The Pragmatic American: Attributions of Crime and the Hydraulic Relation Hypothesis

Abstract

Abstract Attribution theory argues that a "hydraulic relation" exists between dispositional and situational attribution styles, causing people to endorse one style at the expense of the other. That is, attribution theorists predict that there should be a strong negative relationship between attribution styles. We test this prediction using data collected in Hillsborough County (Tampa), Florida, and two national polls. Our investigation shows that, rather than a bifurcated view of crime causation, Americans manifest a complex attributional style that views crime emerging from multiple sources. We discuss how these findings reveal that the American public tends to be not ideological but pragmatic in its view of crime causation and, ultimately, in the crime control policies it is willing to endorse. Keywords: public opinionattribution theorypunitiveness Notes 1. Our focus here is in exploring the alleged sharp distinctions that have been portrayed by those who endorse a dispositional attribution style—a punitive approach to crime—versus those who embrace a situational attribution style. We are not dismissing the fact that there are other pointed cleavages in public opinion when citizens are queried about crime. For example, Unnever, Cullen, and Jonson (Citation2008) show that there are sharp racial divisions in support for the death penalty; the majority of African Americans oppose whereas the majority of whites support capital punishment. 2. Choi et al. (Citation1999) highlight numerous reasons why Westerners such as those in the USA tend to gravitate toward a dispositional attribution style when compared to their Eastern counterparts. For example, they argue that Easterners are more sensitive to situational influences on behavior and view the group as the natural unit of agency whereas Westerners tend to view the individual as the unit of agency and tend to discount situational influences on behavior. However, scholars have called into question the argument that cultural scripts are deterministic causing individuals to make the fundamental attribution error (Funder, Citation2001; Sabini et al., Citation2001). 3. Of note, scholars argue that most of the citizenry have rather distorted and misinformed perceptions of crime (Campbell, Citation2008; Hutton, Citation2005; Roberts & Hough, Citation2005; Roberts & Stalans, Citation2000). Chief among them are the perceptions that crime is always escalating and that the public overestimates the proportion of crime that is violent. 4. Grasmick and McGill (Citation1994) and Cochran et al. (Citation2003) included the following item as a measure of a dispositional attribution style: "Most people who commit crimes were born to be criminals." We did not use this item as an indicator of a dispositional attribution because it negates the premise of offenders committing crimes out of their volition. In our sample, nearly 85% of the respondents disagreed or strongly disagreed with this item. We, however, reran our analysis including this item as a measure of a dispositional attribution style and the results were substantively the same as reported. 5. Although tangentially related to the purposes of our paper, we examined the degree to which including our dispositional and situational attribution styles mitigated the influence of our correlates on the punitiveness scale. The results from this analysis show that adding the two scales to the equation presented in Model 4 of Table 4 reduces the effects of the four significant variables: education by 45%, conservative by 56%, fear of violence by 11%, and victim by 76%. These results can be obtained by request from the first author. 6. Note that we reestimated the structural equation model after deleting the reverse coded rehabilitation item (Punish3) as an indicator of punitiveness from this analysis. The results presented in Figure 2 were substantively the same. They show that the correlation between the dispositional and situational attribution styles remained positive (.07), and the significance level was .06. A reviewer further suggested that we reestimate the model only including Punish2 and Punish5 in the model. Again, the results remained substantively the same; the correlation between the dispositional and situational attribution styles remained positive (.07). In addition, we reran the structural equation model presented in Figure 2 using an ordinal measure of support for the death penalty for adults convicted of murder (1 = strongly disagree to 4 = strongly agree) as our assessment of punitive attitudes. And again, the results remained substantively the same; the correlation between the dispositional and situational attribution styles remained positive (.08). Furthermore, we deleted the reverse coded rehabilitation item as an indicator of punitiveness from the analysis presented in Model 4 of Table 4 and reestimated the regression equation. Once again, the results presented were nearly identical, with the only difference being that the fear of violence measure did not attain statistical significance. Also note that we examined other covariates such as marital status and ethnicity (i.e., Hispanic) and their impact on having a consistent attribution style, a dispositional attribution style, a situational attribution style, and punitiveness (see Table 4). None of these factors significantly predicted any of these measures. 7. For a description of these data see Unnever, Cullen, and Fisher (Citation2007). 8. Future elaborations of these findings may wish to investigate whether a discounting principle takes place for those individuals who strongly identify with a situational attribution. Researchers may want to pursue the reasons why these individuals discount dispositional characteristics even when they are present. 9. As a further analysis of the data, we examined whether Americans who are not adamant supporters of punitive policies—those in the "middle of road"—are more likely to attribute crime to both dispositional and situational factors. We tested this hypothesis by limiting our sample to those who scored between one value below and above the mean score on the punishment scale (mean = 13). This sample includes 263 individuals. We correlated this new measure with our dispositional and a situational attribution style scales. The results show that both dispositional and attributional attribution styles are positively correlated with "middle of the road" individuals, .08 and .08 respectively. In addition, we created two binary variables. The first variable coded people who scored below the mean on both the dispositional and situational scales as "1" and people who scored 1 standard deviation above the mean on the dispositional scale were coded "0." The second binary variable coded people who scored below the mean on both the dispositional and situational scales as "1" and people who scored 1 standard deviation above the mean on the situational scale were coded "0." Thus, these two measures contrast people who embrace both attributional styles with those who either just endorse a dispositional or just a situational attribution style. We correlated the first binary variable with the four items related to punitiveness (Punish 1, Punish2, Punish 4, and Punish5). The results from this analysis show that individuals who endorse a more complex attributional style were significantly less likely to support a punitive approach than those who tended to solely embrace a dispositional attributional style. We then correlated the second binary variable with the item related to rehabilitation (Punish3). The results reveal that individuals who endorse a more complex attributional style were significantly less likely to support a rehabilitative approach than those who tended to solely embrace a situational attributional style. Together, these results suggest that individuals who embrace both attribution styles are statistically distinct from those who tend to embrace only one attribution style and they are more likely to endorse a "middle of the road" approach to crime; that is, they are more pragmatic. 10. Future researchers may wish to examine whether this shift in public opinion was accompanied by a parallel change in the percentage of individuals endorsing a dispositional attribution style rather than an external attribution style. 11. There may be other crime control issues that tend to polarize public opinion more so than those discussed here. For example, Mascini and Houtman (Citation2006) argue that issues such as decriminalization may produce sharper and more polarizing distinctions among conservatives and progressives.