You may have noticed that conservatives and liberals see two different things when they look in the same place. Dr. Linda Skitka, a professor of psychology at the University of Illinois who writes papers about these things, led a team that looked into this great divide. Dr. Skitka published "Dispositions, Scripts, or Motivated Correction? Understanding Ideological Differences in Explanations for Social Problems" with four collaborators in 2002.
I would have liked a shorter title, but there you are. The paper looks at "attributions," the differing types of explanations used by conservatives and liberals for social outcomes. The research shows, first, that in social contexts including income, crime, homelessness, and obesity, liberals prefer (are more likely to adopt) situational explanations, explanations that emphasize the importance of the situation, including the institutional context. Conservatives prefer personal explanations, explanations that emphasize the behavior or character of the protagonist, such as presence or lack of discipline, etc. [Skitka 2002, p. 470]
Second, research shows that once the attribution is decided, liberals and conservatives differ concerning who is entitled to support from the community. Both liberals and conservatives are willing to provide aid when the scenario is uncontrollable, whether the cause was situational or personal. Both are willing to provide aid when the cause is external (situational) and controllable. When the cause is internal (dispositional) and controllable, however, and the protagonist might have avoided being in need, liberals are much more inclined to offer community aid than are conservatives.
Their report evaluates five studies designed to choose among three hypotheses about how the evaluations of liberals and conservatives differ on such questions.
The “dispositional hypothesis” is that some people "by disposition" prefer (are more likely to apply) a situational attribution, while others prefer a personal attribution. Those who prefer attribution to the situation are attracted to liberal analyses, those who prefer personal attributions are attracted to conservative analyses, because these analyses fit their preferences, their dispositions.
Social theorists have gathered evidence that early familial and socialization experiences can create a life-long filter that ultimately shapes the adult. Disciplinary experiences and the child’s reactions to them shape beliefs about human nature and the importance of norm-adherence. Life learning can bend toward a belief that human behavior is sinful (people are bad and the world is dangerous), or toward the belief that human behavior tends toward the positive. (See A Just World with Just Institutions.) [Skitka 2002, p. 470]
The “ideological script hypothesis” proposes that, for whatever motivation, people choose (in the U.S., in any case) one of two political ideologies to adhere to. Once this commitment has happened, people assign attributions in a way that agrees with their chosen ideological script. Under this hypothesis individuals would be expected to apply attributions that are appropriate for their chosen allegiance. This should apply to situations that are encompassed by the script (i.e., to political situations), but not necessarily to situations that aren't covered by the script. [Skitka 2002, p. 471]
The “motivated correction hypothesis” accepts the existing evidence that both conservatives and liberals are inclined initially to make personal attributions rather than situational ones. We are inclined to conclude that behavior occurs because of traits of the actor. We know that we judge others based on stereotypes, but we differ in our inclination to re-evaluate our initial judgments. Our initial judgments are known to happen quickly and with little effort. See, e.g., [Fiske 2012].
People are inclined to review their initial attributions in light of their ideological commitments. Since conservative ideology tends to place a high value on personal responsibility, the conservative initial attribution is commonly reaffirmed. Liberals, however, are likely to replace an initial personal attribution with a situational one, which may require intellectual effort. [Skitka 2002, p. 471]
The first study reported upon in [Skitka 2002] asked subjects to estimate the intelligence of the questioner and the answerer in an imagined quiz contest. (Similar to the "Jeopardy" experiment reported in A Just World with Just Institutions.) In the scenario that was described to the participants (seventy-eight undergrads), the questioner developed five questions to which he/she knew the answers, then asked them of the answerer. The imagined answerer got one of the five answers correct. The subject was informed that the questioner had developed his/her own questions. The scenario included the imagined dialog of each imaginary participant. Subjects rated the perceived intelligence of each imaginary party on a scale from -4 (very unintelligent) to +4 (very intelligent). [Skitka 2002, p. 474]
Averaged across all subjects, the subjects rated the questioner's intelligence at 1.51, and the answerer's at 0.89. Liberals alone rated the questioner and answerer as equally intelligent. Conservatives alone rated the questioner’s intelligence at 1.38, the answerer's at 0.39.
The results verified that liberals prefer situational attributions; conservatives prefer personal attributions. This result is consistent with both the dispositional hypothesis and the motivated correction hypothesis.
The second study was more complex than the first study, but it resulted in no important new evidence. Subjects were given an essay supposedly written by another undergraduate. The essay argued either in favor of or against either welfare reform or tuition hikes. Subjects were told that the imaginary essayist had been assigned to take the position that he/she took in his/her essay. Subjects were asked to judge, on the basis of the essay, what the essayist’s personal beliefs about the issue were.
Subjects judged that those who supposedly wrote essays against the issues (welfare reform or tuition hike) personally believed in the position they had advocated. Subjects judged that those who wrote in favor of welfare reform were personally in favor of welfare reform. They judged, however, that those who wrote in favor of tuition hikes were personally neutral about the idea. These results were the same regardless of the subject’s ideological preference.
The researchers hypothesize that this last result (believing that the person assigned to advocate tuition hikes couldn’t personally favor them) most likely occurred because the subjects would themselves be liable for the tuition increase. [Skitka 2002, p. 475]
The third study analyzed in this report used 1,639 voluntary American respondents, roughly representative demographically of the American adult population. The subjects used internet technology (ca. 2002) to respond to the five scenarios listed below, choosing either the personal or the situational attribution. (They received Web access in exchange for completing various surveys about once per week.) [Skitka 2002, p. 477]
| SCENARIO | EXPLANATIONS | Liberals (%) | Cons’tives (%) |
|---|---|---|---|
| *The prisoner was paroled (because) | he turned over a new leaf. | 39 | 28 |
| the prison was overcrowded. | 61 | 72 | |
| The doctor laughed long and hard at the joke (because) | he has a good sense of humor. | 45 | 44 |
| it was a funny joke. | 55 | 56 | |
| The lawyer tripped over his girlfriend’s feet while learning the new dance step (because) | the lawyer is clumsy. | 45 | 49 |
| it was a challenging dance step. | 55 | 51 | |
| *The man lost his job (because) | he was a poor performer. | 50 | 58 |
| the company had financial problems. | 50 | 42 | |
| The woman gave $200 to her favorite charity (because) | she is a generous person. | 82 | 84 |
| she needed an income tax deduction. | 18 | 16 |
The first and fourth scenarios are marked with an asterisk to indicate that liberals and conservatives gave statistically different responses in those two scenarios. The two starred items are also politically-valenced scenarios. The first touches on crime, the fourth on employment. The fourth offers attributions that are not just different in locus (personal vs situational). The personal attribution also requires the subject to denigrate an anonymous man. The final vignette touches on tax law, which may or may not have been politically charged in 2002, but regardless of that liberal and conservative responses were nearly the same.
Of the three non-starred scenarios, subjects slightly preferred a situational attribution for the doctor and the lawyer, but a personal attribution for the philanthropix.
Liberals were equally likely to attribute the job loss (scenario four) to a dispositional cause or to a situational cause. Conservative respondents preferred the personal attribution (he was a poor performer) with a majority of 58%.
The first scenario, about the release of a prisoner, might be somewhat complex for a conservative respondent. Since conservatives tend to believe that the cause of crime is dispositional, and therefore a stable attribute of the individual, a conservative should be reluctant to believe that the criminal could “turn over a new leaf.”
The researchers believe that these outcomes support the motivated correction hypothesis. In the non-political scenarios, both ideological groups made similar attributions, disconfirming the dispositional hypothesis. The observation that conservatives reasoned through the prisoner release scenario to choose a situational cause suggests, but doesn’t prove, that a motivated correction process occurred.
In their fourth study, the research team reanalyzed telephone interviews that were recorded in 1987. Subjects were asked to respond to the following:
“Some people feel the government in Washington should see to it that every person has a job and a good standard of living. Others think the government should just let each person get ahead on their own. Which is closer to the way you feel or haven’t you thought much about it?”
Subjects were allowed to respond to this, then were prompted in a non-directive way to further comment, and each of their comments was recorded separately. The researchers hypothesized that the first comments of either liberals or conservatives would suggest a personal attribution, but that liberals would tend to move toward a situational attribution in subsequent comments. (This would be motivated correction.) Since most subjects provided either one or two comments, they limited their analysis to the first two comments.
Overall, including first and second comments, the attributional comments heard concerned what had caused the need for assistance, supporting the idea that the reason a person is in need is a concern for both conservatives and liberals. Seventy percent of conservatives’ comments were about the cause of need. Among liberals, 57% of the comments were of this type. The breakdown of comment types is shown in the following table. [Skitka 2002, pp. 478-9]
| Comment Type | Conservatives (%) | Liberals (%) |
|---|---|---|
| Personal attribution | 45 | 25 |
| Situational attribution | 10 | 9 |
| “Correction” | 10 | 20 |
| “Reverse correction” | 5 | 3 |
| Non-attribution | 30 | 43 |
A “correction” means a pair of responses that moved in the direction of a situational explanation, while a “reverse correction” means a move in the direction of a personal explanation.
Significant comments about these results include that conservatives were again more likely to offer personal explanations; liberals were much more likely to “correct” their response, suggesting motivated correction again; liberals were more likely to make comments that weren't concerned with attributions.
The fifth study presented in the paper was designed as a direct test of the motivated correction hypothesis. Motivated correction requires cognitive effort. The three hundred undergraduate subjects were divided into two groups so that one group could be stressed. Both groups were required to evaluate and decide which applicants for medical assistance would receive help, based on brief descriptions of the applicants’ relevant histories. One group performed these evaluations without interference, while the other group had to make the evaluations while simultaneously listening for, counting, and recording repeated musical tones. If motivated correction was taking place, the interference should affect the results of the second group. [Skitka 2002, p. 480]
This experiment was also designed to reveal how the attributed cause would affect the subjects’ judgments. Attribution can be broken down into four types: cause is either internal (personal) or external (situational), and either controlled by the person or uncontrolled. The study took place in the early 2000s, at which time the AIDS epidemic was much in the public view, the way it spread was generally known, and treatments were available but too expensive for many to afford.
Each subject was given eight applicants to judge, one heterosexual applicant corresponding to each cause attribution, and one homosexual applicant for each, as described in the table:
| Attributed cause | Description | Responsibility |
|---|---|---|
| Internal controlled | “contracted AIDS by engaging in high-risk sexual behavior despite knowing how AIDS is transmitted and the risk associated with these behaviors” | 7.69 |
| Internal uncontrolled | “practiced safe sex after learning how the AIDS virus is transmitted, but was exposed to the AIDS virus before it was widely known that AIDS was a sexually transmitted disease” | 4.55 |
| External controlled | “contracted AIDS from a long-term but unfaithful partner” | 4.06 |
| External uncontrolled | “contracted AIDS from a blood transfusion before the AIDS antibody test was developed to screen blood” | 1.53 |
Both homosexual and heterosexual applicants were included in the design because of concern that subjects might infer the applicant’s orientation from the scenario description, thus confusing the interpretation of results. Prior research had indicated that orientation wouldn’t affect judgments if attribution of cause information was provided. This expectation was confirmed in the results.
Subjects were instructed to complete a number of ratings of each of the eight applicants, then to choose any number of the applicants to receive subsidized treatment.
Subjects separately rated the responsibility and the blameworthiness of each applicant. These ratings were highly correlated (r = 0.89). The responsibility ratings are shown in the right-hand column of the preceding table.
The “internal controlled” applicants were rated as most responsible (for their AIDS infection), 7.69 on a scale from 1 (not at all) to 9 (very much). The “internal uncontrolled” were given a responsibility of 4.55, and the “external controlled” were rated 4.06. Least responsible were the “external uncontrolled,” with a responsibility of 1.53. Though there is a broad range of rated responsibility among the final three classes of applicants, the subjects treated the applicants in these three classes similarly, so the analysis treated the three as one group, labeled below as “non—responsible.”
The following table shows the subjects’ decisions. The numbers in the cells are average number of applicants granted medical aid.
| Applicant | Liberals | Conservatives |
|---|---|---|
| - With Distraction - | ||
| Internal controlled | 0.77 | 0.87 |
| Non-responsible | 1.53 | 1.44 |
| - Without Distraction - | ||
| Internal controlled | 1.39 | 1.13 |
| Non-responsible | 1.75 | 1.64 |
In the “without distraction” condition, liberals granted aid to significantly more internal controlled (responsible) applicants than did conservatives. In the “with distraction” condition, liberals granted aid to virtually the same number of internal controlled applicants as did conservatives. This result confirms the motivated correction hypothesis and supports the idea that people apply ideological reasoning when attributing causes in political contexts. Analysis of personal/social situations seems to start with assignment of attribution. The default assignment is personal rather than situational attribution. This seems to be followed by a check against ideological values, which is effortful. If a discrepancy is found, correction is made.
Additional analysis of the data led the researchers to conclude that the “motivated correction” process of the liberal subjects was stalled by the distraction task, which prevented them from developing positive affect toward the applicants who were “responsible” for their own infection.
Although these results identify more instances of liberal motivated correction, the theory suggest that this is a result of the kinds of situations that were included in the study. In situations in which conservative ideology is challenged by the default personal attribution of cause, conservatives are predicted to perform motivated correction, as illustrated by the prisoner-release scenario in the third study. [Skitka 2002, p. 483]
The strategy of reviewing a personal attribution and replacing it may be more generally used than suggested by these studies. For example, failures by the in-group or by in-group members may be explained by situational causes because a personal attribution is undesirable. Similarly, personal attributions may be selected in the case of positive outcomes. Psychological stakes may be higher when values are at stake. [Skitka 2002, p. 484]
This model suggests that it is easier to get a liberal to come to “conservative” conclusions than vice versa. The model suggests that amending personal attributions is work:
“It is much easier to get a liberal to behave like a conservative than it is to get a conservative to behave like a liberal. Liberals act like conservatives when resources are scarce, cognitive load is high, and aid serves secondary rather than primary needs....Conservatives only act like liberals when they are asked to consider helping a person with IC [internal controllable] causes of need who has convincingly reformed (convincing people that the personally responsible have reformed is not an easy feat...).”
It seems that everyone’s default approach is to use personal attributions, but “there is no reason to believe that liberals ‘own’ second-stage reasoning.” [Skitka 2002, p. 483]
Skitka and her collaborators classified attributional reasoning with other automatic cognitive processes and warned that “...some research...indicates that actively trying to suppress thinking about automatically activated concepts (e.g., stereotypes) leads to rebound effects. Once people stop suppressing the category...their subsequent judgments have higher levels of stereotypical content than those who never tried to suppress thoughts about the category.” Like bindweed, swtereotypes have deep roots, so they're hard to get rid of.
Skitka et al. mused about the motivation for second-stage reasoning. An obvious suggestion is that a conflict between the default personal attribution and ideological values motivates it. Would that mean that we routinely monitor certain classes of our judgments? What might motivate a conservative to evaluate situational attributions? [Skitka 2002, p. 484]
This paper also stirred my muse. If, as we can speculate, we do routinely monitor our judgments, do we check them against our beliefs as they are framed by the ideology? What's an Ideology? explains the difference between ideology and reasoning. I've also made a rough start at articulating my concerns about some of the planks of conservative ideology. Any conservative worth his paper and ink could express analogous criticisms of liberal ideology.
This and other research has demonstrated that there is a general tendency to see human behavior as motivated. [Skitka 2002, p. 483] Behavior in response to motivation is AKA free will. This is a major topic of neuroscientist Steven Pinker's "Blank Slate." [Pinker 2002] The topic has enough ramifications and vicissitudes that he wrote a book about it.
The question of motivations versus situations has non-political ramifications, sometimes in unexpected places. If you suffer chronic pain, for example, you will encounter care providers who can't seem to help believing that your behavior is personally and not situationally motivated. When they're right, it's good news. When they're wrong, it isn't they who feel the consequences. Of course technically a chronically irritated body bit is internal....
The linkage of attribution to responsibility also bears heavily on the moral justification for coercive actions toward others. It is sobering to ponder that those among us who are most inclined to condone coercion are also most inclined to a personal causal attribution.
It may be more difficult to envision situational than personal attributions. Our mirror neurons, to take a fundamental example, suggest to us that someone whom we observe performing some act is doing it for the same reason that we have done or would do it. See Wikipedia: Mirror neuron.
There is a connection between a preference for personal attribution on the one hand and belief in a just world or system justification on the other. Perhaps most obvious, if the individual bears responsibility, the systems shouldn't be blamed, as conservatives are wont to say. Less obvious is the cognitive expense of a situational attribution. The default personal attribution is first applied; the scenario is re-processed through the ideological filter and the situational explanation is applied; the many "what abouts" are processed. In the frame of motivated cognition theory, the situational attribution may be much more expensive (cognitively), and so provide a motivation for the less expensive belief. These expenditures needn't be made if you're RWA and full-square behind the status quo or if you're SDO and don't really care.
In the fifth study, subjects decided who should receive funding for (apparently) otherwise-unaffordable life-saving medical help. The subjects, undergraduates, were willing to fictionally doom fictional patients to a miserable fictional death on the basis of probably-transitory “lapses” of judgment in a situation in which infection risk was probably low. In making these sorts of judgments, or at least in this case, the consequence (should I say punishment?) for the decision was incommensurate with the patients’ perceived risk when they made their choices. There is an analogy with the spreading of COVID, which endangers others who haven’t necessarily made any decision. Yet lackadaisy about COVID seems to be more a conservative attitude. I can't reconcile the two situations.
Next: Free and Fair Markets
More information:
[Skitka 2002] “Dispositions, Scripts, or Motivated Correction? Understanding Ideological Differences in Explanations for Social Problems,” Skitka, L. J., Mullen, E., Griffin, T., Hutchinson, S., and Chamberlin, B., “Journal of Personality and Social Psychology,” Vol. 83, No. 2, 2002.
Wikipedia: Just-world hypothesis
Wikipedia: Attribution (theory)
© 2021, Ross A. Hangartner