. As a result and again, it seems fair to conclude that existing research leaves open whether research focusing on adolescence does indeed pay enough attention to ethnic diversity. If research from other fields concerned with psychological issues are any guide, it would appear that journals dedicated to the study of adolescence do not publish more than a handful of articles with ethnic participants.

Debate and uncertainty may surround discussions of whether adolescent research appropriately considers the diversity of adolescents’ ethnic experiences. But, little evidence suggests that the study of adolescence should be immune from calls to take ethnicity more seriously. Research findings, ethical concerns, and practical considerations actually reveal a pressing need to move toward exploring the place of ethnic diversity in adolescent development. For example, the U.S. adolescent population is racially/ethnically diverse, even moreso than the general population. The latest reports from the U.S. Census Bureau (2000) reveal that 36% of adolescents ages 10–19 are minority youth, compared to 28% of the general population. This diversity, of course, pales in comparison to the diversity of adolescence we find when we take a more global perspective (Brown, Larson, and Sarawathi, 2002). The need to consider ethnicity also reflects the principles of good science. Failing to include ethnic minorities in research or making sweeping references to ethnic groups can contribute to misrepresentations, violate central tenets concerning external validity, hamper the ability to generalize findings across subgroups within ethnic categories, and prevent accurate and efficient efforts to replicate research results. Of course, practitioners also necessarily rely on research. They need sufficiently specific research to enable them to make decisions about the applicability and generalizability of research findings to their clients and avoid potential harms attached to erroneous conclusions about commonalities and differences among people. Our understanding of adolescent development further confirms the significance of ethnicity. Many aspects of adolescent development center on identity development, which emphasizes the need to consider how ethnicity can shape, for example, peer influences (Bámaca and Umaña-Taylor, 2006), perceptions of one's body (Newman, Sontag, and Salvato, 2006), academic achievement (Eamon, 2005), educational aspirations (Smith, Schneider, and Ruck, 2005), reproductive health (Milan, Ethier, Lewis, Kershaw, Niccolai, and Ickovics, 2006), and societal stereotypes (Woods, Kurtz-Costes, and Rowley, 2005). Neither last nor least, considering ethnic diversity is the right thing to do. Understanding society's response to difference, and our place in that response, is necessary to take seriously the place of adolescents in society. This means addressing, for example, the harms of perceived exclusion and discrimination (Faircloth and Hamm, 2005), of being perceived as more troubled by those who are ill-equipped to respond effectively to minority youth's needs (Anthony, Anthony, Morel, and Acosta, 2005) and of being placed at increased risk to engage in problem behavior (Darling, 2005). It also means responding to the harms caused by failing to consider normal development and positive youth development (Levesque, 2006, 2007). Given that researchers who engage in the critical task of generating the information that can mold public opinion and serve as the basis for a wide variety of policies and services (such as those involving mental health, education, social welfare, criminal justice, and medicine), they undertake the responsibility to ensure that the soundness of their scientific research rests on its cultural validity.

The expectation that the study of adolescence should take ethnicity more seriously is not new and organizations other than SRA have launched important calls to action. These calls for greater inclusiveness reflect a more general, contemporary emphasis on multiculturalism. For example and even several years ago, revisions of “codes of ethics” of the American Psychological Association (1992), the National Association of School Psychologists (1997), and the American Counseling Association (1995) addressed multicultural issues for researchers and practitioners. The efforts find it unethical to not integrate multicultural and culture-specific awareness, knowledge, and skills into interactions with clients and participants (see, American Psychological Association, 2003). A recent report of the U.S. Surgeon General documented well the existence of several disparities affecting mental health care of racial and ethnic minorities compared with whites: Minorities have less access to, and availability of, mental health services; minorities are less likely to receive needed mental health services; and minorities in treatment often receive a poorer quality of mental health care (U.S. Department of Health and Human Services, 2001). Concerns about the failure to address minorities’ needs also have led to changes in the policies of research-funding institutions. For example, the National Institutes of Health (NIH) mandates that women and minority group members be included in all NIH-supported research (NIH, 2001). The mandate does not accept the cost of including diverse populations or the investigator's geographic area as an appropriate rationale for inadequate representation. At least anecdotal evidence and cursory looks at journals publishing articles relating to child development reveal that these changes and calls for reform seem to have an impact as the developmental sciences attempt to address the underrepresentation of minorities in research and produce more generalizable research (McLoyd, 2003).

The initiatives and calls to action, then, make good sense and some changes do seem to be taking place. It still may be a matter of opinion whether organizations and journals that serve as gatekeepers for adolescent research heed the suggestions and consider ethnicity seriously enough. As a scientific enterprise, however, the study of adolescence could benefit from a systematic review of the extent to which journals, which remain the key outlet for disseminating rigorous empirical research, do consider ethnicity and do present findings in an ethnically informed manner. This study provides empirical data to begin a discussion to address this gap by examining our field's attention to ethnicity. To that end, note that, for our purposes, we use the same ethnic descriptors used by researchers and view ethnic status as denoting a relationship to a socially distinguished group set apart, by itself or others, primarily on the basis of cultural or nationality characteristics. This approach, it is hoped, will help reveal how and the extent to which researchers make a place for ethnicity in adolescent research.

Concerns highlighted by efforts that evaluate the ethnicity of research

Complex and multifaceted constructs like ethnicity, race, culture and minority/majority often raise controversies and can evoke strong reactions. Not surprisingly, dilemmas emerge. Some, for example, question what is meant by these terms and the basis for classifications (Hall, 1997; Helms and Talleyrand, 1997). These challenges question the basis, utility, and consequences of using ethnic and other categorical systems. On the other hand, the scientific demand of adequate participant specification requires researchers to establish the limits of generalizability and provide enough information to aid replication. In the (still recent) past, researchers avoided this dilemma by simply not addressing or reporting ethnicity or race. It is difficult to deny the focus on White subjects, as Guthrie's (1976) groundbreaking work noted in its title, “Even the rat was White.” Attempts to avoid issues were such a common practice that they laid the foundation for calls to action (American Psychological Association, 2003). Now that professional organizations increasingly urge (and some require) authors to report their sample's primary characteristics (age, sex, race/ethnicity) and to address issues of diversity more forthrightly, it remains to be seen how researchers address this dilemma. Do they report their participants’ ethnic characteristics? Do they only pay token attention to requests to report? Do they report but not provide usable data along ethnic lines? Do they compare groups or focus on within group factors? Do they reify classifications and ignore diversity within groups? Do they avoid dilemmas by using only participants’ self-reported race/ethnicity as a basis to then infer an understanding of the constructs? Do they problematize minority groups? Do they highlight ethnicity at the expense of considering other salient sample characteristics (like socioeconomic status and how that status may be experienced differently across and within groups)? Thinking of researchers’ approaches to ethnicity in adolescent development certainly raises numerous questions. Given the multiplicity of issues that could emerge, we begin by highlighting conclusions and concerns raised by researchers who have analyzed systematically the reporting and use of research participants’ racial, ethnic and minority status.

Only one study relates directly to this study's goals. The informal, unpublished SRA study by Larson and his colleagues (Larson, Cauce, and Umaña-Taylor, 2004) provides the sole example of research directly focusing on the place of ethnicity in adolescent research. That study focused on one journal's publications from 1999 to the middle of 2003. The study, which yielded 66 empirical articles dealing with North American youth, mainly analyzed the ethnic characteristics of the articles’ participants, whether the studies made ethnic comparisons, and whether the studies focused on problem behavior. We already have reported some of the study's important findings above. In addition to those findings, it is important to note that the study found that research did not unduly focus on problem behavior; 37% of the 27 articles in which a majority of the sample was youth of color had a major focus on problem behavior. It also found that researchers (89%) did tend to report the ethnic composition of their samples. Only 9% reported comparisons among ethnic groups and 16% analyzed findings in ways deemed ethnically homogenous. These are all quite suggestive findings, but even the authors themselves are quick to note the tentativeness of their findings.

A thorough review of the literature published since 1985 located only one peer-reviewed study of the status of ethnic research that included adolescents as participants. Although that study grouped adolescents with children and focused on one ethnic/racial group, it nevertheless was groundbreaking and influential. That study centered on Child Development articles that included Afro-American children; the focus on African-American children was intentional in that the study sought to determine how studies explored race (McLoyd and Randolph, 1985). The study examined whether researchers compared Afro-American children to other groups, grouped them together without comparison, or focused exclusively on Afro-American children. The study also examined other characteristics of the sample, the research topics and methods used, and whether the study commented on issues of validity and discussed implications relating to Afro-American children. Among the key findings was the conclusion that researchers are twice as likely to compare Afro-American children to other groups rather than examine factors within Afro- American groups. The study also found fluctuation in the proportion of studies that explicitly viewed Afro-American children as deficient. Importantly, McLoyd (2003) offered an updated (but unpublished) report of her impressions of advances in addressing ethnic/racial issues and other pressing needs in the study of child development. That study focused on the Society for Research on Child Development, which has a very long history and a popular journal with literally thousands of articles. Beyond marking important progress in organizational changes, her more recent study is of significance for at least two reasons: (1) it reported that more articles now focus on ethnic/racial groups without comparing them to other groups and (2) it laid the foundation for the SRA journal survey. Given that the current study owes much to the SRA study, it necessarily also owes much to McLoyd's analyses.

The paucity of empirical analyses focusing on how research on adolescence considers ethnic and cultural issues led us to seek guidance from other areas of study. Other fields of study considerably vary in the extent to which they have become self-reflective about their own scholarly discourse. Several important commentaries have addressed the place of ethnicity in empirical research, particularly in the psychological sciences. All highlight the pervasive failure to address ethnicity adequately (Sue, 1999, 2004). Although persuasive, those commentaries tend not to be buttressed by systematic analyses of empirical research. Over the past decade or so, however, researchers have subjected several groups of journals to content analyses, such as journals relating to applied psychology (Case and Smith, 2000), clinical psychology (Iwamasa, Sorocco, and Koonce, 2002), counseling psychology (Ponterotto, 1988; Buboltz, Miller, and Williams, 1999), social psychology (Hunt, Jackson, Powell, and Steelman, 2000), social work (McMahon and Allen-Meares, 1992), family violence (Behl, Crouch, May, Valente, and Conyngham, 2001), addiction (Trimble, 1991), nursing (Drevdhal, Philips, and Taylor, 2006), and vocational psychology (Flores et al., 2006). There also have been a few analyses that have taken a more general approach to journals (e.g., Hall and Marmaba, 2001). Those studies vary considerably in their breadth, depth, and quality. They do, however, reveal concern for addressing ethnicity and offer important points for us to consider.

The first fundamental concern that emerges from prior studies relates to the extent to which published research actually reports ethnic data when it reports other demographic variables. Journals that follow the American Psychological Association's Publication Manual (2001) should follow the suggestion that researchers “report major demographic characteristics such as sex, age, and race/ethnicity, and, where possible and appropriate, characteristics such as socioeconomic status, disability status, and sexual orientation” (p. 81). Content analyses of journals have revealed an increasing trend in the reporting of ethnic and racial characteristics of participants (Imada and Schiavo, 2005). Several factors (see above) may account for the dramatic rise in reporting, but wide variations still exist in the reporting of samples’ ethnic characteristics. Vague reports may comply with requests and reveal an awareness of the need to report ethnic characteristics of samples, but they may not be helpful. Specifics are needed to the extent that we need to know (or at least be able to determine) the relative percentages of the ethnic groups under study and, hopefully, the actual size of the group in order to evaluate the robustness of findings. We cannot generalize findings without specifics; nor can we replicate studies if we do not know their relevant sampling criteria. Thus, a field that would take ethnicity seriously would foster the development of studies that report specific findings that enable readers to make informed decisions about their relevance and utility.

The second issue that arises deals with classifications and analyses relating to them. Research often seems to use quite different terms interchangeably, and this usage appears to have become self-fulfilling. Researchers frequently categorize participants into broad pan-ethnic labels, such as American Indian, Asian, Black, Hispanic, or White. This lumping has been highly criticized on the grounds that it disregards the historical coupling of groups, perepetuates the belief that minority groups are all the same, and ignores how these categories can foster beliefs about the inherent superiority of a particular group (Oboler, 1995; Phinney, 1996). From the perspective of research on adolescence, broad ethnic labels run the risk of producing overgeneralizations about the nature of development among ethnic minority youths and neglecting the unique differences among individuals within various racial, ethnic, and cultural groups. This way of categorizing participants also can mask the developmental influence of racial, ethnic, mixed-race, or bicultural self-identification. These complex issues result in one general point for us to consider. Researchers reporting analyses relating to adolescence are likely to lump groups together (just like other researchers generally do), but these concerns mean that we must consider how researchers who do report the ethnic status of their participants actually analyze the data they have about their participants. This means, for example, that we may wish to consider whether studies focus across groups or within groups. Although either approach may be appropriate depending on the particular study's research questions, properly understanding ethnicity would mean that studies do examine within group differences and that across group differences are done with due care. This concern moves us away from static issues of inclusion or adequate reporting to dynamic issues regarding the implications of considering ethnicity. These efforts consider, for example, the moderating effects of national origin, immigration history, religion, socioeconomic status, neighborhoods, and traditions on normative and maladaptive development. At its core, then, existing research reveals the need to focus beyond mere description to actual analysis.

Lastly, prior research reveals the influence of challenges that will need considerable commitment if we are to address them more satisfactorily. The methods we use, for example, influence our research and eventual findings. As one can tell from the potential grouping just enumerated (e.g., national origin, immigration, religion), it is difficult (though not impossible) to move beyond groups when conducting the type of research we do. The field tends to embrace quantitative research, which brings its own baggage. Most notably, the use of quantitative methods has led to a need to use large samples, which fosters the need to collapse groups together to facilitate the use of significance testing. In addition, research pervasively relies on respondents’ self-reports. Those reports often must be given in forced-choice format and, when coupled with the need for at least modest sample sizes, we may inappropriately ignore, for example, multi-ethnic/multi-racial participants or the reality that ethnic identity can change over time. These all are important challenges, and our ability to address them will derive from the ability to lay a proper foundation resting on our consideration of the other issues we have highlighted above.

This study

This analysis provides a snapshot of the extent to which researchers who study adolescence consider ethnic/minority concerns in the research they eventually publish. We examine how researchers report, operationalize, and use variables denoting ethnic diversity. To do so, this study examines the recent publications of the six leading journals dedicated to the general study of adolescence. This is not an easy task. Given the tendency of one's ethnicity to influence one's perceptions of ethnicity, one certainly must think twice before embarking on such an effort. In addition, existing reviews of published articles reveal the complexity of an attempt to analyze systematically how researchers address ethnic/minority issues. Researchers often are surprisingly imprecise in their descriptions of participants as they often lump different groups together or even do not describe the ethnic characteristics of their samples. When researchers are more precise, that precision often fails to inform their actual analyses. And, when analyses are done, it remains unclear how best to categorize them in terms of the manner they approach ethnicity. As a result, efforts to analyze how researchers address ethnicity necessarily proceed full of limitations that only can be addressed by systematically noting those limitations and carefully revealing how researchers conducted their studies.

This study aimed to address important issues that have emerged from previous studies in ways that would result in the creation of variables that would lend themselves to the least discretion in interpretations and to the highest reliability. As a result, two goals emerged. The first goal sought to determine the extent to which adolescent research does consider ethnicity. This goal seeks to provide a starting point for discussion. We actually do not know the extent to which researchers who study adolescence consider ethnicity in their publications. The second goal was to determine whether and how researchers analyze ethnic data. This analysis was undertaken to determine whether ethnic data was merely reported or actually analyzed to offer findings relevant to ethnic development. The results of these two goals were deemed the necessary starting points for considering whether and how researchers are laying a foundation of research that includes findings relevant to ethnic adolescents.

It is important to emphasize what this study does not do. No results are presented relating to the manner articles focus on problem rather than on normative behaviors. Although this focus clearly was of interest and originally was part of the study, we do not provide results from our effort to address this issue. The reason is that many manuscripts focus on numerous topics and approach topics in very different ways, which makes it difficult to determine the focus of the manuscript along the lines of normative or problem behavior. (For example, some articles examine normative behavior to understand problem behavior; and many problem behaviors are somewhat normative.) We decided not to address this issue because the author found that the variables needed to address them reasonably were too unreliable and that available data relevant to this concern could not be developed and presented with appropriate confidence. This is yet another clear example, as we have noted above, of how the methodological orientation we adopt necessarily limits the analyses we can make.

Method

Sample

This study selected the six, well-established journals with missions to publish articles dealing with the adolescent period: Journal of Adolescence, Journal of Adolescent Research, Journal of Early Adolescence, Journal of Research on Adolescence, Journal of Youth and Adolescence, and Youth and Society. Many other journals do publish research relating to adolescence, but these were selected for their focus on adolescence, their broad missions within that period (e.g., not only health or delinquency), and their status as highly regarded research outlets (see Levesque, 2006). These journals also were selected on the rationale that they would provide us with a good sense of quality research relating to adolescence. Given interest in determining current research's focus on ethnicity, the project analyzed articles published from the year 2000 up to and including 2006. Within the selected journals, the author reviewed each empirical article with or about human participants. The study excluded theoretical articles, program descriptions, book reviews, rebuttals, editorial statements, literature reviews, meta-analyses, speeches, tributes, and brief reports of 2 pages or less. This process identified 1283 empirical articles that serve as the basis for our analyses.

Ethnicity

As with any study using secondary data, we were limited in how we could define ethnicity. The most parsimonious way to approach ethnicity, as noted earlier, was to adopt the categories most often used by researchers. As a result we classified each article in terms of the participants’ ethnic characteristics. See the Appendix for the coding sheet. This approach resulted in four major variables that indexed whether, and if so, how each article presented its participants’ ethnic characteristics.

Sample ethnicity

We recorded the ethnic composition of the research sample reported by authors and grouped them together into the categories most often used by researchers. African American was the most commonly used designation for Black youth, and no studies provided more ethnic descriptors beyond this broad category. Asian Americans also lumped together adolescents from many Asian groups, and research tends not to distinguish among them. European American was the most common designation for White adolescents, and only a few studies identified the ethnic makeup of this broad category. Native Americans also generally were lumped together into one category, with research ignoring variation among different groups. Lastly, Hispanic Americans were somewhat more likely to be defined more precisely (e.g., Mexican American, Puerto Rican, Columbian, etc), but few studies focused on them and the vast majority of studies did lump them into an Hispanic or Latino group; as a result, this study grouped them together as an Hispanic category. For each of these categories and following the SRA study, we coded whether they were the majority of participants (50 to79.9% of the sample) or were the predominant group (over 80%). We coded the ethnic designation as Mixed when the article had no ethnic group constituting a majority.

Specified ethnicity

We noted whether or not articles specifically identified participants’ ethnicity; if they did not, they were coded as unspecified ethnicity. To be coded as a study that specifically identified the ethnic characteristics of participants, the study needed to provide enough information to allow us to determine the actual percentage of at least one group. The focus on at least one group was deemed practical since many articles lump together remaining groups into a broad “other” category. The rationale for the needed specificity was that the study would need to be replicable with a similar group.

Vaguely ethnic

We realized that our “ethnic specificity” variable was quite stringent. To ensure that we did not inflate the number of articles that did not report ethnicity, we also coded for articles that were not specific but still reported, even vaguely, the ethnic composition of their samples. In this category, we included descriptions that could be interpreted differently, such as “mostly white” and “predominantly white” but that still gave us a general sense of participants’ ethnic group. Thus, for some analyses, we will distinguish between the broad category of unspecified ethnicity and vaguely ethnic.

International sample

To qualify as an international sample, the study needed to have been conducted outside the United States, a designation which also included articles that statistically compared adolescents in the U.S. with those from another country. The ethnicity of participants in this category was not analyzed since non-U.S. studies typically do not report the ethnic characteristics of their participants.

Ethnic-focus

The ethnic focus variable coded for six different ways articles can address ethnicity. Articles were coded as ethnic if 50% or more of the sample was comprised of one ethnic group and no comparisons were made to another ethnic group. Articles were multiethnic if the ethnicity of their participants was mentioned and if explicit comparisons were made among ethnic groups (including European American). If ethnicity was mentioned but only as characteristics of the sample (with no further analysis), if no ethnic group constituted a majority percentage of the sample, the focus was deemed as token. Articles fit in the nonethnic category if they provided no ethnic data analyses and explicitly noted that at least the majority of the group was European American. Articles were deemed international if the study was conducted partly outside of the United States. Lastly, articles were coded as unspecified if there were no specific reports of participants’ ethnicity (thus, “majority white” was deemed too vague and thus included here; specific percentages were needed for at least one category). See the Appendix for the coding sheet.

Table 1 Number and percentage of American (U.S.) articles by their reports of ethnicity
Full size table
Table 2 Number and percentage of articles by their samples’ ethnicity
Full size table

Reliability

Several efforts were made to secure appropriate reliability. Three are worth noting. First and as we already have noted, we focused on using constructs that would limit coders’ discretion. Second, all articles were coded twice: One round of coding was conducted to determine the ethnicity of samples and, once that was done for all articles, another round of coding was used to index the ethnic focus of the sample. The results of the two rounds of coding were then compared to identify and resolve (very few) inconsistencies. The accuracy of data entry and tabulations was double checked by two trained coders; three typographical errors were identified and corrected. The third effort to ensure reliability involved an independent analysis conducted by an untrained coder. The coder only was instructed to use the coding scheme (provided in the Appendix) to the best of her abilities, an instruction aimed to determine whether readers could understand the categories and reliably use them to index articles. That analysis included four issues of each of the six journals used in this study and resulted in a hundred percent agreement. Given the success in securing high inter-rater reliability, no other efforts were made to obtain results from other coders.

Results

Table 1 contains the number and percentages of articles that report the ethnic characteristics of their participants. Note that we have excluded international articles and focused on those conducted entirely in the United States. Using the figures provided, we can determine that, overall, 93% of articles published since 2000 and in the six leading journals dedicated to the study of adolescence do report the ethnicity of their participants. Of that high percentage, 87% provide specific characteristics of their sample's ethnicity. When examining each journal's percentage of articles that are specific relative to their total number of articles with exclusively U.S. participants, we find modest variation among journals: 83, 89, 92, 91, 84, and 84 percent. We do find somewhat more variation when we consider the percentage of articles, relative to the total number of articles, that do not report ethnicity at all: 15, 5, 2, 4, 5, and 9 percent. Overall, these findings suggest that no journal skews our findings and that journals do tend to report specific ethnic characteristics of their participants.

Table 3 Number and percentage of articles by their ethnic and international focus
Full size table

Table 2 presents a summary of the number of articles across journals by their mentioning the ethnicity of participants and the location of the study (solely U.S. vs. international). A few important points are especially worth highlighting. The number of studies using international samples was quite remarkable. Over 40 percent of the studies published in these journals have samples from countries other than the United States; all but one journal (JEA) had over a quarter of its articles devoted to studies of adolescents in other countries. That overall figure is higher than the average of all journal articles that specifically describe another group in the United States. Articles with at least a majority of European American adolescents constitute an overall average of 28% of articles published in these journals, with only one journal (JEA) publishing slightly more than 40% of manuscripts with a majority of European American participants. Seventeen percent of articles published had more minority participants than European American participants; all journals had a similar percentage except JA which had only 7% of its articles sampling minority, U.S. youth. Over a quarter (26%) of articles do not have a European American majority of participants. When we add this figure to the number of articles with subjects outside the U.S., we find that the majority of articles (68%) do not have a European American majority. Even if we were to assume that the articles that we marked as unspecified all made use of European American adolescents, we still are left with the result that the majority of articles, 60%, do not have a European American majority. We also find that, although journals may publish a high percentage of articles that do not have a European American majority, it does not mean that articles equally address all groups of ethnic adolescents. Articles tend not to have majorities from several American ethnic groups, especially those often subsumed under the broad categories of Asian, Hispanic, and Native American. These groups may be represented in “mixed” studies and lumped in “other” categories in the studies that have other majorities. As expected, the lumping does not allow us to draw usable conclusions regarding specific ethnic groups.

Table 3 presents a summary of the number and proportion of articles across journals by their focus on ethnic groups. We find that articles are slightly more likely to have an ethnic and multiethnic focus than a nonethnic focus. Together, multiethnic and ethnic manuscripts constitute one quarter of all manuscripts while nonethnic manuscripts constitute 21% of all manuscripts. Our analysis permits us to determine the extent to which articles merely described the ethnic composition of their study without analyzing results along group lines (the token group, 4%) and the extent to which studies failed to provide enough information to allow us to determine the number of at least one group of participants (the unspecified group, 8%). Given that we cannot say with any precision whether the token or unspecified groups tell us much about ethnicity, we may wish to group them with the non-ethnic group. This would make sense in that it would not over-estimate the percentage of articles that have an ethnic focus, would underestimate those that do not, and would give us a sense of the number of manuscripts that do focus on an ethnic group. If we add articles that merely are “token” or unspecified to the nonethnic group, we find that 33% of manuscripts do not have a specifically ethnic, multiethnic or international focus. The conclusion to be drawn from this analysis is that a conservative estimate reveals that the majority of articles do have an ethnic or international focus.

Table 4 Number and percentage of American (U.S.) articles by their ethnic focus
Full size table

Given the high amount of international research, it could be reasonable to wonder whether the inclusion of international articles skews our results. To address this possibility, Table 4 removes articles that were conducted outside the United States and offers a re-analysis of the extent to which articles present research focused on an ethnic group. If we remove studies that have an international focus, we find that, overall, at least 44% of articles report analyses that relate to ethnicity. We also find that nearly 20% of articles focused specifically on an ethnic group without comparing that group to another. We also are left with the impression that a notable percentage of articles (14%) do not specify their participants’ characteristics enough, despite increasing calls to describe the ethnic and cultural composition of samples. The failure to report specific characteristics does not mean that researchers did not have an ethnic focus. Studies that did not report the specific ethnic make-up of their samples actually often conducted ethnic analyses; it was just unclear how many participants in their group were from ethnic groups. This most likely led to an underestimate of the percentage of articles with an ethnic focus. We are left with the finding that over half (64%) of articles do have an ethnic focus.

Discussion

Organizations and researchers interested in adolescent research as a field of study recently have turned their attention to reflect critically on the manner the field addresses ethnicity. This reflection has led organizations to consider changes that would allow them to better address minority members’ needs. It also has resulted in evaluations of publications themselves. The focus on research has led, for example, to efforts investigating whether articles report the number of minority participants, the number of studies that focus solely on ethnic groups, and the research designs used when studies do examine data collected from minority adolescents. For the study of adolescence, these evaluations obviously focus on very basic issues. Although basic, even these simple issues surprisingly have not been addressed in a systematic fashion that considers the most common outlets for research on adolescence. This project emerged from these efforts; it sought to determine the extent to which the study of adolescence does commit itself to understanding ethnic minority adolescents.

Although other fields of study report a striking paucity of research focusing on ethnic diversity (Hall and Maramba, 2001), the same cannot be said of research on adolescence. Outlets for disseminating research on adolescence appear more inclusive than others. To be sure, our study does not address the quality and influence of specific articles. We do know, however, that no journal is overly contributing research that ignores issues of ethnicity. In addition, our findings suggest that all journals contribute a high volume of research relating to ethnicity. In this regard, this investigation adds much to Larson et al.'s (2004) informal study that had focused on one journal and found that 53% of studies have samples constituted of a European American majority. We found a considerably lower rate. Depending on one's methodological preferences, between 32 to 40% of 1283 empirical articles had samples constituted of a European American majority. We also found that, after excluding international articles, over 40% of our articles have an ethnic or multi-ethnic focus.

We found one unexpected result relating directly to diversity but which has not been the subject of much commentary. It is not clear how other fields report results from international studies, but it is clear that research on adolescence appears marked by an international focus. This international focus reveals that much of the research base includes even more international analyses than ethnic or nonethnic ones. This is an encouraging development for the study of adolescence. This development may not mean that European American adolescents’ experiences do not serve as the standard on which to judge other adolescents’ experiences. But, it does mean that researchers provide a knowledge base that does not ignore cultural and ethnic influences on development. Theories and research consider cultural and ethnic forces, albeit sometimes too implicitly. This development is of significance. Studies with participants other than European American adolescents can help determine and perhaps even challenge the relevance of European American theories that are assumed to dominate research. Such consideration of ethnic and cultural issues can increase the relevance and comprehensiveness of research and theories, not to mention lead to a more realistic understanding of adolescence.

Our findings may reveal cause for optimism in the field's recognition of ethnicity's significance to adolescent development, but that optimism is somewhat tempered if we take a broader look at our findings. Researchers pervasively do take into account cultural and ethnic forces in their analyses and the foundation of adolescent research does include considerable diversity. Yet, the actual place and use of ethnicity appears limited. Four examples illustrate the limits of the overall status of research addressing ethnicity.

The first example of the limited recognition of ethnic diversity involves the tendency to lump groups together when their number of participants is low, often for convenience and for not removing participants from studies. Perhaps as problematic is the removal of participants when their numbers are insufficient for statistical analyses. The demands of empiricism seem to be leading to a new kind of invisibility. Some groups of adolescents are studied, but they eventually are dropped from analyses or lumped with groups with their only similarity being that they are small in number. These issues become even more challenging when investigations seek to understand multiple identities (gender, sexual, ethnic, age, etc), especially those linked to multiple forms of minority status. This is not to say that more contextualized studies are impossible; recent studies, for example, provide insightful analyses of intersections among ethnicity and other key categories, such as gender (Elling and Knoppers, 2005; Kapungu, Holmbeck, and Paikoff, 2006). But these efforts to examine intersections are remarkable for their rarity, perhaps because these efforts often face enormous practical, statistical and methodological challenges.

The second potential limitation in the field's approach to ethnicity involves the reality that ethnicity often occupies a secondary role as a moderator or qualifier of theoretical propositions. The practice of comparing ethnic groups certainly has had its critics (see McLoyd, 2003). Most notably, criticisms emerge not so much because studies are comparative but because of their characteristic interpretation of data and the restrictiveness of the provided information. For example, comparisons may result in assumptions that one group serves as the standard on which to judge others; indeed the very use of the term “minority” implies that a dominant “majority group” exists and, in the U.S., that typically means that the majority assumes a position of dominance (see Trimble, 1991). There is much to be said, for example, of research that specifically examines one ethnic group and the many physical, community, social, and psychological factors that can influence developmental outcomes (see Umañia-Taylor, Diversi, and Fine, 2002; Spano, Rivera, and Bolland, 2006; Ge, Brody, Conger, and Simons, 2006). The use of comparative analyses, however, does not necessarily mean that they are unhelpful in understanding ethnic adolescents. Comparative research often is of interest when it finds, for example, ethnic similarities and differences in related areas, such as recent findings revealing similar ethnic influences on adolescents’ body image (Nishina, Ammon, Bellmore, and Graham, 2006) while also finding that broad social forces, such as community poverty, differently influence different ethnic groups’ actual weight (Wickrama, Wickrama, and Bryant, 2006). In addition, research that separately analyzes groups on similar variables can lead to important results that highlight the significance of different factors differently (or similarly) influencing adolescent development. Thus, we now know that similar psychological factors could mediate relations between ecological factors and ecological outcomes among different ethnic groups (Prelow, Weaver, and Swenson, 2006). Research questions still may determine the need for analyses; but it still means that researchers need to exercise care when they compare groups and ensure the validity of their constructs. We can learn much when we find differences among and invariance across groups; but analyses may lose their potential effectiveness if ethnicity occupies an ad hoc role.

The third example of potential limitations involves the manner research pervasively treats ethnic groups as discrete categories to which people belong and that explain some aspects of functioning. Yet, ethnicity is a complex multidimensional construct. Researchers rarely examine the strength, salience, and meaning of ethnic identity; and they also rarely examine the experiences and attitudes associated with minority status. As the field becomes immersed in constructivist ideas about the nature of identity and ethnicity, ideas about how to define and measure ethnic identities (including situated identity and acculturative status) become considerably complex, which challenges the use of commonly used methods and constructs. As researchers examine the construct validity and psychometric properties of instruments, our notions of underlying constructs change. Researchers who do consider and assess these aspects of ethnicity are able to determine better how ethnicity does impact outcomes of interest. These efforts certainly point us toward interesting directions, but it would not necessarily be wise to require research to move in any of these directions. The use of ethnic identity and acculturation measures can create difficulties in defining appropriate sample frames. Moves toward greater specification may themselves become problematic and lose utility in that, the finer the definition, the less likely there will be another group like it. Addressing issues of generalizability is something fraught with peril, but the challenges do not vitiate the need to recognize and better respond to the multidimentionality of ethnicity.

A last example of the failure to take diversity as seriously as it could deals with international research, which in our study was conceptualized as research studies that make use of non-U.S. samples. We already have noted that the mere use of international samples does reveal considerable diversity. Yet, many of the arguments concerning the need to enhance our understanding of ethnicity in the United States could be extended to other countries. Our analysis did find some, very recent, efforts to identify the ethnicity of participants in studies conducted outside the United States. There also are interesting studies emerging that look at acculturation and immigration in international samples (see Vazsonyi, Trejos-Castillo, and Huang, 2006). Overall, however, international research, which constitutes a large percentage of published research on adolescence, ignores issues of ethnicity. Yet, samples may not be, for example, representative of the cultural and ethnic diversity within relevant countries; and groups within countries may well have different views and experiences and, as a result, attach different meaning to the investigated factors. These issues certainly bring limitations in the study of adolescence to yet another level of complexity.

Limitations

Just as our study uncovered important limitations in the manner researchers study ethnicity and adolescence, this study itself has important limitations worth emphasizing. Our study is limited to the reviewed journals. Other journals publish articles that have adolescent participants. For example, journals dealing with children (e.g., Child Development) often publish articles that include adolescents, as do public health journals (e.g., Journal of Adolescent Health) and those pertaining to criminology (e.g., Journal of Research on Crime and Delinquency), not to mention the dozens of journals relating to education. In addition, several journals now focus on ethnic and minority issues and it does appear that those journals increasingly draw studies that would have otherwise been published in the more general journals or not published at all (Imada and Schiavo, 2005). The inclusion of more journals could have revealed different patterns. The focus on 6 journals specifically targeting the study of adolescence, however, does support the proposition that we have identified a good segment of research.

Another important limitation deals with the reporting of participants’ characteristics. As we have seen, several studies do not seem to be reporting the actual number of minority participants. The failure to report specific characteristics does not mean that studies lacked an ethnic focus. Studies that did not report the ethnic make-up of their samples often conducted ethnic analyses; it was just unclear how many in their overall group were from ethnic groups. This most likely led to an underestimate of the percentage of articles with an ethnic-focus. Importantly, the studies that did not report specific numbers of participants from at least one group (including European Americans) tended to be quite large and, not surprisingly, their data served as the foundation for many articles. For example, several studies using the Add Health data do not readily report the ethnic composition of their studies but still do analyses that control for ethnicity. Although this may contribute to underestimates of the place of ethnicity in adolescent research, these articles were not counted as ethnic (or not). Those articles simply were not part of the analysis because of their lack of precision in their sample descriptions. The rationale for their exclusion was simple: we would know about trends in findings but would have a difficult time evaluating their real significance. A lesson to be learned is that authors wishing to influence the field must provide specifics.

Our findings likely are impacted by the number of researchers in this area and their relative productivity. Our analysis could have benefitted from knowing the actual number of researchers conducting research with ethnic groups, the number of different data bases used, the proportion of manuscripts submitted each year, and the proportion of these manuscripts that are accepted or rejected. It is possible, for example, that overall publication rates are low, but that journals may in fact be accepting these studies at higher rates relative to others. It is difficult to discern the existence of biases in the peer-review process or among editors or if attention to ethnicity relates to the number of published articles dealing with ethnic adolescents. These considerations gain significance to the extent that perceptions are important. The belief that the field does not publish as much as it should in certain areas could sway researchers, research, and publications away from considering the place of ethnicity in adolescent development.

Yet another significant limitation deals with the potentially detrimental byproduct of research that conducts content analyses like the one that we have just done. We have payed attention to sampling and methodological issues. That focus may be had at the detriment of the needed attention to broader systemic issues. Included in these issues would be concerns about marginalization and discrimination. Also included would be the risk of becoming preoccupied with differences among people at the expense of similarities. But, the concerns also go to our methods. Some have argued that researchers rely too much on internal validity, and that they do so at the expense of external validity (Sue, 1999). Others argue that current approaches to the study of ethnicity reveal a lack of theory driven research, lack of appropriate and valid psychometric tools, and an overreliance on convenience samples (Sue, Bingham, Porche-Burke, and Vasquez, 1999). These are quite legitimate concerns. Indeed, empirical research shows that even our very definitions and theories of ethnicity influence our findings (Umañia-Taylor, Diversi, and Fine, 2002). In a real sense, we are stuck with the limitation that those who study ethnicity need to realize that their object does not exist independently of their description of it; the mind has a tendency to find what it looks for.

figure a

Conclusions

This review certainly reveals important, albeit uneven, progress. The vast majority of studies including U.S. participants report the ethnic make-up of their samples. Some articles, however, do not describe their participants’ characteristics specifically enough, which makes us unable to decipher an accurate number or percentage of ethnic groups within their study. Overall, the leading outlets designating themselves to publishing research on adolescence do appear inclusive in their use of ethnic participants. But, that inclusiveness has only started to shape analyses. Having now recognized that ethnicity may matter and influence findings, our attention should turn to actual analyses and the extent to which the study of adolescence leads to the development of constructs that reflect the complexity of ethnicity and its relationship to developmental phenomena. Ethnic categorizations have very real social and political consequences to real people, and taking ethnicity seriously can help ensure that the scientific study of adolescence, both in the United States and abroad, can contribute to better consequences.