For Larry, on the other hand, the value for getting ahead does not exclude the possibility of various other kinds of people getting ahead, for he seems to be thinking in terms of an
expanding
economy in which working men can have a strong role (Item 68) and in which depressions are unnecessary (Item 5).
Adorno-T-Authoritarian-Personality-Harper-Bros-1950
62
57 . 48 68 ? 52
226 . 52
c. Groups taking Fora 45:c
Extension Testing Class Women
San Quentin Men 110 Psychiatric Clinic women
Psychiatric Clinic Men
W orking-Class Men and
Women?
D. Groups taking Fora 40:c
George Washington University Women California Service Club Men Middle-Class Men
Middle-Class women
W orking-Class Men
WOrking-Class WOmen
Los Angeles Men
Los Angeles Women
Mean ra
E. Groups taking Foras 40 and 45: Employment Service Men Veterans
(Form 40)
(Form 45) Maritime School Men (Form 40) (Form 45)
Mean ra
Mean r for all groups taking Forms 40 and 45
59 . 60
. 14d 71 ? 53
50 ? 55d 50 ? 75
340 . 51
132 . 48 63 . 64 69 ? 67
154 . 76 61 . "74 53 ? 86
117 . 61 130 ? 52
779 ? 66
55 ? ooe
51 ? 49 164 . 4le 179 ? 38
449 . 47 1568 ? 57
Mean ra
~n obtaining the over-all means, the individual group means were not weighted by N, and Z,. was not used.
bit will be recalled that in Form 60 the E scale contained 4 A-S items, there being no separate A-S scale.
CpEC is correlated with EA in groups taking Form 40, with total EA+B scale in groups taking Form 45.
dFor the PEC scale in the Psychiatric Clinic groUPs, the number of -en was 45, the number of men 29~ due to a substitution of forms.
erhese correlations are based on the EA scores of subjects taking Form 45 . as well as those taking Form 40.
E:PEC
A-S:PEC
? THE AUTHORITARIAN PERSONALITY
negligible. The same holds for class differences: in the Working-Class Men and Women the correlations are very high (. 74 and . 86), but in other groups which draw heavily upon the working class, notably San Quentin and the Maritime School, the correlations are very low (. 14 and . 4). The reasons for the variations in the size of r seem to. lie more in the specific nature of the group than in its sex or class status. It is interesting in this connection that two groups in which the E-PEC correlation was very high, the Working Class Men (. 74) and Women (. 86) also had the two lowest PEC means (see Table 8(V) ). W e may hypothesize that the E-PEC correlation will be highest when, other things such as the S. D. being equal, the group contains a sizeable minority of strong liberals. Judging from some of the other groups, the number of strong conservatives has less influence on the correlation. W e shall return to this question shortly.
The general level of the E-PEC correlations demonstrates that ethnocen- trism and conservatism, as measured by the present scales, are significantly but imperfectly related. 7 In everyday terms, we may say that conservatives are, on the average, significantly more ethnocentric than liberals are. The more conservative an individual is, the greater the likelihood that he is ethno- centric-but this is a probability and not a certainty. Since the existence of an affinity between these ideological patterns has often been observed previ- ously, the present correlations are perhaps less a startling discovery than an indication of the validity of the scales. To those who have been unaware of the E-PEC relationship, the significance of the correlations must be stressed. To those who tend to equate conservatism and ethnocentrism as psycho- logical trends in the individual, it must be pointed out that the correlations are far from perfect. Even with a much more reliable measure of PEC, the correlation with E could hardly average over . 70-a value inadequate for predictive purposes. It becomes necessary, then, to understand what pro- duces the close association between these ideological patterns in the indi- vidual, as well as what systematic factors-apart from ignorance or misinfor- mation-make the E-PEC correlation less than I . o.
A theoretical basis for the close tie between conservatism and ethnocen- trism is suggested by certain similarities in their major underlying trends. Support of the prevailing politico-economic ideology and authority is, ap- parently, often a part of the generalized ethnocentric tendency to submit to accepted authority in all areas of social life. Similarly, ethnocentric rejection of outgroups is expressed in the politico-economic sphere by resistance to social change and by the tendency to subsume progressive political ideologies under the general heading of "foreign" outgroups and ideas (threats to in- group authority). The interconnection between the two ideologies and the difficulty of separating them even for purposes of study are revealed by a
? 7 These results are, on the whole, consistent with those of the other studies mentioned at the start of this chapter.
? POLITICO-ECO~OMIC IDEOLOGY AND GROUP MEMBERSHIPS I 8I
number of scale items. For example: "America may not be perfect, but the American Way has brought us about as close as human beings can get to a perfect society. " To support this idea is, it would seem, to express both politico-economic conservatism and the ingroup idealization so character- istic of ethnocentrism. The item, "The worst danger to real Americanism during the last 50 years has come from foreign ideas and agitators," is another example of politicalized ethnocentrism: again we find moral stereotypy and externalization of blame for social problems onto the threatening outgroup.
There are also theoretical reasons for expecting a relationship between liberalism and anti-ethnocentrism. Both tend to involve a critical attitude toward prevailing authorities and traditions. The identification with the masses (workers, "the common man," "the weak and downtrodden") so often a central theme in left-wing political ideology, finds expression also in opposition to ethnocentrism and outgroup suppression. Indeed, the forma- tion of leftist political views in youth often begins with a sense of the injus- tice of anti-Semitism or anti-Negroism; when sympathetic imagery of sub- ordinate groups is extended to include the working class, the transition from "group relations" to "politico-economic" ideology has begun. The further development of liberal-radical views is ordinarily based on imagery and atti- tudes identical to those underlying anti-ethnocentric ideology: opposition to hierarchy and to dominance-submission, removal of class and group barriers, emphasis on equalitarian interaction, and so on.
We have also to consider the "correlation-lowering" factors. Why, in view of the theoretical argument above, is the E-PEC correlation not higher? The present data suggest, but are not adequate to test, several hypotheses. The correlation charts (scattergrams) for each sample reveal that extreme liberals (low scorers on PEC) are for the most part low as well on E. But the "middles" on PEC are extremely diversified with respect to standing on E. It is possible that the group which is low on E but middle on PEC consists largely of the "politically pacifistic" liberals discussed previously. Practically none of the subjects were low on PEC and high on E (ethnocentric liberals); such il1dividuals would, however, be well worth intensive study.
The high scorers on PEC are more variable on E than are the low scorers. While most of those high on PEC are also high onE, a considerable number are middle and a small but consistent percentage low onE (nonethnocentric conservatives). In other words, strong political liberalism is a pretty good indicator of anti-ethnocentrism, but political conservatism is less consistently related to ethnocentrism.
In attempting to explain the variability of conservatives with respect to ethnocentrism, we are reminded of the distinction between "genuine" and "pseudo-" previously drawn with respect to patriotism and traditional- conventional values (Chapter IV). One can be politically conservative, just as one can be patriotic (in the sense of firm attachment to American culture
? THE AUTHORITARIAN PERSONALITY
and tradition), without being ethnocentric. vVe should like to use the term "genuine conservative" to refer to the individual with this broad pattern of thought. He is "genuine" because, whatever the merits of his political views, he is seriously concerned with fostering what is most vital in the American democratic tradition. He believes, for example, in the crucial importance of the profit motive and in the necessity of economic insecurity; but he wants the best man to win no matter what his social background. He is resistant to social change, but he can be seriously critical of the national and political ingroups and-what is more important-he is relatively free of the rigidity and deep-lying hostility characteristic of ethnocentrism.
The ethnocentric conservative is the pseudoconservative, for he betrays in his ethnocentrism a tendency antithetical to democratic values and tradi- tion. He is the E-PEC "correlation raiser" because, as discussed above, his politico-economic views are based on the same underlying trends-submis- sion to authority, unconscious handling of hostility toward authority by means of displacement and projection onto outgroups, and so on-as his ethnocentrism. It is indeed paradoxical that the greatest psychological poten- tial for antidemocratic change should come from those who claim to repre- sent democratic tradition. For the pseudoconservatives are the pseudo- democrats, and their needs dispose them to the use of force and oppression in order to protect a mythical "Americanism" which bears no resemblance to what is most vital in American history. .
An additional hypothesis may be proposed regarding individuals high on E but middle on PEC. These may well be pseudoconservatives who have kept up with changes in the actual politico-economic situation by making changes in traditional (individualistic) conservative ideology. They empha- size competitiveness as a value, yet they support the concentration of eco- nomic power in big business-the greatest single threat at present to the individual competing businessman. They emphasize economic mobility and the "Horatio Alger" myth, yet they support numerous forms of discrimina- tion that put severe limitations on the mobility of large sections of the popu- lation. They may also believe in extending the economic functions of government, not for humanitarian reasons but as a means of limiting the power of labor and other groups.
This is not merely a "modern conservatism. " It is, rather, a totally new direction: away from individualism and equality of opportunity, and teward a rigidly stratified society in which there is a minimum of economic mobility and in which the "right" groups are in power, the outgroups subordinate. Perhaps the term "reactionary" fits this ideology best. Ultimately it is fascism. While certainly not a necessary sequel to laissez-faire conservatism, it can be regarded as a possible (and not uncommon) distortion of con- servatism-a distortion which retains certain surface similarities but which changes the basic structure into the antithesis of the original. Since most
? No.
1. 5. 8.
13. 15. 22. 27. 36. 44. 52. 61. 63. 68. 71. 76. 78.
I t e m
(Value of dollar) (Depressions)
(Charity)
(Businessmen, artists) (Middle of the road) (Political candidate) (Rebellious ideas)
(Gov' t. responsibility) (Socialized medicine) (Taxes, corporations) (Economic security) (Gov' t. interference)? (Unions stronger)
(Ford, Morgan)
(Gov' t. activity) (Ability will tell)
Groupa Mack L a r r y Mean
6 7 6. 10 5 1 3,33 3 7 5. 46 1 1 2. 29 7 5 4. 35 7 7 6. 38 5 6 3. 86 2 1 3. 22 2 6 2. 38 2 3 3. 66 6 6 3. 75 5 1 4. 01 6 2 4. 58 7 6 5. 30 2 2 3. 32 7 6 4. 74
Group a D. P .
1. 16 2. 76 1. 48 1. 70 2. 90 0. 32 2. 84 3. 01 1. 69 2. 29 2. 68 2. 39 2. 30 2. 00 2. 76 1. 99
POLITICO-ECONOMIC IDEOLOGY AND GROUP MEMBERSHIPS 183
Americans are "middle on PEC," it becomes crucial to understand the psy- chological dispositions which help to determine new ideological directions in the individual.
The above distinction regarding ideological patterns within the political left and right are presented as hypotheses to help explain the scale results. 8 If these hypotheses are not borne out, others will be needed. For it is clear that political ideologies do not fall neatly along a simple liberalism-conserva- tism dimension; that the relation between ethnocentrism and "conservatism" is extremely complex; and that the individual's receptivity to political ideol- ogy, as to "group relations" ideology, is based to a large extent on deep- lying personality trends.
D. V ALIDATlON BY CASE STUDIES: THE RESPONSES OF MACK AND LARRY ON THE PEC SCALE
In an attempt to judge the validity of the PEC scale we may here, as in Chapters III and IV, compare the scale responses of Mack and Larry and consider them in relation to material from their interviews (Chapter II). The PEC-scale scores of these two subjects, the group mean, and the D. P. for each of the 16 PEC items included in Form 78 are shown in Table II (V).
TABLE 11 (V)
RESPONSES OF MACK AND LARRY ON THE PEC SCALE
Mean per item
aThe group means and D. P. 's are based on all four groups taking Form 78.
8 Further hypotheses, plus supporting evidence, are presented in Chapter XVII, which deals with the interview material.
4. 56 4. 19 4. 17
2. 14
? THE AUTHORITARIAN PERSONALITY
The two men are much more similar in their PEC-scale scores than was the case with their A-S and E scores. Larry's mean score, 4? I9, is at the group mean and Mack's, 4? 56, is not very far above. When the group of Public Speaking Men to which these subjects belonged is considered by itself, Mack is just inside the high quartile, Larry is just below it. On I I of the I6 items the scores of the two men do not differ by more than one point. Attention to the individual items, however, shows that the similarities are confined to certain areas of politico-economic ideology; in other areas there are sharp differences.
Mack and Larry are most similar in their consistent support of general conventional-conservative values. Both agree, usually rather strongly, with Items I (Value of a dollar), I 5 (Middle of the road), 22 (Political candidate), 27 (Rebellious ideas), ? I (Ford, Morgan), and 78 (Ability will tell). This seems to be in keeping with the interviews, in which both men expressed the usual conservative criticisms of the New Deal. Both men, it appears, accept the view that a man's getting ahead depends most of all upon his living according to the values of thrift, determination, work, honesty, conformity, and the like.
Examination of the scores on other items, however, indicates that Mack and Larry differ with respect to the context in which their conservative values occur. For Mack the context appears to be one of upward social mobility on a class-ingroup basis, for Larry it appears to be one of nineteenth- century liberalism. This seems to be expressed in their wide disagreement on Items 5 (Depression) and 68 (Unions stronger). Mack's belief that de- pressions are "only natural" can be interpreted as an expression of the broader idea that, in the nature of things, the benefits to be had in our society are insufficient to go around and that it is no more than proper that the major portion of them should go to the "right people," that is, to an ingroup. This ingroup does not seem to include organized labor (Item 68) nor the various minority groups which he rejected in his responses on the E scale. This would appear to be another manifestation of Mack's tendency, so marked in his interview, to make rigid ingroup-outgroup distinctions in his thinking about politico-economic matters as well as about social relations generally.
For Larry, on the other hand, the value for getting ahead does not exclude the possibility of various other kinds of people getting ahead, for he seems to be thinking in terms of an expanding economy in which working men can have a strong role (Item 68) and in which depressions are unnecessary (Item 5). The absence of any ingroup-outgroup distinction, and optimism with respect to the possibilities of economic abundance were outstanding features of Larry's interview.
Neither man shows the usual conservative opposition to the government's participation in the economic life of the nation: for Items 36 (Government responsibility), 44 (Socialized medicine), 52 (Taxes, corporations), 63 (Gov-
? POLITICO-ECOXOJ\IIC IDEOLOGY AND GROUP MEMBERSHIPS 185
ernment interference), and '76 (Government activity) the two men have the same low mean score of 2. 6o. Their reasons, however, seem to be different. When the responses to the total scale are considered, it appears that Mack deviates from true, laissez-faire conservatism by taking a stand further to the political right. The fact that he rejects labor unions and believes in the inevitability of depressions suggests that the strong central control which he favors is not to have as its function economic planning to benefit all of the people; rather, it seems, he is thinking of rule by an ingroup from which the majority of the population would be excluded. That he does reject the majority of the population has been seen in his responses to theE scale. When it is considered that he is antigovernment only when it comes to interference with the individual's freedom "to make money and spend it as he likes" (Item 63) we are led to the conclusion that his idea of central control is a combination of government and the most powerful business interests. Thus it appears that Mack comes as close as he can, within the confines of the r6- item PEC scale, to expressing that pattern of pseudoconservatism which emerged from the analysis of his interview.
Larry, on the other hand, deviates from the usual conservative position by moving further to the left. Not only does he insist upon the social obligations of government (Items 36 and 76) but he would accept limitations upon the profits of individuals and corporations (Items 52 and 63). These views can be reconciled with his strong conservative values and pro-business senti- ments, it seems, only by means of the belief that there is plenty for all, that it is the task of government to see to it that no one has too little or too much, and that this situation will permit people who, like himself, are willing to work hard, to get as much as they really need without causing others to suffer.
It would appear from this analysis, and from the analysis of Mack's and Larry's interviews in Chapter II, that the difference of o. 37 in their PEC- scale means is not great enough to represent the actual distance between them on a right vs. left dimension of politico-economic ideology. However, as the discussion in the preceding section has shown, the differences between pseudoconservatism, which we find in Mack, and genuine conservatism as represented by Larry are qualitative as well as quantitative, and it is to the credit of the PEC scale that it pointed out these differences while indicating at the same time that Mack is somewhat more extreme in a quantitative sense.
E. THE RELA TION BETWEEN ETHNOCENTRISM AND MEMBERSHIP IN V ARIOUS POLITICAL AND ECONOMIC GROUPINGS
We have considered in previous sections the nature of political ideologies as measured by the PEC scales, and the relation of these ideologies to ethno- centrism. The data also revealed numerous group differences in degree of
? 186 THE AUTHORIT ARIAN PERSONALITY
ethnocentrism. It was therefore natural to ask next: How do various political and economic groupings differ with respect to ethnocentrism? Information on such groupings (political party preference, parents' preference, organiza- tion memberships, income level, etc. ) was obtained on the front page of the questionnaire, so that it was possible to compute the mean (average) E score for each group. 9
We may consider first the relation between ethnocentrism and political group preference.
Political group preference was determined by means of several questions. First: "What political pany or group do you like best? " Second: "How do you feel about each of these political groups? Democrats (Anti-New Deal), Democrats
(New Deal), Willkie-type Republicans, Traditional Republicans. " In each case the subject was asked to check one of four choices: agree much, agree, disagree, disagree much. On the basis of his pattern of response to these questions, each subject was assigned a single "group preference" by two raters working together (semi-independent judgments). Each subject could be placed with relative ease into one or another of the following categories: (I) "Total" Democrats (suppon- ing both factions within the pany); (2) Anti-New Deal Democrats; (3) New Deal Democrats; (4) Willkie Republicans; (5) Traditional Republicans; (6) "Total" Republicans (supporting both factions); (7) New Deal Democrats, Willkie Republicans (supponing these two groupings and opposing the others, without indicating specific party preference; (8) Communists; (9) Socialists; ( IO) P. A. C. (National Citizens and C. I. O. Political Action Committee); (I I) Undecided (wrote in "undecided" to first question, omitted the second); (12) Anti-all panies; (I3) None, nonpartisan (gave this answer to first question, omitted the second); (I4) Self-contradictory (e. g. , supported Traditional Repub- licans and New Deal Democrats, gave no over-all party preference); (I5) Blank.
Table 12 (V) gives the number of cases (N) and the average A-S or E score for each political grouping, as well as for each sample tested and for the group of samples taking each form of the questionnaire. It is thus possible to compare, say, the New Deal Democrats in one sample with the other political groupings in that sample, or with the over-all totals for all New Dealers tested, or with the over-all totals for all subjects tested. 10
TheN's and means in the bottom row (horizontal) may be examined first. The ratio of Democrats (columns 1-3) to Republicans (columns 4-5) is roughly 10 to 7, a value which approximates the registration figures in the
1944 California elections. However, the proportion of New Deal Democrats and of "combined liberal groups" (column 17) is unduly large. It seems safe to conclude that the present sample shows a slight but significant bias
9 It would have been an important additional validation of the PEC scales to show differences among these groups with respect to PEC. Unfortunately, limitations of time prevented this further step.
10 It will be noted that group-membership data was not statisticized for several of the groups tested (approximately one-third of the total N), due to time limitations. However, the groups in Table 12 (V) appear to be a fairly representative selection of the total sample.
? POLITICO-ECONOMIC IDEOLOGY AND GROUP MEMBERSHIPS 187
in a liberal direction-a sampling factor which tends to lower the over-all E mean.
The various groupings appeared to fall into two clearly differentiated categories on the basis of average degree of ethnocentrism. First, the "con- servative" category (column r6: Total Democrats, Anti-New Deal Demo- crats, Traditional Republicans, Total Republicans), with E means ranging from 4. 17 to 4. 72 and averaging 4? 39? Second, the "liberal" category (New Deal Democrats, Willkie Republicans, New Deal-Willkie combined, Com- munists, Socialists, P. A. C. ), with E means ranging from 1. 25 to 3. 6o and averaging 3? 41. The difference of 0. 98 between the liberal and conservative categories is statistically very significant (far above the 1 per cent level). n
The rank order of the individual groupings, in terms of E mean, is similar to their rank order on a right-left political dimension. The traditional wings of the Democratic and Republican parties are the most conservative as well as the most ethnocentric (E means of 4. 2 to 4? 7)? The New Deal Democrats and the Willkie Republicans, representing in the main a slightly left-of- center political position, have a similar stand on ethnocentrism (means of
3. 6 and 3? 5)? Interestingly enough, those who support both the New Deal and Willkie, without making an over-all party choice, have a much lower E mean of 2. 4-an indication perhaps that greater political sophistication in liberals is accompanied by greater militancy regarding democratic group relations. The Socialists (those who gave this as their party preference, re- gardless of their views on the other groupings) were similar to the previous group with a mean of 2. 6. The most militantly anti-ethnocentric groups were the P. A. C. and the Communists, with E means of 2. 0 and 1. 25 respectively.
The great difference between the "liberal" and "conservative" categories, as well. as the rank order of the individual groupings, offer important evi- dence for the validity of the E scale and the E-PEC correlations reported above. The relationship between ethnocentrism and liberal-conservative group membership is very similar to that between ethnocentrism and liberal- conservative ideology in the individual. There is, on the average, a syste- matic relation between E, PEC, and political group preference, to the extent of a correlation of approximately -5-
0nce again, both the group trend and the individual differences must be emphasized. The relationships, though significant, are far from perfect. There is, for one thing, considerable individual variability within each group-
11 Critical ratios have not been computed for the various group differences discussed here. The following rules of thumb may be used in estimating the significance of the differences: Assume that the Standard Deviation for any grouping is equal to the S. D. for the sample containing it. This estimate is a maximal one so that any errors will tend to lower spuriously the value of the C. R. obtained. For groups with N's of about 50 each, differences of o. 6 are likely to be significant at the 5 percent level; when the N's are wo, a difference of 04 is adequate. These approximate standards hold for all tables in this section. An additional argument for the significance of these differences is their relative consistency from group to group.
? 188
THE AUTHORIT ARIAN PERSONALITY
TABLE 12 (V)
liEAN A? S OR E SOORESa FOR GROUPS SHOWING VARIOUS OVER? ALL POLITICAL PARIT PREFERENCES
Ol (1>
Group
Groups taking Form 78:
u. c. Public Speaking Class
WOmen 63. 63 43. 83 283. 11
u. C. Public Speaking Class
Men 0--35. 07173. 25
393. 00134. 19 343. 66
Extension Psychology Class WOmen
Professional WOmen
Totals
Groups taking Form 60:
73. 3943. 30
25. 35 83. 43o--o
Univ, of Oregon Student WOmen 2 3. 32 2 3. 24142. 73
134. 2832. 770 143. 2422. 370 10 3. 67 7 2. 32 0
Univ. of Oregon and Univ. of
California Student Women Univ, of Oregon and Univ. of
California Student Men Totals
Groups taking Form. 45:
Maritime School Men Psychiatric Clinic Men Psychiatric Clinic Women san ~entin Menb
Totals
Groups taking Form 40:
3 4. 59 4 4. 15192. 95
Geo. Washington Univ, women 13 4. 37 17
Mariti? e School Men Middle-Class Women Middle-Class Men Working-Class Men
Totals
Over-all totals
14 4. 14 15 5. 60
1 3. 80 0 2 5. 80 0
"'
~
<l)
j
8
N Mean N Mean
0 -- 0 82. 100 0--0
,. ~
(1>"' 01 z . . (1> . o0
-~ 0
~~ z"(1>
M
N Mean N \lean N :? ,lean N Mean N Mean N Mean
113. 3344. 18
13. 80 63. 70 112. 57
o -- 16. 00302. 06123. 4332. 63
63. 88
7 3. 6614
50 3. 72 8 2. 10 0
4. 19 86 2. 70
69 3. 17 24 3. 84
63. 1444. 46
93. 2612. 49
9 2. 08 6 3. 45
24 2. 79 11 3. 73 37 3. 72 12 2. 44 0
3. 07 12 2. 95 3. 70 45 2. 88
4. 49 74 4. 32 ? ? 203. 97 4. 07 26 3. 51 4. 74454. 60
4. 42 120 4. 08
4. 86 44 3. 83 5. 11 69 4. 95 3. 35 43 2. 76
&the following scales )Jere used . in the various forms: Form 78: A-S scale (10 items); Form 6():
E scale (12 items); Form 45: E scale (10 items); Form 40: E scale (5 items).
hrhe San QJentin group was not included in obtaining the totals, The means for this group were so much larger than those of the others. for reasons which seemed to have little to do with party preference, that their inclusion would distort the over-all picture,
ing, and there is much variation in group mean from one sample to the next. The New Deal Democrats, for example, obtained E means ranging from 2. 06 to 4?
57 . 48 68 ? 52
226 . 52
c. Groups taking Fora 45:c
Extension Testing Class Women
San Quentin Men 110 Psychiatric Clinic women
Psychiatric Clinic Men
W orking-Class Men and
Women?
D. Groups taking Fora 40:c
George Washington University Women California Service Club Men Middle-Class Men
Middle-Class women
W orking-Class Men
WOrking-Class WOmen
Los Angeles Men
Los Angeles Women
Mean ra
E. Groups taking Foras 40 and 45: Employment Service Men Veterans
(Form 40)
(Form 45) Maritime School Men (Form 40) (Form 45)
Mean ra
Mean r for all groups taking Forms 40 and 45
59 . 60
. 14d 71 ? 53
50 ? 55d 50 ? 75
340 . 51
132 . 48 63 . 64 69 ? 67
154 . 76 61 . "74 53 ? 86
117 . 61 130 ? 52
779 ? 66
55 ? ooe
51 ? 49 164 . 4le 179 ? 38
449 . 47 1568 ? 57
Mean ra
~n obtaining the over-all means, the individual group means were not weighted by N, and Z,. was not used.
bit will be recalled that in Form 60 the E scale contained 4 A-S items, there being no separate A-S scale.
CpEC is correlated with EA in groups taking Form 40, with total EA+B scale in groups taking Form 45.
dFor the PEC scale in the Psychiatric Clinic groUPs, the number of -en was 45, the number of men 29~ due to a substitution of forms.
erhese correlations are based on the EA scores of subjects taking Form 45 . as well as those taking Form 40.
E:PEC
A-S:PEC
? THE AUTHORITARIAN PERSONALITY
negligible. The same holds for class differences: in the Working-Class Men and Women the correlations are very high (. 74 and . 86), but in other groups which draw heavily upon the working class, notably San Quentin and the Maritime School, the correlations are very low (. 14 and . 4). The reasons for the variations in the size of r seem to. lie more in the specific nature of the group than in its sex or class status. It is interesting in this connection that two groups in which the E-PEC correlation was very high, the Working Class Men (. 74) and Women (. 86) also had the two lowest PEC means (see Table 8(V) ). W e may hypothesize that the E-PEC correlation will be highest when, other things such as the S. D. being equal, the group contains a sizeable minority of strong liberals. Judging from some of the other groups, the number of strong conservatives has less influence on the correlation. W e shall return to this question shortly.
The general level of the E-PEC correlations demonstrates that ethnocen- trism and conservatism, as measured by the present scales, are significantly but imperfectly related. 7 In everyday terms, we may say that conservatives are, on the average, significantly more ethnocentric than liberals are. The more conservative an individual is, the greater the likelihood that he is ethno- centric-but this is a probability and not a certainty. Since the existence of an affinity between these ideological patterns has often been observed previ- ously, the present correlations are perhaps less a startling discovery than an indication of the validity of the scales. To those who have been unaware of the E-PEC relationship, the significance of the correlations must be stressed. To those who tend to equate conservatism and ethnocentrism as psycho- logical trends in the individual, it must be pointed out that the correlations are far from perfect. Even with a much more reliable measure of PEC, the correlation with E could hardly average over . 70-a value inadequate for predictive purposes. It becomes necessary, then, to understand what pro- duces the close association between these ideological patterns in the indi- vidual, as well as what systematic factors-apart from ignorance or misinfor- mation-make the E-PEC correlation less than I . o.
A theoretical basis for the close tie between conservatism and ethnocen- trism is suggested by certain similarities in their major underlying trends. Support of the prevailing politico-economic ideology and authority is, ap- parently, often a part of the generalized ethnocentric tendency to submit to accepted authority in all areas of social life. Similarly, ethnocentric rejection of outgroups is expressed in the politico-economic sphere by resistance to social change and by the tendency to subsume progressive political ideologies under the general heading of "foreign" outgroups and ideas (threats to in- group authority). The interconnection between the two ideologies and the difficulty of separating them even for purposes of study are revealed by a
? 7 These results are, on the whole, consistent with those of the other studies mentioned at the start of this chapter.
? POLITICO-ECO~OMIC IDEOLOGY AND GROUP MEMBERSHIPS I 8I
number of scale items. For example: "America may not be perfect, but the American Way has brought us about as close as human beings can get to a perfect society. " To support this idea is, it would seem, to express both politico-economic conservatism and the ingroup idealization so character- istic of ethnocentrism. The item, "The worst danger to real Americanism during the last 50 years has come from foreign ideas and agitators," is another example of politicalized ethnocentrism: again we find moral stereotypy and externalization of blame for social problems onto the threatening outgroup.
There are also theoretical reasons for expecting a relationship between liberalism and anti-ethnocentrism. Both tend to involve a critical attitude toward prevailing authorities and traditions. The identification with the masses (workers, "the common man," "the weak and downtrodden") so often a central theme in left-wing political ideology, finds expression also in opposition to ethnocentrism and outgroup suppression. Indeed, the forma- tion of leftist political views in youth often begins with a sense of the injus- tice of anti-Semitism or anti-Negroism; when sympathetic imagery of sub- ordinate groups is extended to include the working class, the transition from "group relations" to "politico-economic" ideology has begun. The further development of liberal-radical views is ordinarily based on imagery and atti- tudes identical to those underlying anti-ethnocentric ideology: opposition to hierarchy and to dominance-submission, removal of class and group barriers, emphasis on equalitarian interaction, and so on.
We have also to consider the "correlation-lowering" factors. Why, in view of the theoretical argument above, is the E-PEC correlation not higher? The present data suggest, but are not adequate to test, several hypotheses. The correlation charts (scattergrams) for each sample reveal that extreme liberals (low scorers on PEC) are for the most part low as well on E. But the "middles" on PEC are extremely diversified with respect to standing on E. It is possible that the group which is low on E but middle on PEC consists largely of the "politically pacifistic" liberals discussed previously. Practically none of the subjects were low on PEC and high on E (ethnocentric liberals); such il1dividuals would, however, be well worth intensive study.
The high scorers on PEC are more variable on E than are the low scorers. While most of those high on PEC are also high onE, a considerable number are middle and a small but consistent percentage low onE (nonethnocentric conservatives). In other words, strong political liberalism is a pretty good indicator of anti-ethnocentrism, but political conservatism is less consistently related to ethnocentrism.
In attempting to explain the variability of conservatives with respect to ethnocentrism, we are reminded of the distinction between "genuine" and "pseudo-" previously drawn with respect to patriotism and traditional- conventional values (Chapter IV). One can be politically conservative, just as one can be patriotic (in the sense of firm attachment to American culture
? THE AUTHORITARIAN PERSONALITY
and tradition), without being ethnocentric. vVe should like to use the term "genuine conservative" to refer to the individual with this broad pattern of thought. He is "genuine" because, whatever the merits of his political views, he is seriously concerned with fostering what is most vital in the American democratic tradition. He believes, for example, in the crucial importance of the profit motive and in the necessity of economic insecurity; but he wants the best man to win no matter what his social background. He is resistant to social change, but he can be seriously critical of the national and political ingroups and-what is more important-he is relatively free of the rigidity and deep-lying hostility characteristic of ethnocentrism.
The ethnocentric conservative is the pseudoconservative, for he betrays in his ethnocentrism a tendency antithetical to democratic values and tradi- tion. He is the E-PEC "correlation raiser" because, as discussed above, his politico-economic views are based on the same underlying trends-submis- sion to authority, unconscious handling of hostility toward authority by means of displacement and projection onto outgroups, and so on-as his ethnocentrism. It is indeed paradoxical that the greatest psychological poten- tial for antidemocratic change should come from those who claim to repre- sent democratic tradition. For the pseudoconservatives are the pseudo- democrats, and their needs dispose them to the use of force and oppression in order to protect a mythical "Americanism" which bears no resemblance to what is most vital in American history. .
An additional hypothesis may be proposed regarding individuals high on E but middle on PEC. These may well be pseudoconservatives who have kept up with changes in the actual politico-economic situation by making changes in traditional (individualistic) conservative ideology. They empha- size competitiveness as a value, yet they support the concentration of eco- nomic power in big business-the greatest single threat at present to the individual competing businessman. They emphasize economic mobility and the "Horatio Alger" myth, yet they support numerous forms of discrimina- tion that put severe limitations on the mobility of large sections of the popu- lation. They may also believe in extending the economic functions of government, not for humanitarian reasons but as a means of limiting the power of labor and other groups.
This is not merely a "modern conservatism. " It is, rather, a totally new direction: away from individualism and equality of opportunity, and teward a rigidly stratified society in which there is a minimum of economic mobility and in which the "right" groups are in power, the outgroups subordinate. Perhaps the term "reactionary" fits this ideology best. Ultimately it is fascism. While certainly not a necessary sequel to laissez-faire conservatism, it can be regarded as a possible (and not uncommon) distortion of con- servatism-a distortion which retains certain surface similarities but which changes the basic structure into the antithesis of the original. Since most
? No.
1. 5. 8.
13. 15. 22. 27. 36. 44. 52. 61. 63. 68. 71. 76. 78.
I t e m
(Value of dollar) (Depressions)
(Charity)
(Businessmen, artists) (Middle of the road) (Political candidate) (Rebellious ideas)
(Gov' t. responsibility) (Socialized medicine) (Taxes, corporations) (Economic security) (Gov' t. interference)? (Unions stronger)
(Ford, Morgan)
(Gov' t. activity) (Ability will tell)
Groupa Mack L a r r y Mean
6 7 6. 10 5 1 3,33 3 7 5. 46 1 1 2. 29 7 5 4. 35 7 7 6. 38 5 6 3. 86 2 1 3. 22 2 6 2. 38 2 3 3. 66 6 6 3. 75 5 1 4. 01 6 2 4. 58 7 6 5. 30 2 2 3. 32 7 6 4. 74
Group a D. P .
1. 16 2. 76 1. 48 1. 70 2. 90 0. 32 2. 84 3. 01 1. 69 2. 29 2. 68 2. 39 2. 30 2. 00 2. 76 1. 99
POLITICO-ECONOMIC IDEOLOGY AND GROUP MEMBERSHIPS 183
Americans are "middle on PEC," it becomes crucial to understand the psy- chological dispositions which help to determine new ideological directions in the individual.
The above distinction regarding ideological patterns within the political left and right are presented as hypotheses to help explain the scale results. 8 If these hypotheses are not borne out, others will be needed. For it is clear that political ideologies do not fall neatly along a simple liberalism-conserva- tism dimension; that the relation between ethnocentrism and "conservatism" is extremely complex; and that the individual's receptivity to political ideol- ogy, as to "group relations" ideology, is based to a large extent on deep- lying personality trends.
D. V ALIDATlON BY CASE STUDIES: THE RESPONSES OF MACK AND LARRY ON THE PEC SCALE
In an attempt to judge the validity of the PEC scale we may here, as in Chapters III and IV, compare the scale responses of Mack and Larry and consider them in relation to material from their interviews (Chapter II). The PEC-scale scores of these two subjects, the group mean, and the D. P. for each of the 16 PEC items included in Form 78 are shown in Table II (V).
TABLE 11 (V)
RESPONSES OF MACK AND LARRY ON THE PEC SCALE
Mean per item
aThe group means and D. P. 's are based on all four groups taking Form 78.
8 Further hypotheses, plus supporting evidence, are presented in Chapter XVII, which deals with the interview material.
4. 56 4. 19 4. 17
2. 14
? THE AUTHORITARIAN PERSONALITY
The two men are much more similar in their PEC-scale scores than was the case with their A-S and E scores. Larry's mean score, 4? I9, is at the group mean and Mack's, 4? 56, is not very far above. When the group of Public Speaking Men to which these subjects belonged is considered by itself, Mack is just inside the high quartile, Larry is just below it. On I I of the I6 items the scores of the two men do not differ by more than one point. Attention to the individual items, however, shows that the similarities are confined to certain areas of politico-economic ideology; in other areas there are sharp differences.
Mack and Larry are most similar in their consistent support of general conventional-conservative values. Both agree, usually rather strongly, with Items I (Value of a dollar), I 5 (Middle of the road), 22 (Political candidate), 27 (Rebellious ideas), ? I (Ford, Morgan), and 78 (Ability will tell). This seems to be in keeping with the interviews, in which both men expressed the usual conservative criticisms of the New Deal. Both men, it appears, accept the view that a man's getting ahead depends most of all upon his living according to the values of thrift, determination, work, honesty, conformity, and the like.
Examination of the scores on other items, however, indicates that Mack and Larry differ with respect to the context in which their conservative values occur. For Mack the context appears to be one of upward social mobility on a class-ingroup basis, for Larry it appears to be one of nineteenth- century liberalism. This seems to be expressed in their wide disagreement on Items 5 (Depression) and 68 (Unions stronger). Mack's belief that de- pressions are "only natural" can be interpreted as an expression of the broader idea that, in the nature of things, the benefits to be had in our society are insufficient to go around and that it is no more than proper that the major portion of them should go to the "right people," that is, to an ingroup. This ingroup does not seem to include organized labor (Item 68) nor the various minority groups which he rejected in his responses on the E scale. This would appear to be another manifestation of Mack's tendency, so marked in his interview, to make rigid ingroup-outgroup distinctions in his thinking about politico-economic matters as well as about social relations generally.
For Larry, on the other hand, the value for getting ahead does not exclude the possibility of various other kinds of people getting ahead, for he seems to be thinking in terms of an expanding economy in which working men can have a strong role (Item 68) and in which depressions are unnecessary (Item 5). The absence of any ingroup-outgroup distinction, and optimism with respect to the possibilities of economic abundance were outstanding features of Larry's interview.
Neither man shows the usual conservative opposition to the government's participation in the economic life of the nation: for Items 36 (Government responsibility), 44 (Socialized medicine), 52 (Taxes, corporations), 63 (Gov-
? POLITICO-ECOXOJ\IIC IDEOLOGY AND GROUP MEMBERSHIPS 185
ernment interference), and '76 (Government activity) the two men have the same low mean score of 2. 6o. Their reasons, however, seem to be different. When the responses to the total scale are considered, it appears that Mack deviates from true, laissez-faire conservatism by taking a stand further to the political right. The fact that he rejects labor unions and believes in the inevitability of depressions suggests that the strong central control which he favors is not to have as its function economic planning to benefit all of the people; rather, it seems, he is thinking of rule by an ingroup from which the majority of the population would be excluded. That he does reject the majority of the population has been seen in his responses to theE scale. When it is considered that he is antigovernment only when it comes to interference with the individual's freedom "to make money and spend it as he likes" (Item 63) we are led to the conclusion that his idea of central control is a combination of government and the most powerful business interests. Thus it appears that Mack comes as close as he can, within the confines of the r6- item PEC scale, to expressing that pattern of pseudoconservatism which emerged from the analysis of his interview.
Larry, on the other hand, deviates from the usual conservative position by moving further to the left. Not only does he insist upon the social obligations of government (Items 36 and 76) but he would accept limitations upon the profits of individuals and corporations (Items 52 and 63). These views can be reconciled with his strong conservative values and pro-business senti- ments, it seems, only by means of the belief that there is plenty for all, that it is the task of government to see to it that no one has too little or too much, and that this situation will permit people who, like himself, are willing to work hard, to get as much as they really need without causing others to suffer.
It would appear from this analysis, and from the analysis of Mack's and Larry's interviews in Chapter II, that the difference of o. 37 in their PEC- scale means is not great enough to represent the actual distance between them on a right vs. left dimension of politico-economic ideology. However, as the discussion in the preceding section has shown, the differences between pseudoconservatism, which we find in Mack, and genuine conservatism as represented by Larry are qualitative as well as quantitative, and it is to the credit of the PEC scale that it pointed out these differences while indicating at the same time that Mack is somewhat more extreme in a quantitative sense.
E. THE RELA TION BETWEEN ETHNOCENTRISM AND MEMBERSHIP IN V ARIOUS POLITICAL AND ECONOMIC GROUPINGS
We have considered in previous sections the nature of political ideologies as measured by the PEC scales, and the relation of these ideologies to ethno- centrism. The data also revealed numerous group differences in degree of
? 186 THE AUTHORIT ARIAN PERSONALITY
ethnocentrism. It was therefore natural to ask next: How do various political and economic groupings differ with respect to ethnocentrism? Information on such groupings (political party preference, parents' preference, organiza- tion memberships, income level, etc. ) was obtained on the front page of the questionnaire, so that it was possible to compute the mean (average) E score for each group. 9
We may consider first the relation between ethnocentrism and political group preference.
Political group preference was determined by means of several questions. First: "What political pany or group do you like best? " Second: "How do you feel about each of these political groups? Democrats (Anti-New Deal), Democrats
(New Deal), Willkie-type Republicans, Traditional Republicans. " In each case the subject was asked to check one of four choices: agree much, agree, disagree, disagree much. On the basis of his pattern of response to these questions, each subject was assigned a single "group preference" by two raters working together (semi-independent judgments). Each subject could be placed with relative ease into one or another of the following categories: (I) "Total" Democrats (suppon- ing both factions within the pany); (2) Anti-New Deal Democrats; (3) New Deal Democrats; (4) Willkie Republicans; (5) Traditional Republicans; (6) "Total" Republicans (supporting both factions); (7) New Deal Democrats, Willkie Republicans (supponing these two groupings and opposing the others, without indicating specific party preference; (8) Communists; (9) Socialists; ( IO) P. A. C. (National Citizens and C. I. O. Political Action Committee); (I I) Undecided (wrote in "undecided" to first question, omitted the second); (12) Anti-all panies; (I3) None, nonpartisan (gave this answer to first question, omitted the second); (I4) Self-contradictory (e. g. , supported Traditional Repub- licans and New Deal Democrats, gave no over-all party preference); (I5) Blank.
Table 12 (V) gives the number of cases (N) and the average A-S or E score for each political grouping, as well as for each sample tested and for the group of samples taking each form of the questionnaire. It is thus possible to compare, say, the New Deal Democrats in one sample with the other political groupings in that sample, or with the over-all totals for all New Dealers tested, or with the over-all totals for all subjects tested. 10
TheN's and means in the bottom row (horizontal) may be examined first. The ratio of Democrats (columns 1-3) to Republicans (columns 4-5) is roughly 10 to 7, a value which approximates the registration figures in the
1944 California elections. However, the proportion of New Deal Democrats and of "combined liberal groups" (column 17) is unduly large. It seems safe to conclude that the present sample shows a slight but significant bias
9 It would have been an important additional validation of the PEC scales to show differences among these groups with respect to PEC. Unfortunately, limitations of time prevented this further step.
10 It will be noted that group-membership data was not statisticized for several of the groups tested (approximately one-third of the total N), due to time limitations. However, the groups in Table 12 (V) appear to be a fairly representative selection of the total sample.
? POLITICO-ECONOMIC IDEOLOGY AND GROUP MEMBERSHIPS 187
in a liberal direction-a sampling factor which tends to lower the over-all E mean.
The various groupings appeared to fall into two clearly differentiated categories on the basis of average degree of ethnocentrism. First, the "con- servative" category (column r6: Total Democrats, Anti-New Deal Demo- crats, Traditional Republicans, Total Republicans), with E means ranging from 4. 17 to 4. 72 and averaging 4? 39? Second, the "liberal" category (New Deal Democrats, Willkie Republicans, New Deal-Willkie combined, Com- munists, Socialists, P. A. C. ), with E means ranging from 1. 25 to 3. 6o and averaging 3? 41. The difference of 0. 98 between the liberal and conservative categories is statistically very significant (far above the 1 per cent level). n
The rank order of the individual groupings, in terms of E mean, is similar to their rank order on a right-left political dimension. The traditional wings of the Democratic and Republican parties are the most conservative as well as the most ethnocentric (E means of 4. 2 to 4? 7)? The New Deal Democrats and the Willkie Republicans, representing in the main a slightly left-of- center political position, have a similar stand on ethnocentrism (means of
3. 6 and 3? 5)? Interestingly enough, those who support both the New Deal and Willkie, without making an over-all party choice, have a much lower E mean of 2. 4-an indication perhaps that greater political sophistication in liberals is accompanied by greater militancy regarding democratic group relations. The Socialists (those who gave this as their party preference, re- gardless of their views on the other groupings) were similar to the previous group with a mean of 2. 6. The most militantly anti-ethnocentric groups were the P. A. C. and the Communists, with E means of 2. 0 and 1. 25 respectively.
The great difference between the "liberal" and "conservative" categories, as well. as the rank order of the individual groupings, offer important evi- dence for the validity of the E scale and the E-PEC correlations reported above. The relationship between ethnocentrism and liberal-conservative group membership is very similar to that between ethnocentrism and liberal- conservative ideology in the individual. There is, on the average, a syste- matic relation between E, PEC, and political group preference, to the extent of a correlation of approximately -5-
0nce again, both the group trend and the individual differences must be emphasized. The relationships, though significant, are far from perfect. There is, for one thing, considerable individual variability within each group-
11 Critical ratios have not been computed for the various group differences discussed here. The following rules of thumb may be used in estimating the significance of the differences: Assume that the Standard Deviation for any grouping is equal to the S. D. for the sample containing it. This estimate is a maximal one so that any errors will tend to lower spuriously the value of the C. R. obtained. For groups with N's of about 50 each, differences of o. 6 are likely to be significant at the 5 percent level; when the N's are wo, a difference of 04 is adequate. These approximate standards hold for all tables in this section. An additional argument for the significance of these differences is their relative consistency from group to group.
? 188
THE AUTHORIT ARIAN PERSONALITY
TABLE 12 (V)
liEAN A? S OR E SOORESa FOR GROUPS SHOWING VARIOUS OVER? ALL POLITICAL PARIT PREFERENCES
Ol (1>
Group
Groups taking Form 78:
u. c. Public Speaking Class
WOmen 63. 63 43. 83 283. 11
u. C. Public Speaking Class
Men 0--35. 07173. 25
393. 00134. 19 343. 66
Extension Psychology Class WOmen
Professional WOmen
Totals
Groups taking Form 60:
73. 3943. 30
25. 35 83. 43o--o
Univ, of Oregon Student WOmen 2 3. 32 2 3. 24142. 73
134. 2832. 770 143. 2422. 370 10 3. 67 7 2. 32 0
Univ. of Oregon and Univ. of
California Student Women Univ, of Oregon and Univ. of
California Student Men Totals
Groups taking Form. 45:
Maritime School Men Psychiatric Clinic Men Psychiatric Clinic Women san ~entin Menb
Totals
Groups taking Form 40:
3 4. 59 4 4. 15192. 95
Geo. Washington Univ, women 13 4. 37 17
Mariti? e School Men Middle-Class Women Middle-Class Men Working-Class Men
Totals
Over-all totals
14 4. 14 15 5. 60
1 3. 80 0 2 5. 80 0
"'
~
<l)
j
8
N Mean N Mean
0 -- 0 82. 100 0--0
,. ~
(1>"' 01 z . . (1> . o0
-~ 0
~~ z"(1>
M
N Mean N \lean N :? ,lean N Mean N Mean N Mean
113. 3344. 18
13. 80 63. 70 112. 57
o -- 16. 00302. 06123. 4332. 63
63. 88
7 3. 6614
50 3. 72 8 2. 10 0
4. 19 86 2. 70
69 3. 17 24 3. 84
63. 1444. 46
93. 2612. 49
9 2. 08 6 3. 45
24 2. 79 11 3. 73 37 3. 72 12 2. 44 0
3. 07 12 2. 95 3. 70 45 2. 88
4. 49 74 4. 32 ? ? 203. 97 4. 07 26 3. 51 4. 74454. 60
4. 42 120 4. 08
4. 86 44 3. 83 5. 11 69 4. 95 3. 35 43 2. 76
&the following scales )Jere used . in the various forms: Form 78: A-S scale (10 items); Form 6():
E scale (12 items); Form 45: E scale (10 items); Form 40: E scale (5 items).
hrhe San QJentin group was not included in obtaining the totals, The means for this group were so much larger than those of the others. for reasons which seemed to have little to do with party preference, that their inclusion would distort the over-all picture,
ing, and there is much variation in group mean from one sample to the next. The New Deal Democrats, for example, obtained E means ranging from 2. 06 to 4?
