A different view of authoritarianism was presented by Dr. Karen Stenner in her 2005 book, “The Authoritarian Dynamic,” Karen Stenner, Cambridge University Press, 2005. It describes Stenner’s work in this field over more than a decade.
Stenner’s work is incisive and important. Her book describes her multi-year process of study and is of interest not only for the results it presents, but also because she carefully explains the rationale for the form of her research.
One of Stenner’s major points is that the concepts and classifications used in instruments such as Altemeyer’s RWA and Pratto’s SDO are not sufficiently precise to serve the several levels of analysis needed to understand these political phenomena. She distinguishes authoritarianism from status-quo conservatism and from economic laissez-faire. She sees authoritarianism as a desire for oneness and sameness which looks to authority as the means to achieve them. She defines authoritarianism in terms of child-raising values and demonstrates a surprisingly strong link between those values and the various forms of intolerance that we associate with “authoritarianism.”
She argues that political conservatism in its varieties and authoritarianism are separable things which should be clearly distinguished, but which aren’t in much of the research.
Stenner's work involves many layers of analysis that are intermixed in work of her peers. It is a substantial challenge to describe one in terms of the other. It is also challenging to de-synthesize her message. Her book should be well worth the study for anyone who is deeply interested in this field, but it is a substantial undertaking. I’ve chosen for my purposes to highlight particular insights she has developed.
In this essay I will describe Stenner’s theoretical model; present evidence for her major hypothesis that expression of authoritarian disposition is influenced by perceived threat; describe and critique her propositions about child-rearing values and authoritarianism; and explain my reservations about her proposals.
I’m particularly excited to pass along what Stenner and her team heard from subjects when they conducted interviews with the most and least authoritarian subjects they found in a large mail-in survey. These live and raw events bring home what the abstractions represent.
In Stenner’s formulation, authoritarianism is a “predisposition concerned with the appropriate balance between group authority and conformity, on the one hand, and individual autonomy and diversity, on the other.” Under the sway of this disposition, people adopt a set of functionally related stances or strategies. These stances address a basic human dilemma, life as a member of a group versus pursuit of individual goals. (This view is similar to John Duckitt’s. See “A Causal Chain for Conservatism?”) Authoritarianism is characterized by rejection of diversity and insistence on sameness. Coercion becomes involved because it is seen as necessary to achieve sameness. [Stenner 2005, p. 14]
Authoritarian disposition can turn into a worldview that emphasizes the social value of obedience and conformity. Such a worldview is often called conservative or authoritarian.
A “normative threat” is a condition that activates the predisposition toward authoritarianism, and can include:
Common authority and shared values define a group, the basis for “us” and “them.” In this way norms and group identity are intermeshed. Normative threat activates authoritarian predisposition, making authoritarianism dynamic. [Stenner 2005, p. 17]
Stenner believes that authoritarians want to belong to a system that creates oneness and sameness, and that this goal for them is more important than the particular normative order. However, any system that can claim the loyalty of authoritarians will always be characterized by sameness and constraint. [Stenner 2005, p. 18]
The RWA scale (Right Wing Authoritarianism) confounds authoritarianism with conservatism, which Stenner seems to regard as a set of normative beliefs. The RWA questions include references to specific norms and authorities, and therefore the RWA metric doesn’t match her definition of authoritarianism as a need for groupiness. Authoritarianism is about sameness, not about resisting change, so it is different from status-quo conservatism. However, RWA can be a satisfactory scale for authoritarianism in some research. [Stenner 2005, p. 21]
Authoritarianism is about “authority and uniformity versus autonomy and difference.” A better measure for it than RWA can be found in child-rearing values. These are determined by asking subjects which is preferred, e.g., "that a child follows the rules” versus “that he follows his own conscience,” or “that he has respect for his elders” versus “that he thinks for himself.” Scales based on such questions seem to be equally valid regardless whether the subject was raised in a particular way, whether he/she has raised children this way, and whether or not he/she has raised children. Such questions, however, are of questionable validity if presented to subjects who are not far removed from being raised, in which case the question can be asked as “which appeals to you more?” Such questions, Stenner says, reveal fundamental values. [Stenner 2005, p. 23]
In her theoretical view, these fundamental values are concerned with preserving a normative order. The values can motivate various attitudes and beliefs, depending upon social, cultural, and institutional context. Attitudes serve needs and perform functions. (See “Motivation for Conservatism”). The needs served may include “adjustment,” “ego defense,” “value expression,” or “knowledge.” Ego defense functions are activated under threat. This provides a mechanism for the activation of dynamic authoritarianism. [Stenner 2005, p.25]
If you’ve been reading carefully, you’ve noticed that “authoritarianism” has already been used in this essay to designate a certain disposition and to designate both attitudes and behaviors. Stenner uses a substantially different definition of authoritarianism than the other scholars whose work we're looking at. In an attempt to keep meanings clear, I’m going to use my own coinage to refer to Stenner’s “authoritarianism,” specifically, the acronym AN, which stands for “Authoritarian Nurturance preference.” I'll use AN to refer to Stenner's authoritarianism score. I'll also use AN to refer to someone with a high score. I’ll use the acronym LN to refer to someone with a low AN score, based on Stenner’s use of “libertarian” for that purpose. When quoting Stenner, I’ll keep her terms.
The “Cumulative General Social Survey” has been ongoing since 1972. It is a large survey of Americans conducted by the National Opinion Research Center at the University of Chicago. It was taken twice a year from 1972 to 1994, the years which Stenner and her collaborator Feldman relied on for results they published in 1997. They found that increased variability in public opinion, high levels of protest, and recent turnover of political control between parties corresponded with increases in expressed racism, aversion to free speech, support for compulsory school prayer, and support for capital punishment among those more disposed toward authoritarianism. On the other hand, the occurrence of family financial distress, criminal victimization, or personal trauma (divorce, illness, or deaths) didn’t increase authoritarianism but instead induced “more tolerant and inclusive” attitudes. This supports the posited role of “normative threat” in expressed authoritarian attitudes. [Stenner 2005, p. 31]
The “National Election Study” is another long-running survey of American opinion, this one involving surveys before and after presidential elections, conducted out of the University of Michigan. Using data from this dataset from 1992, these researchers again compared childrearing values and perceived societal threat against expressed authoritarian attitudes, and again found increases in intolerance, in militarism, in support for the death penalty, in favoring order over freedom, and in derogation, stereotyping, and discrimination against out-groups, particularly among those disposed to authoritarianism. Threat was indicated by perceptions of ideological diversity, negative evaluations of presidential candidates and political parties, perceptions of national economic decline, and fear of nuclear war. The most significant of the threats were perceptions of ideological diversity and negative evaluations of politicians and parties. Perceptions of personal threat (such as financial insecurity) were “inconsequential” with respect to authoritarian attitudes and “more often than not” reduced authoritarian attitudes.
From these and other studies, Stenner suggests that what we refer to as “conservatism” or “authoritarianism” may actually be made up of various combinations of AN (measured by child-rearing values), status-quo conservatism, and laissez-faire conservatism. These three components have separate motivations and effects. Conservatism is motivated by a desire to dampen change. Laissez-faire conservatism in her view may be motivated by a preference for freedom over equality. [Stenner 2005, pp. 138 et cf.]
Well by and by, way after many years are gone,
And all the war freaks die off, leavin’ us alone
We’ll raise our children in the peaceful way we can
It’s up to you and me, brother,
To try and try again.
Ain’t Wastin’ Time No More
Greg Allman
Stenner organized a study, the “Durham Community Survey 1997” or DCS97, to gather more data. DCS97 began as a mail-out/mail-back survey. It was mailed out to 1,200 randomly-chosen adult residents of the Durham, NC community. (Durham is a city of about 280,000, part of the larger metroplex known as the “Research Triangle.” It is a college town with a number of large universities in the area, including Duke, North Carolina Central, and UNC Chapel Hill.) The return rate for the questionnaire was 35%. Those returns included 361 non-Hispanic white subjects. These made up the subject group of primary interest to the investigation, which was concerned with the attitudinal makeup of majority group members. [Stenner 2005, p. 38]
After the survey was tabulated, the researchers selected the thirty highest and thirty lowest scoring questionnaires on the authoritarianism/child-rearing questions. These are referred to here as “AN” and “LN” subjects. (“A” is for “authoritarian“ while “L” is for “libertarian,” the term.) They attempted to arrange in-person, at-home interviews with each of the sixty, and succeeded with 27 of the ANs and 19 of the LNs. This resulted in forty completed interviews.
The interviews were to be loosely structured. The interviewers were provided with thirteen broad and open-ended questions around which to structure the interviews.
Six interviewers were used, five undergrads and one grad student from Duke. Interviews were done by pairs of interviewers. The primary interviewer was to arrange the interview over the phone, including a cash payment to be negotiated with the subject in the range $10-30. The primary was also to ask the questions and otherwise interact with the subject during the interview. The interview partner was to operate the voice recorder, take notes during the interview, and provide social support for the primary. Both interviewers were required to log their answers to structured and open-ended questions about their subjects and the events of their interviews.
Four of the student interviewers were white (three of these female), two of them black (one female). Interviewers were assigned to pairs and to subjects entirely at random. This allows the effects of interviewer race to be inferred. Stenner insists that all six interviewers were low-key and highly innocuous. [Stenner 2005, p. 205]
Some effects were observed before any interviews occurred.
Interviews were negotiated according to a detailed script in order to standardize the negotiations. According to interviewers’ evaluations, the AN subjects were more reluctant to be interviewed and more concerned about the size of their payments. When the interviewer was black (there is substantial evidence that Americans can identify a “black voice” over the phone), this effect was increased, and more suspicion and hostility were evinced. (One anecdote reported that an authoritarian subject repeatedly gave confusing or incorrect directions to his/her home after having agreed to be interviewed.)
When the interviewer was black, the AN subject was twice as likely to ask about his/her payment before the team was allowed inside. Both members of the interview pairs felt that AN subjects were more hostile and suspicious during the interviews. (Stenner argues that this reflects characteristics of the subjects rather than behavior of the interviewers.) Five of the subjects complained to the interviewers that they had never received payment for the mail-in questionnaire they had completed eight months earlier. All five of these were ANs. [Stenner 2005, p. 209]
Metrics were later taken from the interview transcripts. ANs on average said 1,400 fewer words during their interviews, used 600 fewer distinct words, and used shorter words, than LNs. These measures were unaffected by the race of the interviewer.
The grade level and reading ease of the transcripts were evaluated using a built-in function of Microsoft Word. The utterances of ANs were three grade levels lower than those of LNs, and the reading ease of the ANs' transcripts was eleven percent better, reflecting the relative simplicity of what they said. These metrics were affected by interviewer race. When both of the interview pair were white, the grade level and reading ease of ANs was almost the same as for LNs. When the primary interviewer was black, the grade level was five lower than that of the LNs, and the reading ease was 16 to 17 percentage points higher. [Stenner 2005, p. 215]
There were also differences between the words used by ANs and LNs. Terms were classified according to their theme using software. The greatest contrast involved terms referring to inclusion and exclusion. AN subjects speaking to a white interviewer were less likely than LNs to use terms indicating social exclusion, isolation, or disconnection. AN subjects speaking to a black interviewer were more likely than LNs to use such terms. AN subjects were no more inclined than LNs to use aggressive words if their interview partner was white. If the interview partner was black, however, many more aggressive terms were used by the ANs. ANs used fewer critical and complaining terms than LNs if both interviewers were white. If the interview partner were black, however, many more critical and complaining words were used by the ANs. [Stenner 2005, p. 220]
The team evaluated the interview transcripts for three domains of intolerance, racial, political, and moral. The transcripts were coded according to standards set by Stenner. The coder identified occurrences of specific events which Stenner defined, but the coder worked at arm’s length from Stenner, and was not told what her client’s theoretical position was, nor the hypothesis she was testing. It is not clear whether or not the coder was aware of the racial characteristics of the interviewers. These coded events were then correlated with authoritarianism/libertarianism (AN/LN) across the set of subjects. (An example of a coded event was “Subject seems to be racist.”) [Stenner 2005, p. 239]
| AN choice | LN choice |
|---|---|
| that a child obeys his parents | that he is responsible for his own actions |
| that he has good manners | that he has good sense and sound judgment |
| that he is neat and clean | that he is interested in how and why things happen |
| that he has respect for his elders | that he thinks for himself |
| that he follows the rules | that he follows his own conscience |
Each line of this table was presented as a forced choice to the subject (“Please choose one.”) The items were presented in random order. Half of the authoritarianism score was the number of AN answers the subject chose. The other half of the authoritarianism score was calculated by having the subject rank-order both the AN and the LN choices. The precise procedure and the reason for it are both unclear to me.
The following table excerpt shows results for one of the coded events related to racial intolerance.
| CODED EVENT | Total | Both white | Black interviewer | Black partner |
|---|---|---|---|---|
| S seems to be racist | .34** | .50** | -.14** | - |
The first column describes the coded event. The “Total” column provides the correlation between the occurrence of the event and the AN/LN status of the subjects. Since LNs had a lower score and ANs a higher score, a positive correlation means that the event happened more commonly with AN subjects than with LN subjects.
Asterisks indicate the “p” value or statistical significance of the correlation. No asterisks appear if there was more than a ten percent chance that the calculated correlation is the result of random variation. One asterisk means that there is more than a five percent but less than a ten percent chance that the correlation was the result of randomness. Two asterisks appear if that chance is less than five percent. More stars means stronger evidence.
The “Both white” column shows the correlations for those interviews in which both interviewers were white. The “Black interviewer” column, for those in which the primary interviewer was black. The “Black partner” column, for those in which the interview partner, the observer, was black. The hyphen in the cell means that the event (“S seems to be racist”) wasn’t found in interviews with a black partner often enough to yield a meaningful correlation estimate.
In this example there is a fairly large correlation, .34, between AN status and “seems to be racist.” (This is in the opinion of the coder, based upon the standards supplied by Stenner.) Besides being fairly large, the correlation is statistically significant (ss) at the p = .05 level. (There is less than a five percent chance that a correlation value as large as .34 would have occurred if in fact there was no correlation.) In the interviews in which both interviewers were white, the correlation was higher, .50, still ss at .05. AN subjects were more likely to seem racist when there were no blacks in the room. This contrasts with the results when the primary interviewer, the person with whom the subject was talking, was black. In this case, the correlation is -.14 and ss at the .05 level, meaning that AN subjects were less likely than LN subjects to appear racist when speaking to a black interviewer.
In sum: AN subjects were much more likely than LN subjects to seem racist. This was particularly so when both interviewers were white. When the interviewer was black, however, ANs were less likely to seem racist than were LNs. [Stenner 2005, p. 241]
The results in the three tables that follow report on evidence of intolerance in the three areas of race, politics, and morality. Positive correlation means the events were more frequent with high-AN subjects. Negative correlation means the events were more common with low-AN subjects.
| CODED EVENT | Total | Both white | Black interviewer | Black partner |
|---|---|---|---|---|
| S seems to be racist | .34** | .50** | -.14** | - |
| S uses “us” and “them” in discussing race | .27** | .54** | .01* | .41 |
| S defensive, “some of my best friends black” | .20* | .40** | -.04** | -.06 |
| S disapproves of interracial contact | .19** | .33* | .09* | .04 |
| S welcomes interracial contact | -.32** | - | - | - |
| S concerned race relations getting worse | .18 | -.67** | .08** when either is black | |
| S dismayed about racial tension | -.10 | -.35** | .05** when either is black | |
| S says blacks have good deal, selves to blame | .45** | .76** | .00** | .51 |
| S concedes discrimination, system to blame | -.30* | -.92** | -.05** | -.47 |
| S claims disadvantaged for being white | .11 | .43** | -.17** | .01 |
| S has negative stereotypes of blacks: violent, lazy | .23** | .10* | .77** | - |
AN subjects expressed racial intolerance in the presence only of whites. This indicates that the ANs know and are willing to respect the norms of political correctness in the presence of black people, but have not accepted or internalized them.
The last item in the table, expression of negative stereotypes, contrasts with all the other items. Negative stereotypes of blacks were much more common when an AN subject was speaking to a black primary interviewer. Stenner believes that talking to a black person subconsciously activated these stereotypes. This view is further supported (according to Stenner’s analysis of the texts) by the circumstance that these subjects seamlessly slipped from talking about blacks to talking about crime, violence, drugs, etc. [Stenner 2005, p. 241]
Analysis of the racial events shows that ANs as a group were disposed to say different things in the presence of whites than in the presence of blacks. Does this imply that these people are two-faced? It hardly seems fair to come to this conclusion when each subject had just one interview, and had no opportunity to display two-faced behavior. Yet I think the answer is probably “Yes, on average.” The numbers and the experimental design almost preclude alternative explanations. (There is a possibility that something about the interviewers besides their melatonin affected the speech of the subjects.) The evidence supports the prediction that many of these high-AN subjects would speak differently to whites than to blacks.
| CODED EVENT | Total | Both white | Black interviewer | Black partner |
|---|---|---|---|---|
| S recognizes flaws of United States/Americans | -.50** | - | - | - |
| S not ethnocentric, admires other countries | -.41** | - | - | - |
| S says “love it or leave it” re United States | .12** | - | - | - |
| S very emotional re American flag | .17** | -.00 | .23** when either is black | |
| S does not follow politics | .25* | - | - | - |
| S follows politics | -.09 | -.29* | - | -.31** |
| S thinks politicians all about money/power | .04 | .53** | .02 | .16 |
| S thinks politicians mean well, unappreciated | -.21* | -.40** | .00** | - |
| S sympathetic to super-patriot/militia movement | .14 | .65** | -.07** when either is black | |
| S no sympathy for patriot/militia issues | -.25* | -.65** | -.06* when either is black | |
LNs were about half again more likely than ANs to accept that there are flaws among Americans and with the country. ANs were prone to wax emotional over the flag, but, continuing the pattern, only in the presence of a black interviewer or observer. This seems to illustrate the principle that group members associate more strongly with group symbols when ‘under threat.’ (The threat here is merely the presence of a black person in the discussion.)
ANs were more likely to have derogatory opinions of politicians and at the same time be uninterested in the political process. With white interviewers, ANs were half again more likely than LNs to opine that politicians are motivated by money and power. ANs were almost two-thirds more likely to express sympathy with ‘super-patriot’ or militia groups, but only when talking with a white interview team. [Stenner 2005, p. 250]
Themes of the ‘super-patriot’ ideology include distrust of the federal government, white supremacism, and the threat posed by the increasing influence of non-whites. The reluctance of ANs to express these sympathies to black interviewers is consistent with the notion that these ANs were familiar with the super-patriot ideology. The ANs were highly likely to express these sympathies in an all-white interview. One of the subjects expressed his views:
“I agree with some of ‘em...Now you talkin’, say, a militia, uh, I don’t, I don’t disagree with ‘em. But a militia’s just like IBM or Duke University or somethin’. You got bad ones in all of ‘em. Now these militias, course the government’s called militias bad, I understand why they are, ‘cause some of ‘em, they’re crazy to start with! But, if they’re doin’ right, I think they’re OK really. I say, I just, I have to, I have to side with them people. I don’t believe they did nothin’ that bad for the army and the FBI to come up there. I just, they weren’t done right, to my opinion.” [Stenner 2005, p. 256]
Stenner commented:
“...through to the present, authoritarians have been understood as individuals who possess a rather tenuous commitment to democratic processes, and who are extremely uncomfortable with, even uncomprehending of, the core components of a democratic system. If authoritarianism is concerned with sacrificing individual freedom and diversity to group authority and conformity, and if authoritarians have the motives and capacities I have ascribed to them, it is easy to see that the central elements of democracy are not just anathema, but actually insensible to authoritarians. Disagreement, dissent, and disobedience; determination of the ‘common good’ by debate and negotiation between partisans of competing worldviews; none of this is comprehensible, let alone palatable, from the authoritarian perspective.” [Stenner 2005, p. 250]
| CODED EVENT | Total | Both white | Black interviewer | Black partner |
|---|---|---|---|---|
| 1. S worries about nation’s loss of morals, moral decay | .52** | .80** | .71 | .08* |
| 2. S ridicules idea of nation’s moral decay | -.47** | -.99** | -.11** when either is black | |
| 3. S says Americans have strong shared values | -.01 | -.35** | .18** when either is black | |
| 4. S says govt. to reverse nation’s moral decline | .39** | - | - | - |
| 5. S wants schools to reverse moral decline | .20* | .57** | .04** when either is black | |
| 6. S says family breakdown source of society’s ills | .29** | - | - | - |
| 7. S says kids don’t respect authority, no discipline | .28** | .51** | - | .08** |
| 8. S says TV corrupts kids, bad influence | .17** | - | - | - |
| 9. S very concerned with crime and safety | .38** | .20 | .80* | .29 |
| 10. S turns everything into crime issue | .26** | -.01 | .72** | .04 |
Issues of morality and crime were the ones that most aroused ANs. “...fears regarding morality and crime, claims about the critical need to reestablish some normative order, and elaboration of plans for accomplishing this reversal occupied the bulk of authoritarians’ ‘psychic space’, consuming a vastly disproportionate share of their time and energy in these discussions.” [Stenner 2005, p. 256]
As with race issues, ANs spoke differently when one or both interviewers were black. Especially with a black interview observer ANs more strongly claimed that Americans have strong shared values, they spoke less about their plans to remedy the moral rot, and were less disposed to criticize American children. ANs were almost indistinguishable from LNs when blacks were present. Stenner attributes this to their sense that they were among “others.” She suggested an alternative hypothesis, that the concern over moral decay is in actuality a sense of the corrupting influence of blacks in the society. She finds the “us/them” explanation more likely, but not provable from her data. [Stenner 2005, p. 258-9]
ANs (event 1.) were .80 more likely than LNs to be worried that morals in our country were in decline. One subject offered his view that standards of public morality had “evaporated.” LNs in contrast (event 2.) were almost sure to deride this idea. ANs seemed to be much inclined to believe in a fixed morality transcendent of human needs or preferences, which is consistent with their common embrace of conservative Christianity. The problem to the ANs was not what the proper rules might be, but how to enforce them. [Stenner 2005, p. 261]
ANs (event 4.) often had a plan for improving the morals of the nation. These plans usually involved using the coercive abilities of government, including the schools. One subject stated that “we definitely can’t do a totalitarian state,” but, “if the people can’t maintain civility ourselves then somebody has to institute it.” [Stenner 2005, p. 263] ANs were more than half again more likely to assert that the schools must play a big role in halting the rot (event 5.). One subject proposed that “there has to be some kind of required course that’s going to teach us...to have higher moral vales than we have.” [Stenner 2005, p. 263-4] This won’t apparently be a discussion session or sensitivity training. ANs often endorsed the use of force to “teach” pupils the limits of acceptable behavior. Stenner comments that “it does seem clear from their various comments that the authoritarians, far more than the libertarian subjects, believed that fear of retribution, external constraints, and physical force were appropriate means, indeed the ideal means, for regulating social behavior.” [Stenner 2005, p. 264]
Event 6. shows that ANs were inclined to believe that family breakdown was an important factor in the moral breakdown. (Moral decline, TV violence, and music lyrics were big GOP talking points around this time.) Their model family was a patriarchal one.
In the racism results, racist stereotypes seemed to break free while ANs were talking about other topics to a black interviewer. The last two codes in the morality/crime results also show an enormous increase in talk about crime when the interviewer was black. Again, Stenner judges from context that the connection between blackness and crime must be deeply felt by these ANs. “Clearly, then, as others have demonstrated before me, for many white Americans crime has a black face.” [Stenner 2005, p. 260]
Stenner had several comments about these subjects:
“...such dominance of strong dispositions [to associate blackness with crime] over social learning [that social norms forbid this] might also suggest, inversely, that extreme authoritarians would harbor negative racial stereotypes even if socialized in a perfectly enlightened environment of tolerance.” [Stenner 2005, p. 261] (Stenner remarked elsewhere that the oft-noted negative correlation of authoritarianism with education may reflect not only the ameliorative effects of knowledge, but may reflect a lack of interest and aptitude for education which is characteristic of authoritarianism. [Stenner 2005, p. 198])
“...authoritarians clearly found it difficult to reconcile political disagreement and debate with being a patriotic American.” [Stenner 2005, p. 267]
“...authoritarians are inclined to ‘reification’ of social norms and processes [picture the Ten Commandments engraved on stone],...they tend to think of social reality as ‘encompassing a superodinate normative dimension, an external locus where events are determined, where moral authority resides, and to which men must adapt themselves....Thus institutions, rules, customs and norms are treated not as human products...but rather as if installed by some external force, ‘superordinate and infused with transcendental authority.’....” [Stenner 2005, p. 267]
ANs were correspondingly alarmed by departures from absolutism, and looked to authorities to intervene. As I read her comments above on reification of norms and the related ideas, it called to my mind that this is precisely what most of the Christian churches have been preaching essentially since their beginning. (See “God On Our Side.”)
Stenner asserts that several flavors of intolerance (racial, political, and moral) roughly covary, that they are all motivated by an impulse toward sameness, and that they are promoted by normative threat. A normative threat is a condition that activates the predisposition toward authoritarianism, and includes [Stenner 2005, p. 269]
Elsewhere though, she says “I have isolated two critical components of normative threat: failed political leadership and, especially, great variance in public opinion.” [Stenner 2005, p. 283]
Normative threat, by this theory, affects both authoritarians and libertarians, the name she applies to those with low authoritarianism predispositions. In authoritarians normative threat causes increased expression of intolerance and increased calls for government intervention. In libertarians, it increases the endorsement of tolerance and motivates defense against suppression.
I am persuaded that existing theories of authoritarianism are muddled and would benefit by more precision. I’m also persuaded that it’s a dynamic process. The conditions that alarm authoritarians increase intolerance and the desire to suppress diversity and non-compliance with the authoritarians’ codes. I’m not, however, persuaded by Stenner’s argument that she has identified the true authoritarianism in child-rearing values, nor that her concept of normative threat is clearly defined or demonstrated by the work reported in this book.
Stenner has provided strong evidence of the dynamic nature of these processes and strong evidence that intolerance tends to vary regardless of the domain (racial, political, or moral). For example, Stenner experimented with a group of political science majors in 1995, in her CRE95 study. [Stenner 2005, p. 279-81] Subjects were given one of two stories to read, disguised as news reports. One story reported that NASA expected extra-terrestrial aliens to arrive soon. This story commented that the diversity that would arrive with the aliens would exceed any that we had so far experienced. The other story reported that NASA had concluded there are no extra-terrestrials. This story commented that, since there will be no space visitors, the diversity we are familiar with is as much as there will ever be.
The reading of one or other of these stories was followed by an assessment of the subjects’ expressed racial intolerance. The stories had quite a big effect. Among those reading the ‘no aliens’ story, racial intolerance increased with level of AN, yielding a statistically significant r of .76. Among those reading the ‘aliens are coming’ story, however, racial intolerance was substantially lower among high-AN subjects than among LNs. The statistically significant r was -.47. In another comparison within the same study, the ‘no aliens’ story was associated with a strong relation between high AN and high punitiveness, with a statistically significant r of .74. The ‘aliens coming’ story resulted in no difference between ANs and LNs in their punitiveness. In both these comparisons the supposed coming of aliens lessened the intolerance of high ANs.
I find this result startling. It certainly demonstrates that expressed tolerance can be manipulated in surprising ways. However, is the ‘aliens are coming’ story a normative threat? It doesn’t fit obviously into any of the categories of normative threat that Stenner provided, though I might put it into the category of “questioning or questionability of norms.” In that case, the ‘aliens coming’ story would be an enormous normative threat, and Stenner’s hypotheses would predict a result opposite to the result observed. Stenner believes that the ‘aliens coming’ scenario was seen as a non-normative threat, and the non-normative threat expanded the definition of the in-group for the ANs.
In another investigation Stenner used a different method to present threat or assurance. She participated in a Multi-Investigator Study (MIS99), in which subjects were telephoned randomly. The time allocated to her questions was limited, and instead of a 500-word story, subjects were read an 80-word synopsis of a fictitious news story. Three of the synopses were normatively threatening and three were normatively reassuring. Following the synopsis, the authoritarian predisposition (AN) of the subjects was tested. [Stenner 2005, p. 285 et cf.]
With this data in hand, Stenner was able to to see whether the authoritarian disposition measure was affected by normative threat and reassurance. It was. The “alpha” of the child-rearing questions, a measure of their internal consistency, was twenty percent higher among subjects who had received threat compared with those who had received reassurance from their synopsis. This result supports Stenner’s hypotheses, in that normative threat is said to activate or strengthen authoritarianism. Again, the result certainly supports the contention that subtle manipulations can change this process.
I question three points here. First, it’s not obvious to me how and why three of the synopses are threatening and three reassuring. Second, Stenner repeatedly specifies that when she says “authoritarianism,” she means a disposition, and a disposition, I would think, would not change because of the manipulations of the MIS99 study. Third, Stenner used at least three different operational definitions of authoritarianism in three studies. I believe it is a matter to be demonstrated that these measure the same thing.
| 1a. Belief diversity | “The story was that American public opinion on a wide range of issues—from how children should be raised to how the political system should be run—is becoming increasingly divided. The American people are starting to disagree about more things, and disagree more strongly. It seems that public consensus is deteriorating. And worst of all, this disunity in American society looks certain to worsen in the future, with more and more disagreement about what is right and wrong.” |
| 1b. Belief consensus | “The story was that American public opinion on a wide range of issues—from how children should be raised to how the political system should be run—is becoming increasingly united. The American people are starting to agree about more things, and agree much more strongly. It seems that public consensus is growing. And best of all, this consensus in American society looks certain to improve in the future, with more and more agreement about what is right and wrong.” |
| 2a. Stable diversity | “The story was that America is going through a period of steady social stability. Advances in science and technology have slowed down dramatically, and we now see stabilization in our political system, our jobs, and our families. The article was not suggesting that American society is pulling together, it was suggesting that while we might have different goals and values, we have a stable society that will endure as a constant as we ease into the next century.” |
| 2b. Changing together | “The story was that America is going through a period of rapid social change. Advances in science and technology have brought about enormous changes in our political system, our jobs, and our families. The article was not suggesting that American society is falling apart. Rather, it was suggesting that we’re moving forward at a very fast pace, finding new ways to meet our common goals and values as we speed into the next century.” |
| 3a. Bad leadership | “The story was that American presidents have generally not lived up to our expectations. With just a few exceptions, both our Republican and Democratic presidents have been remarkably lacking in strength, vision, and principle. Our presidents, from both political parties, have generally been unworthy of the trust we placed in them, and have not been real leaders in any real sense of the word. And worse still, with no electoral reforms, we’re bound to confront even poorer-quality candidates in the future.” |
| 3b. Good leadership | “The story was that American presidents have generally lived up to our expectations. With just a few exceptions, both our Republican and Democratic presidents have shown great strength, vision, and principle. Our presidents, from both political parties, have generally been worthy of the trust we placed in them, and have been leaders in every sense of the word. And better yet, electoral reforms mean we can look forward to even better-quality candidates in the future.” |
Reading these synopses raises further questions to me. Is it required that the story change the subject’s belief state in some way? Must the subject believe the story, or is the fact that someone made those claims enough to activate the disposition? It seems to me that answers to these questions are needed before a safe prediction can be made about how a subject will perceive such stimuli.
Can you tell which of the six synopses is threatening? I’ve read her explanations and rationale for them, but I can only tell which is threatening by remembering that the “a” questions are threats, the “b” questions are assurances.
Dr. Stenner appears to be staking a lot on a pre-hoc assessment of how these stories will be heard. She certainly has a more-informed opinion about that than I, and she may well be right, but these apparent leaps of interpretation cause me to question her concept of “normative threat.” I wonder what it means, and I question whether it can bear the weight she’s put upon it.
Three definitions of “authoritarianism” were used in the three studies I’ve mentioned. They’re shown for comparison in the following tables.
| AN answer | LN answer |
|---|---|
| that a child obeys his parents | that he is responsible for his own actions |
| that he has good manners | that he has good sense and sound judgment |
| that he is neat and clean | that he is interested in how and why things happen |
| that he has respect for his elders | that he thinks for himself |
| that he follows the rules | that he follows his own conscience |
| AN answer | LN answer |
|---|---|
| that a child obey his parents | that he is responsible for his own actions |
| that a child has respect for his elders | that he thinks for himself |
| that a child follows the rules | that he follows his own conscience |
| AN answer | LN answer |
|---|---|
| obey | question |
| rules | progress |
| obedience | curiosity |
The DCS97 survey questions are asking which is more important to give to a child. To me that suggests a final goal of parenting. The MIS99 questions, on the other hand, could refer to a specific situation. The CRE95 questions were designed for students who presumably have recently left the oversight of parents, and who might have no firm ideas about raising children. They can be interpreted as referring to the student himself. That is, they may be asking what the student prefers for himself. This would then be a measure of the student’s conformity rather than of child-rearing values.
Stenner claims that AN measures a need for unity and uniformity, but at another time it is a disposition to authoritarianism. To me it is clearly something rather different. The child-rearing questions are to my eyes mostly about compliance vs judgment, rules vs decisions, although other things are certainly involved. I might describe it as “locus of control” or “locus of authority.” This conceptualization ties more closely with Stenner’s own concept of “normative threat,” I think, than does the concept of “authoritarianism.”
I’d like to understand in what way normative threat is “normative.”
The theory as I understand it is that if you believe that external authority is the determinant of acceptable behavior, this will be reflected in your child-raising beliefs. You may well be inclined to belong to a group which supports this belief, and which coaches you through a life lived according to these beliefs. You may believe that many areas of life must be controlled and not up to the individual's discretion. When you become aware of others acting in ways that violate your norms (which to you are the only norms), you interpret it as the result of inadequate oversight. “Human nature” tells you that the misbehavior will get worse while oversight is inadequate. The normative threat, then, is to order writ large. The threat is “normative” in that violation of norms is threatening everything else, according to your worldview.
If my understanding is correct, child-rearing values are far from a disposition, but are in fact an effect alongside intolerance. They seem to be a good predictor of intolerance. But those particular child-rearing values that associate with intolerance and coercion are also associated with a set of beliefs about human behavior that has long been recognized. (“Is human nature fundamentally good or bad?”) What facilitates or causes those beliefs isn’t answered without being asked.
Also, if my understanding is correct, authoritarianism is far from a desire for sameness, but rather is a specific desire that the whole world adhere to your norms, more reminiscent of Torquemada than of Mister Rogers.
It’s easy to accept that child-raising goals could be important. My wife and I were gratuitously offered child-raising advice by strangers at least twice carrying our new-born through the hospital corridors. Attitudes about child-raising may be an important but neglected aspect of conservative ideology. Child-raising isn’t often regarded as a political act.
Child-rearing attitudes are not only an effect of learnings and motivations, but also a cause of diverse ideological and behavioral effects. (A person known to me once attended a shower for a religiously conservative mother. Among many other things said over brunch, babies were described as soldiers-of-God-in-training.)
I find it quite startling that these few questions about child rearing reveal so much about a person, and that they seem to accurately and cleanly measure a phenomenon closely related to RWA/SDO. I agree with Stenner's view that investigation of the mechanism whereby authoritarianism is dynamic is important. Nevertheless I question child-rearing-as-authoritarianism and I am eager to understand more about the threat that she is studying. I appreciate what I’ve learned from her.
Next: Under the Hood
More information:
[Stenner 2005] “The Authoritarian Dynamic,” Karen Stenner, Cambridge University Press, 2005.
© 2021, Ross A. Hangartner