Beliefs about Marriage, Homosexuality, and Race

The 2004 election campaigns clearly placed marriage on the political agenda. Nevertheless, the regulation of marriage tends to be a lowsalience issue in American politics and is generally discussed in broad symbolic terms such as “traditional values.” Even without a high level of salience, the public maintains relatively strong preferences about marriage regulations; this is particularly evident for laws regulating same-sex and interracial marriages (Barlow and Probert 2004; Bowman et al. 2004; Firebaugh and Davis 1988; Kuklinski and Cobb 1998; Moran 2001; Quillian 1996; St. Jean 1998; Schuman et al. 1997). Indeed, in 2000, just over 40% of voters in Alabama rejected the repeal of the state constitutional ban on miscegenation (Knigge and Moody 2003) and 13 states adopted constitutional bans on same-sex marriage during the 2004 elections. And, as noted, these opinions are strongly shaped by religious beliefs and adherence to those beliefs (Barlow and Probert 2004; Moran 2001; Schuman et al. 1997; Sherrill and Yang 2000; Westervelt 2001). However, the data in Figures 1 and 2 lead us to suspect attributions about the biological origins of behavior play an even larger role in shaping opinions on marriage policy. To understand this phenomenon more clearly we turn to attribution theory.

Attribution Theory

Attribution theory suggests that individuals develop causal theories to explain the world around them, including explanations of why people behave the way they do and why some people perform better in society than others (Fiske and Taylor 1991; Heider 1944, 1958). Observed behaviors are attributed to the person (internal or dispositional) or the environment (situational). The manner of response to behaviors generally depends on whether dispositional or situational attributions are made.

Weiner (1979, 1985) proposed an additional dimension, namely controllability. Whereas dispositional and situational factors refer to perceived causes of behavior, controllability has to do with whether the behavior is perceived as controllable or uncontrollable. The emphasis on controllability has been especially useful for analyses of causal attributions related to stigmatized groups. For instance, Crocker, Cornwell, and Major (1993) found that overweight women who believed their condition was controllable were more likely to perceive prejudicial responses regarding their stigma as warranted. Similarly, DeJong (1980) discovered that evaluations of obese people are more negative when their obesity is attributed to voluntary overeating as opposed to a psychological malady. More generally Weiner, Perry, and Magnusson (1988) established that stigmas with a behavioral origin such as child abuse, AIDS, and Vietnam War Syndrome were more likely attributed to controllable causes than stigmas with a physical basis such as Alzheimer's, heart disease, and cancer. Furthermore, in some cases attributing stigmas to a controllable cause elicited greater anger and negative affect toward the stigmatized.

Existing research has applied these principles to understanding the social stigma of homosexuality. Researchers hypothesized that people who considered homosexuality as a controllable state would report more negative attitudes toward homosexuals than those who viewed homosexuality as uncontrollable. Several studies have confirmed this, finding that people who viewed homosexuality as genetic or biologically based reported less negative attitudes about homosexuals (Aguero, Block, and Byrne 1984; Schneider and Lewis 1984; VanderStoep and Green 1988).

The theory suggests that a stigma perceived as controllable leads to more negative emotions and evaluations of the relevant group or individual. A genetic or biologically based attribution about the origins of homosexuality suggests people believe homosexuality is not controllable. However, a belief that homosexuality is acquired through social context, learned, or a personal preference suggests a belief that homosexuals can control this condition, and therefore can change their sexual orientation. If sexual orientation can be changed, then perhaps there is less of a need for public policies that protect individual orientations. Thus, the causal attribution individuals make should influence their support for same-sex marriage.

We can apply the same theoretical lens to opinions about legal bans on interracial marriage. However, in this instance, an explicitly negative biologically based attribution, one also considered uncontrollable, we expect to lead to less support for African-American rights. In fact, Kinder and Mendelberg (1995, 406) refer to beliefs that blacks have lower social and economic achievement because of the “inadequacies of blacks themselves” as biologically based forms of prejudice, but do not couch their arguments in attribution theory (see also Nelson 1999). Using their language or formulations from attribution theory makes little difference. If individuals believe intelligence differs between races, and attribute these differences to biological, uncontrollable determinants such as in-born ability to learn or level of motivation, this attribution should in turn increase support for laws banning interracial marriage. Why support interracial marriage when it may produce offspring considered biologically inferior (Stoddard 2003)? This logic grows from distorted attributions linking perceived deficiencies in blacks to uncontrollable biological sources. To avoid spreading biological defects, individuals who report such attributions are far less likely to support interracial dating or marriage, and are likely to support legal bans on interracial marriage (Schuman et al. 1997). We then hypothesize that those individuals who attribute the relative poor position of blacks in American society to a biologically based defect, such as an innate inability to learn, will be more likely to support laws that ban interracial marriage.

Data and Analysis

For our model of opinion on same-sex marriage we use a question from a May 2004 Gallup survey of national adults. In that survey Gallup asked respondents: “Should homosexual marriages be recognized by the law as valid?” Almost 37% of respondents stated homosexual marriages should be valid, while 64% reported they should not be recognized.

Our central independent variable is taken from the question: “In your view, is homosexuality something a person is born with, or is homosexuality due to factors such as upbringing and environment?” Respondents were allowed to refuse these choices and indicate both or neither. We only use those respondents who indicated “born with” (42%) or “upbringing/environment” (45%). These responses were coded “0” and “1,” respectively. Based on previous studies of attitudes on gayrelated policy we also included a number of control variables (Bowman et al. 2004; Brewer 2002, 2003a, 2003b, Pew Research Center 2003; Sherrill and Yang 2000). These include gender, marital status, having children, race, religiosity, income, education, political ideology, and partisanship. We expect that women, whites, the less religious, the wealthy, the highly educated, liberals, Democrats, those with no young children, and the unmarried will be more supportive of laws that recognize same-sex marriage (Brewer 2002, 2003a, 2003b, Sherrill and Yang 2000).

For our model of support for laws banning interracial marriage we use the following question from the General Social Survey 1972 to 2002 cumulative file: “Do you think there should be laws against marriages between (Negroes/Blacks/African-Americans) and whites?” On average, over the 20 years the question was asked, approximately 77% of respondents reported there should not be and 23% indicated there should be. As noted in Figure 1, the percentage of respondents saying “yes” has declined steadily over time.

The main independent variable is based on the question: “On average Blacks have worse jobs, income, and housing than white people. Do you think these differences are because most Blacks have less in-born ability to learn?” Those indicating yes are coded as “0” (16%) and those indicating no are coded as “1” (84%).

Based on previous studies of support for interracial marriage bans (Firebaugh and Davis 1988; Kinder and Mendelberg 1995; Knigge and Moody 2003; Quillian 1996; Schuman et al. 1997; Tenzer 1990; Wilson 1986; Yancey 2001), and previous studies of opposition to policies that benefit African Americans (Bobo and Kluegel 1993; Carmines and Layman 1998; Nelson 1999; Schuman et al. 1997), we also included a number of control variables for respondent characteristics. These include gender, race (not including African Americans because they were not asked the key questions), income, education, political ideology, partisanship, religiosity, and whether or not a respondent resided in the South. We expect that women, non-whites, the wealthy, the highly educated, liberals, Democrats, the less religious, those with no young children, the unmarried, and non-Southerners will be less supportive of laws against interracial marriage (Bobo and Kluegel 1993; Firebaugh and Davis 1988; Kinder and Mendelberg 1995; Quillian 1996; Schuman et al. 1997; Tenzer 1990; Wilson 1986).