A preliminary study of how multiple exposures to messages about genetics impact on lay attitudes towards racial and genetic discrimination

Abstract

Media depictions of genetics have led to concerns that this coverage will lead to increased belief in genetic determinism and increased discrimination, including racism. Previous studies of single exposures to messages about genetics or messages about genetics and race have shown some increases in discrimination and racism. Since attitude change is linked to repeated exposure to many messages, this study aimed to identify the effect of multiple exposures to multiple messages about genetics on attitudes towards determinism, discrimination and racism. Results showed an increase in genetically based racism, no increase in general racist affect and no significant increase in belief in determinism. Based on these results, we suggest that genetically based racism is a combination of racist affect with belief that perceived differences in human characteristics are solely or primarily influenced by genetics and that a move towards genetically based racism has implications for social policy.

Keywords:

• genetics
• media
• determinism
• discrimination
• genetically based racism

The Human Genome Project and the first maps of the human genome have been the focus of extensive mass media coverage (Conrad and Weinberg 1996, Djick 1998, Wilcox 2003) and this research has begun to have an impact on clinical practice. These media depictions have led to concerns that coverage of genetic studies will increase discrimination, including racism, even as health promoters have begun to include genetic information in public health messages (Duster 1990, Rothman 1998, Graves 2001, Silva 2005, Hansen 2006). Some research has begun to explore what effects these media depictions have had on lay attitudes about genetics. A few studies have examined the impact of messages and news headlines on genetic determinism, and these studies have found that a single exposure to a message about genetics did not increase genetic determinism (Condit and Williams 1997, Condit et al. 2000, 2001, 2004a, Bates et al. 2004). One of these studies showed differences in non-race based discriminatory affect related to the content of the messages (Condit and Williams 1997). Another study has shown an increase in racism, but in that study the messages explicitly linked genetics, race and health (Condit et al. 2004a). Given that attitude change tends to be a product of cumulative exposure to multiple messages, it is important to assess the impact of multiple exposures to genetics-based messages on genetic determinism and to begin to explore in detail the components of messages that result in changes in discriminatory attitudes. No experimental studies about the impact of multiple exposures to messages about genetics on attitudes about race have been published.

This paper addresses the issue of multiple exposures to messages about genetics and how those messages impact on attitudes toward race, genetic determinism and genetic discrimination. First, our concepts of genetic determinism, genetic discrimination, genetically based racism and general racism will be situated at the intersection of previous research on race and genetics and research on media influence on racist attitudes. Second, the paper will present the results of a study involving multiple exposures of participants to messages about genetics. Those results suggest that, even when overall levels of racism are not increased by media messages about genetics, these messages may increase the extent to which lay people attribute perceived differences among groups to genetics (i.e. the genetic basis for their racism). Third, based on previous research and the results of the study, we speculate on the implications of such a shift and outline various pathways through which messages about genetics can influence racist attitudes and offer suggestions for future research and practice.

Intersections of race, genetics and racism

Scientific studies of possible genetic bases for race and race-based pharmacogenomics highlight three interrelated sets of issues. First, the same ethical and social concerns raised about genetics and messages about genetics generally also arise in studies of genetics and race. Second, these studies appear in socio-cultural settings where concerns about race and racism exist. Third, the intersection of genetics and race raises new concerns in addition to the salient issues from ongoing debates about the social impact of genetics and ongoing debates about the nature of racism in contemporary society.

Genetics

Scholars in the social sciences and the humanities have raised concerns about increased knowledge of genetics in lay people. Some concerns focus on pragmatic issues concerning the right to privacy, informed consent and creating mechanisms for insurance coverage of genetic therapies (Kevles and Hood 1992, Collins 1999, Clayton 2003, Guttmacher and Collins 2003). Some historical and critical studies focus on the link between eugenics and genetics as a basis for concern (Kevles 1980, Duster 1990, Larson 1995, Hasian 1996). Among the variety of critiques of representations of genetics in the media, two prominent concerns are that media portrayals will increase genetic determinism and media portrayals will increase discrimination based on one's genes.

First, critics worry about the possibility of genetic determinism: that lay people exposed to genetic information will begin to believe people are reducible to their genes. Dorothy Nelkin and Susan Lindee (1995) argue that mass media depictions of genetics send a message of genetic essentialism that “reduces the self to a molecular entity, equating human beings, in all their social, historical and moral complexity, with their genes” (p. 2). Silva (2005) notes, “Genes are positioned as the roots of identity, behavior, and health across a wide range of public media” (p. 101). Ellen Wright Clayton notes, “People tend to see genetic information as more definitive and predictive than other types of data, in the sense that ‘you cannot change your genes’ and that ‘genes tell all about your future’” (2003, p. 563). Second, critics worry that determinism will lead to discrimination. Nelkin and Lindee argue that the popularity of genetics is grounded in its promise to “read” one's genes and predict future fate as a result. These predictions could be used to limit an individual's opportunities: “If an employer, or educator, or insurer can make the case that the ‘predicted’ future status of their client matters, then discrimination – denial of opportunity for medical care, work or education – can occur with impunity” (p. 167). Ruth Hubbard and Elijah Wald (1993) argue that the images used to describe genes imply hierarchies that justify discriminating based on genetic condition: “When molecular biologists speak of genes as ‘control centers’ or ‘blueprints,’ this is testimony to the hierarchical models they use rather than a description of the ways in which organisms function” (p. 64).1

Critics have raised concerns that messages about genetics will lead to a belief that all human characteristics are determined by genes and that this belief will justify discriminating against individuals because of the presence or absence of specific gene variants. While concerns about determinism and discrimination are prominent in their own right, it has been argued that messages about genetics will increase racial discrimination: in other words, if people accept genetic determinism, this will lead them to believe that perceived differences between races are also shaped by genetics (Rothman 1998).

Racism and media

According to Martin Bernal (1997), “we live in an age and society in which race has become obsessional and has pervaded every nook and cranny of our social life and culture” (p. 75). One means by which race and racism pervades our culture is mass media. Media scholars have shown mass media routinely deploy stereotypes of racial minorities (Dates and Barlow 1990, Jhally and Lewis 1992, Turner 1994). The work of Robert Entman has examined racial stereotyping in the media through a number of perspectives and methods (Entman 1990, Entman and Rojecki 2000). His work has demonstrated that, apart from news coverage of sports, depictions of African-Americans appear primarily in stories about inner-city crime, and that African-American politicians are portrayed as playing to special interests. These studies show how media stereotypes influence racial attitudes and increase suspicion between racial groups. Additionally, the work of Entman and others has shown that direct expressions of racist attitudes – so-called “old-fashioned” racism – has declined and been replaced by indirect, or “modern” racism (Schuman et al. 1998, Entman and Rojecki 2000, Schneider 2004). Modern racism manifests itself as a lack of awareness or understanding of the problems that minority racial groups encounter, and it consists of three components: (1) anti-black affect or emotional hostility toward African-Americans; (2) resistance to African-Americans' political demands; and (3) the belief that racial discrimination no longer affects African-Americans (Entman 1990).

Many theorists of racism have argued that the belief in a biological or inherited component of race is central to racist attitudes: race is biological and therefore immutable, and this immutability justifies discrimination (Wetherell and Potter 1992, Holt 2000, Reilly et al. 2003). Perceptions of a biological basis for race might be exacerbated by messages about genetics, which would then increase racist attitudes. In addition to the contribution to racism generally, the intersection of race and genetics has the potential to raise specific concerns.

Genetics and race

Studies linking genetics and race and reports about them in the media raise concerns about genetic determinism, discrimination and racism. Yet, these studies also raise the possibility of a new problem: racism that is grounded in a belief in the genetic basis of race. Understanding this possibility requires addressing two areas: the scholarly examination of these issues, including scientific, sociological and ethical discussion of genetics and race, and lay understandings of the intersection of genetics and race.

Scientific, ethical and social studies

Some medical researchers argue that racial categories represent an adequate proxy for genetic variation among humans (Risch et al. 2002). Medical personnel are using race to identify groups of people who are most likely to have genes that impact the effectiveness of various drugs (Exner et al. 2001, Caraco 2004, Taylor et al. 2004). Other geneticists and medical researchers hold that race is a biologically meaningless category and contest the utility of race as a proxy for medically important genetic variations in many common chronic diseases (Goodman 2000, Schwartz 2001, Collins 2004, Rebbeck et al. 2006). Studies supporting the use of race as a genetic proxy reach their pinnacle in the production of race-based medications. The premier race-based medication is NitroMed's BiDil, a heart disease medication intended solely for African-Americans (Bloche 2004). In 2004, researchers in the African-American Heart Failure Trial compared the use of BiDil with a standardized heart disease treatment and found that BiDil was somewhat more effective (Taylor et al. 2004). This trial has been viewed as meeting the final set of requirements the FDA had set for approving the drug for public use (Bloche 2004). Research that assumes or explores race-based genetic differences and attempts to produce race-based medication continues in large part because of financial considerations (Foster et al. 2001, Daar and Singer 2005).

Studies of the ethical and social impacts of this scientific research highlight three main concerns. First, studies linking genetics and race potentially can stigmatize groups as well as individuals: this means scientists must develop means of informing and gaining assent from the communities they study and that countries must develop legal protections for these groups (Weijer and Miller 2004, Kahn 2006). Second, using race as a proxy for genetic differences obfuscates the complex interactions between health, poverty and lived experience (Goodman 2000, Sankar et al. 2004). Third, the pursuit of genetic explanations for perceived racial differences “reifies racial and ethnic classifications by reinforcing the notion of biological difference rooted in genetics. This reification leads to stigmatization of racial and ethnic minorities” (Braun 2002, p. 160).

Lay understandings

Most studies of public understandings of race and genetics have occurred within the context of pharmacogenomics, especially the possibility of using race as a proxy for genetic difference in medical and pharmaceutical practice. Individuals are suspicious of proposals to prescribe drugs on the basis of race (Bevan et al. 2003, Condit et al. 2003). Lay people have indicated that the effectiveness of race-based medication would be hampered by racism inherent in the medical community and because “admixture” – a combination of ancestors from different socially defined races – made effective prescription based on race impossible (Bevan et al. 2003).

The studies on pharmacogenomics indicate a distrust of medical institutions that would use race as a diagnostic tool and doubts about the efficacy of race as an accurate tool for prescribing medication. However, they do not highlight the role that lay people attribute to genes in shaping differences between racial groups. Lay people use genetics and ancestry, along with cultural factors, as the conceptual means of delineating racial categories (Lynch 2006). Ramsey et al. (2001) argue that individuals who already believe in race-based differences would switch between genetic, cultural and individual-based accounts to explain their belief. Condit et al. (2004b) propose a four-part model of how the public understands the role of genetics in shaping “race”:

1) lay people identify race primarily by physical features, but these identifications are categorized into a variety of groupings that may be regional, national, or linguistic; 2) they believe that physical appearance is caused largely by genetics, and therefore that race has a genetic basis; 3) lay people believe that perceived differences in traits thought of as “non-physical” are caused by factors other than genes, so that; 4) while lay people do perceive races as hierarchically arrayed, they do not necessarily attribute the basis of these hierarchies to genetics. (p. 253)

To the extent that race is embodied in physical attributes, race can be treated as genetic, but differences between races, from health outcomes to complex social behaviors, would have cultural and individual roots as well.

Media depictions of this research play a role in shaping lay attitudes. A study of lay responses to fictional direct-to-consumer advertisements for race-based medication showed that a majority of respondents did not exhibit an increase in deterministic or discriminatory attitudes (Bates et al. 2004). A study of possible public service announcements linking genetics, race and heart disease did show an increase in racism (Condit et al. 2004a), and a study of news coverage of genetics and race found a majority of newspaper stories conveyed the idea that races are biologically distinct (Lynch and Condit 2006).

Scholars studying the ethical impact, and public understanding, of research into genetics and race identified a unique issue raised by scientific research into this area. In addition to concerns raised about genetics and concerns raised about race, this research and media messages about it raise the concern that a new type of racism will arise that is grounded in a belief that races are genetically distinct. This genetically based racism would be independent of general racist attitudes but would interact with them as well as genetic determinism and discrimination (Condit et al. 2004b). Messages about genetics might help reify race by emphasizing a genetic basis for most human characteristics; messages that increased genetic determinism would therefore foster a genetically based racism. Given the reviewed arguments about genetic discrimination, racism and the biological or genetic bases for racism, we hypothesize that exposure to multiple media messages about genetics will increase genetic discrimination, racism and genetically based racism. This study tests that possibility through an experimental manipulation that assesses the impact of multiple exposures to messages about genes.

Methods

Previous research has established four related issues pertinent to understanding lay responses to messages about genetics and race that can be conceptualized as genetic determinism, genetic discrimination, racism and genetically based racism. These concepts have been operationalized, transformed into a set of written survey questions to measure the degree to which participants in our study professed attitudes that fit each conceptual category. The genetic determinism instrument measures the degree to which individuals believe that genes are the primary force shaping their lives (unpublished data). The genetic discrimination instrument measures the degree to which individuals hold attitudes supporting discrimination on the basis of one's genes, and it consists of five sets of questions measuring the areas where individuals might feel discrimination is justified (Parrott et al. 2005). Several different measures of racism exist. Two of them are Entman and Rojecki's (2000) racial denial scale, a well-validated and often-used instrument to measure racist attitudes and beliefs, and the modern racism scale, which aims to measure racist beliefs despite attempts by participants to give socially desirable responses that reflect contemporary disapproval of racism (McConahay 1986). The genetically based racism instrument measures the degree to which individuals believe that hierarchically organized differences between races are determined by genetics: for example, one question asks to what degree members of one racial group will commit crimes more than members of another group because of their genes (Condit et al. 2004b).

Participants for this study (N = 104) were recruited from introductory communication classes at a large public university in the southeastern United States. The study was approved by the university's Institutional Review Board, and informed consent was obtained from all participants. The study exposed participants to many messages that were designed to provide a relatively representative sample of messages about genes that people were likely to see in daily life, and to do so across three time intervals. Participants were required to come to the laboratory three times at one-week intervals and were given modest financial compensation ($20.00). A pre-test, post-test experimental design was employed owing to concerns about maintaining a comparable random sample across three time periods (different fall-out rates were possible among different race and income groups).

At the beginning of the study, participants filled out a written survey that included the genetic determinism instrument, the genetic discrimination instrument, a truncated racial denial scale, a truncated modern racism scale, and the genetically based racism scale. They then viewed six headlines displayed on large boards around a room as they watched a 12-minute videotape that included three commercials mentioning genes, as well as part of a documentary about genetically modified food. They then read a paragraph from an article associated with one of the news headlines. They returned one week later to view new headlines, a new news paragraph, and a repeat of the video. They returned a third time after an additional week to view headlines, a new news paragraph, and a repeat of the video. They then answered the survey again. A total of 13 participants who did not attend the second session were excluded from further participation.

Newspaper headlines were included because most people are exposed to more headlines about genetics than to full news articles (Wilcox 2003, Caulfield and Bubela 2004). Headlines were selected by using Lexis-Nexis to conduct a search of major newspapers from November 2002 to April 2003 using search terms “gene” and “for”. Seventy-five documents were retrieved and we selected the first 12 (sorted by topic relevance) that were about genetics and that were published in US newspapers, after removing duplicate topics (e.g. “obesity”) or duplicate sources. Headlines were rotated at each time period, so that participants saw each headline twice. News paragraphs were drawn from the same articles as the headlines. Videotapes included the available commercials with genes as content, a short out-take from a documentary about genetic modification of food crops and a news story covering the mapping of a human chromosome.

Results

Participants

Participants were asked to self-designate race, gender, ethnicity and income from closed option categories. As Table 1 indicates, participants were more female, more white, less Hispanic/Latino and had a lower income than the national average. They were also much younger, with an average age of 21.5. As the table indicates, there was a higher dropout rate among non-white, Hispanic/Latino and low-income participants.

Table 1. Study 1 participants at time 1 (N = 104) and time 3 (N = 91).

Gender t1	Female 60 (57.7%)	Male 44 (42.3%)
Gender t3	Female 52 (57.1%)	Male 39 (42.9%)
“Race” t1	White 87 (84%)	African Am. 4 (3.8%)	Asian 6 (5.8%)	Bi/multiracial 1 (1%)	Hispanic 1 (1%)
“Race” t3	White 79 (87%)	African Am. 4 (3.3%)	Asian 4 (3.3%)	Bi/multiracial 1 (1%)	Hispanic 0 (0%)
Ethnicity t1	Hispanic 4 (3.8%)	Non-Hispanic 100 (96.2%)
Ethnicity t3	Hispanic 2 (2.2%)	Non-Hispanic 89 (97.8%)
Income t1	< $20,000 47 (45.2%)	$20–35,000 9 (8.7%)	$35–50,000 5 (4.8%)	$50–80,000 9 (8.7%)	> $80,000 33 (31.7%)
Income t3	< $20,000 39 (43.3%)	$20–35,000 6 (6.7%)	$35–50,000 5 (5.6%)	$50–80,000 12 (13.3%)	> $80,000 28 (31.1%)

Genetic discrimination

The original genetic discrimination scale consisted of five subscales that measured a single (i.e. unidimensional) aspect of genetic discrimination. Each genetic discrimination subscale was assessed on a five-point, Likert-type scale (1 = strongly disagree to 5 = strongly agree). Further, exploratory factor analyses were employed to discover the latent dimensions of a set of items and ensure that each subscale measured a single factor. Within these factor analyses, varimax rotation was selected to provide a clear identification of each item with a specific factor. Results for these scales do not support the hypothesis of increased genetic discrimination following multiple exposures to messages about genetics.

Crime subscale

Two items (e.g. all persons who are arrested should have their DNA put on file in police departments) originally formed a criminal subscale. Across both time 1 and time 3, these items formed a unidimensional factor. Because participants responded to items within this scale fairly similarly, the items were internally consistent with one another (as determined by Cronbach's alpha) and the scale was thus reliable (time 1 α = 0.88, time 3 α = 0.87). Paired sample t-tests, which are used to compare two related groups (i.e. time 1 and time 3) in association with a dependent variable (i.e. genetic discrimination) revealed that genetic discrimination relating to crime did not significantly differ from time 1, where the mean answer on the scale was 1.36, to time 3, where the mean answer was 1.32. This difference is within chance variation (t = –1.35, df = 90, p = 0.18).

Insurance subscale

The insurance subscale (e.g. insurance companies should not discriminate against those who have genetic diseases) was comprised of five items. In time 1 and time 3, these items formed a unidimensional factor and were internally consistent, time 1 α = 0.90, time 3 α = 0.85. Paired sample t-tests revealed that insurance genetic discrimination did not significantly differ from time 1 (mean = 3.83) to time 3 (mean = 3.84, t = –0.101, df = 90, p = 0.92).

Personal reproduction and social reproduction subscale

The original social reproduction subscale consisted of three items (e.g. physicians should advise all prospective parents who have genetic flaws against having children), whereas the individual subscale included four items (e.g. I would not marry someone who has a high risk of getting a genetic disease). In the present study, at both time 1 and time 3, these seven items combined to form one unidimensional factor and were internally consistent, time 1 α = 0.90, time 3 α = 0.82. Paired sample t-tests revealed that the combined social and individual subscales did not significantly differ from time 1 (mean = 2.23) to time 3 (mean = 2.17, t = 0.943, df = 90, p = 0.35). Scales were also run separately because the sample size was too low to allow factor stability. There were no significant differences across time.

Organizational subscale

Three items originally formed the organizational subscale. In the time 1 factor analysis, items 18 (employers should have the option to not hire someone with a genetic disease) and 27 (employers should have the option to not hire someone who is more likely than average to get a genetic disease) cross-loaded with the insurance subscale items. In time 3, item 27 cross-loaded again with the items in the insurance subscale. Because a single factor did not emerge, no further analyses for the organizational subscale were conducted.

Racism

Truncated versions of the racial denial scale and the modern racism scale were used to assess racist attitudes and beliefs at time 1 and time 3. Each racism subscale was measured via a five-point, Likert-type scale with higher values indicating stronger agreement with each item (1 = strongly disagree, 5 = strongly agree). Exploratory factor analyses employing varimax rotation were also conducted to ensure that each subscale assessed one factor. The results of the racial denial scale and the modern racism scale do not support the hypothesis of increased racist attitudes following multiple exposures to the messages.

Racial denial scale

Across time 1 and time 3, the original four items (e.g. most Blacks who are on welfare programs could get a job if they really tried) formed a unidimensional factor and were internally consistent scales, time 1 α = 0.73, time 3 α = 0.82. Paired sample t-tests revealed that racial denial levels did not significantly differ from time 1 (mean = 3.35) to time 3 (mean = 3.33, t = 0.478, df = 90, p = 0.63).

Modern racism scale

The four modern racism items (e.g. it is easy to understand the anger of Black people in America) formed a unidimensional factor and was a reliable scale at both time 1 (α = 0.75) and time 3 (α = 0.81). Paired sample t-tests indicated that racial denial levels did not significantly differ from time 1 (mean = 2.92) to time 3 (mean = 3.01, t = –1.39, df = 90, p = 0.169).

Genetically based racial discrimination

Eight items (e.g. one race may be stronger than another because of genetics) form the genetically based racial discrimination scale. These eight items comprised a unidimensional factor in the current study in both time 1 and time 3, which constituted an internally consistent scale, time 1 α = 0.89, time 3 α = 0.93. Paired sample t-tests revealed that genetically based discrimination significantly increased from time 1 (mean = 2.55) to time 3 (mean = 2.80, t = –3.30, df = 90, p < 0.001). The effect size using means and standard deviations for both time periods is small-to-moderate, r = 0.123 (Cohen's d = 0.248). The results support the hypothesis that multiple exposures to messages about genetics increase genetically based racism.

Genetic determinism

The original genetic determinism instrument consisted of seven items that measured the belief that genes are the primary force shaping people's lives (e.g. genes make some people more likely to benefit from medicine than others). Two principal axis factor analyses using varimax rotation (one for each administration of the scale) each revealed that these items represented a single factor. Both scales were also internally consistent, time 1 α = 0.78, time 3 α = 0.80. Paired sample t-tests revealed that genetic determinism did not significantly increase from time 1 (mean = 3.97) to time 3 (mean = 4.08, t = –1.87, df = 90, p = 0.06). The modest sample size makes this result equivocal, assuming a cut-off p-value of 0.05.

Linking types of racism and messages about genetics

The results of this study present an intriguing picture of how lay individuals respond to repeated exposure to messages about genetics. The study suggests two things. First, it suggests that multiple exposures to these messages about genetics increase genetically based racism, but there is not an increase in levels of racism. The increase in genetically based racism, without a concurrent increase in levels of racist attitudes or genetic discrimination, may be explained by a switch in explanations justifying racism. It is possible that people are adding genetics to or switching from other explanations that justify their belief that differences exist between groups and their explanation of the source of that difference. This finding is consonant with the results of a study by Ramsey et al. (2001), which showed that people who have high levels of racism are able to “explain” the differences among groups with a variety of rationales: individual motivation, culture, family, neighborhoods or genetics. The current results suggest that exposure to information about genetics leads those with high levels of racism and genetic discrimination to attribute the cause of perceived differences to genetics.

Second, while the results are not definitive with regard to the increase in genetic determinism and the non-significant finding is consistent with previous single-exposure studies, the result in this multiple exposure study approaches significance, and this suggests the need for additional, more precisely targeted research. We suggest first that such studies need to specify levels of genetic determinism with regard to specific characteristics. Recent research suggests that people do not assign a global level of genetic causation to all human characteristics, but rather assign different levels of causation to different characteristics (Parrott et al. 2003, Condit et al. 2004b, Condit and Parrott 2004). The use of the global determinism measure may mask changes in the levels of determinism that are relevant to genetically based racism. Second, we suggest that future studies might profitably explore a deliberately planned range of levels of genetic determinism in the messages. The current study sought to provide a representative or at least typical set of messages, in order to provide the most naturalistic conditions. These messages include a wide range of levels of determinism, but on average may be relatively similar to the middle range of genetic causation that typifies the lay public. It would now be useful to explore whether more deterministic messages produce increased levels of genetic determinism. This might clarify the equivocal findings of the present study.

The pattern of results in the extant studies leads us to propose that attitudes about genetics and race are shaped along the following paths of influence (see Figure 1). We propose that perceived differences between “racial” groups plus a predisposition towards messages about hierarchical rankings of those differences lead to racism. At this stage of influence, a hypothetical individual exposed multiple times to messages linking genes, race and health might perceive, or believe he or she perceives, a difference between racial groups, and if the individual also believes in a hierarchical ranking of those groups, they will exhibit greater racist affect. Also, it is possible that the hypothetical individual will show an increase in the tendency to place groups in hierarchies as a result of this exposure, although we believe this is a secondary route of influence that messages about genetics, race and health might have.

Figure 1. Pathways influencing attitudes about genetics and race.

The racism developed through the perception of “racial” differences and hierarchicalization will combine with various beliefs about what causes human characteristics, whether they are genetic causes, socio-cultural causes or individual causes. This combination produces accounts of racial hierarchies that can be described as either genetically based racism or socio-culturally based racism. If the hypothetical individual exposed to messages that increase their racist affect believes society and culture are the primary determinants of human characteristics, he or she will account for hierarchies using a socio-cultural racism, while individuals who believe genes are the primary determinants will produce an account shaped by genetically based racism. Also, as individuals are exposed to different ways of accounting for so-called “racial” differences, the type of account they give might change: for example, if an individual shifts from a genetic determinism to a social determinism, the type of racist affect they exhibit will change as well. Finally, it is possible that messages emphasizing genetic causation will increase genetic determinism, which will contribute to an increase in genetically based racism or a switch to accounts of racial hierarchies shaped by genetically based racism. When individualist accounts of behavior are dominant, racism appears as uncaused, and the study by Ramsey et al. (2001) has indicated such cases do occur.

These pathways of influence are consistent with the existing pattern of results, in which messages linking genetics, race and health increase racism, and messages emphasizing genetic causation do not increase racism, but do increase genetically based racism. The shift to genetically based racism also would entail a significant shift in beliefs about the possibility and desirability of remedying differences. If differences between groups are a product of individual will or socio-cultural differences, they are not necessarily permanent, while genetically based differences would be permanent. Thus, programs that aim at remedying differences – whether it is the remediation of educational disparities through Head Start programs or addressing health disparities in the Health People 2010 initiative – would be of value from a socio-cultural perspective on difference but not a genetically based perspective. Also, the different perspectives on difference would alter the degree to which one thinks that individuals may vary from their perceived group “type”. In a socio-cultural account, individual variation from a stereotypical norm may make more sense than from a biological or genetic account.

Clearly, further studies will be needed. The impact of the shift from socio-cultural and individually based racism to a genetically based racism needs further examination: questions about how genetically based racism shapes views of social programs and individual variation are salient. Also, further study will need to address the attitudes of different racial and age groups. In our study, the majority of participants were very young (average age of 21.5 years). The Human Genome Project and increased attention to genetics in the media began over 15 years ago when they were young children. This might have led to early exposure to messages about genetics that had significant influence on their beliefs and attitudes towards genetic determinism and genetic discrimination. Also, participants were a predominantly white group of people attending a university in the southeast US. That region's history of racism might influence participant attitudes towards race. Further study that includes a diversity of ages, ethnicities and geographic regions is needed.

Despite these limitations, this study suggests the immediate need for more research on the impact of messages about genetics on lay attitudes, and the proposed model suggests a specific agenda for that research. Until the impact of those messages can be more definitively determined, we suggest caution when the public health or clinical community produces messages about genetics, especially those that link health, race and genetics for the public. The development of potential message strategies for mitigating negative impacts should be a priority.

Acknowledgements

This research was supported by grant no. 5 R01 HG02191-02 from the National Institutes of Health to CMC. The authors wish to thank Alan Templeton, Tasha Dubriwny, Kristan Poirot, Alison Trego and Richard Nabring for their assistance in data collection and instrument development.

Notes

1. See also Lippman (1992), Rothman (1998). For a different reading of the “blueprint” metaphor of genes that emphasizes non-deterministic and non-discriminatory interpretations, see Condit (1999).