C. Results with Successive Forms of the F Scale
1. Statistical Properties of the Preliminary Scale (Form 78)
The preliminary F scale, made up of the 38 items listed above, was administered as a part of questionnaire Form 78 to four groups of subjects in the spring of 1945. These groups were described in Chapter III,[8] and they are listed in Table 11 (III).
The scoring of the scale followed the procedures used with the A-S, E, and PEC scales. Except in the case of negative items, a mark of +3 was scored 7, +2 was scored as 6, and so on. Items 12, 20, and 28 are negative (they state the unprejudiced position), and here, of course, a mark of +3 was scored 1, and so on. Table 2 (VII) gives the reliability coefficients, mean scores per item, and Standard Deviations for these four groups. The mean reliability of .74 is within the range ordinarily regarded as adequate for group comparisons, but well below what is required of a truly accurate instrument. It might be said that, considering the diversity of elements that went into the F scale, the degree of consistency indicated by the present figure is all that could be expected of this preliminary form of the scale. The question was whether by revision of the scale it might be possible to attain the degree of reliability that characterizes the E scale, or whether we might be dealing here – as seemed to be the case in the PEC scale – with areas of response in which people are simply not very consistent.
Table 2 (VII)
Reliability of the F Scale (Form 78)a
Property Group Over-all b
A B C D
Reliability .78 .56 .72 .88 .74
Mean (total) 3.94 3.72 3.75 3.43 3.71
Mean (odd half) 3.80 3.59 3.60 3.22 3.55
Mean (even half) 4.08 3.87 3.91 3.64 3.88
S.D. (total) .71 .57 .70 .86 .71
S.D. (odd half) .87 .71 .85 .94 .84
S.D. (even half) .69 .65 .76 .84 .74
N 140 52 40 63 295
Range 2.12–5.26 2.55–4.87 2.39–5.05 1.68–5.63 1.68–5.63
a The four groups on which these data are based are:
Group A: U.C. Public Speaking Class Women.
Group B: U.C. Public Speaking Class Men.
Group C: U.C. Extension Psychology Class Women.
Group D: Professional Women.
b In obtaining the over-all means, the individual group means were not weighted by N.
It may be noted that the Professional Women show considerably more consistency than do the other groups of subjects, their reliability coefficient of .88 being in the neighborhood of that regularly obtained with the E scale. Since these women are considerably older, on the average, than our other subjects, it may be suggested that the higher reliability is due to their greater consistency of personality.
There appears to be no ready explanation for the low reliability found in the case of the Public Speaking Men. It may be noted that the Standard Deviation and the range for this group were also unusually small. Adequate explanation would require data from a larger sample of men and from an improved F scale.
Examination of Table 2 (VII) shows that there are no extremely high and no extremely low scores in any of the groups and that the obtained means are near the neutral point. The relatively narrow distribution of scores – narrow as compared with those obtained from the other scales – may be in part a result of lack of consistency within the scale: unless the items are actually expressive of the same general trend, we could hardly expect an individual to respond to the great majority of them with consistent agreement or consistent disagreement. On the other hand, it is possible that the present sample does not contain subjects who are actually extreme with respect to the pattern which the F scale was designed to measure. This circumstance (lowered »range of talent«) would tend to lower the reliability coefficients.
The F scale correlated .53 with A-S and .65 with E, in Form 78.
2. Item Analysis and Revision of the Preliminary Scale
Data obtained from the initial four groups of subjects were used in attempting to improve the F scale – to increase its reliability and to shorten it somewhat, without loss in its breadth or meaningfulness. As with the other scales, the Discriminatory Power of an item provided the major statistical basis for judging its worth. Since it was intended that the F scale should not only have internal consistency but should also correlate highly with overt prejudice, attention was given both to the item's relation to the total F scale and to its ability to discriminate between high and low scorers on the A-S scale. An item's Discriminatory Power in terms of A-S (D.P.A-S) is simply the difference between the mean score of the high A-S quartile on that item and the mean score of the low A-S quartile on the item. Table 3 (VII) gives for each item the mean score, the Discriminatory Power in terms of high vs. low scorers on F (D.P.F), the D.P.F's order of merit, the D.P.A-S, the latter's order of merit and, finally, the item's rank in a distribution of the sums of the D.P.F plus the D.P.A.A-S. This final rank order was a convenient index of the item's statistical »goodness« for our over-all purpose.
Table 3 (VII)
Means and Discriminatory Powers of the F-Scale Items (Form 78)a
Item Mean D.P.Fb Rank D.P.ASc Rank Final
D.P.F D.P.AS Rankd
(D.P.F +
D.P.AS)
2. (Astrology) 2.60 1.74 (22) 1.24 (11) (18)
3. (Force to preserve) 3.04 1.98 (18) 1.05 (17) (15)
6. (Women restricted) 2.93 1.75 (21) 0.41 (32) (26)
9. (Red-blooded life) 3.99 2.04 (15) -0.08 (35) (29)
10. (Pearl Harbor Day) 2.22 2.20 (9) 1.37 (6) (8)
12. (Modern church) 4.67 0.19 (38) -1.18 (38) (38)
14. (Rats ... germs) 4.44 1.60 (26.5) 0.85 (24) (23.5)
17. (Familiarity) 3.33 1.86 (19) 1.56 (4) (10)
19. (One should avoid) 3.63 0.76 (36) 0.70 (27) (35)
20. (Progressive education) 3.28 1.07 (33) -0.25 (37) (37)
23. (Undying love) 3.62 2.61 (4) 1.17 (13) (5)
24. (Things unstable) 5.01 0.79 (35) 0.88 (22) (33)
28. (Novels or stories) 3.02 1.29 (30) 0.76 (26) (27)
30. (Reports of atrocities) 4.20 0.43 (37) 0.66 (28) (36)
31. (Homosexuals) 3.22 2.16 (10) 1.18 (12) (13)
32. (Essential for learning) 3.31 1.67 (24) 1.10 (16) (20)
35. (Law in own hands) 2.50 1.42 (29) 0.62 (29.5) (28)
38. (Emphasis in college) 3.91 1.20 (31) 1.14 (15) (25)
39. (Supernatural force) 3.97 2.54 (6) 1.26 (9.5) (4)
42. (For one reason) 2.06 1.05 (34) 0.59 (31) (34)
43. (Sciences like chemistry) 4.35 2.79 (3) 0.97 (18) (6)
46. (Sex orgies) 3.64 2.11 (12.5) 0.93 (20) (14)
47. (Honor) 3.00 2.09 (14) 1.65 (3) (7)
50. (Obedience and respect) 3.72 3.09 (1) 1.55 (5) (2)
53. (Things too intimate) 4.82 1.99 (17) -0.23 (36) (32)
55. (Leisure) 5.20 2.11 (12.5) 1.26 (9.5) (11)
56. (Crime wave) 4.60 1.16 (32) 0.62 (29.5) (31)
58. (What a man does) 3.48 1.70 (23) 0.87 (23) (22)
59. (Always war) 4.26 2.59 (5) 1.91 (2) (3)
60. (Important values) 4.17 1.60 (26.5) 0.31 (34) (30)
65. (World catastrophe) 2.58 1.55 (28) 0.90 (21) (23.5)
66. (Books and movies) 4.10 2.48 (7) 0.38 (33) (19)
67. (Eye to profit) 3.71 2.21 (8) 0.78 (25) (17)
70. (Plots by politicians) 3.27 1.85 (20) 1.15 (14) (16)
73. (Infection and disease) 4.79 2.02 (16) 1.34 (8) (12)
74. (Tireless leaders) 5.00 1.66 (25) 0.94 (19) (21)
75. (Sex crimes) 3.26 2.81 (2) 2.07 (1) (1)
77. (No sane person) 4.12 2.12 (11) 1.36 (7) (9)
Mean/Person/Item 3.71 1.80 0.89
a The four groups on which these data are based are: Group A: U.C. Public Speaking Class Women (N = 140); Group B: U.C. Public Speaking Class Men (N = 52); Group C: U.C. Extension Psychology Class Women (N = 40); Group D: Professional Women (N = 63). In obtaining the over-all means, the individual group means were not weighted by N.
b D.P.F is based on the difference between the high quartile and the low quartile on the F scale distribution.
c D.P.AS is based on the difference between the high quartile and the low quartile on the A-S scale distribution. E.g., the D.P.AS of 1.24 on Item 2 indicates that the mean of the low quartile on A-S was 1.24 points lower than the mean of the high quartile on A-S.
d For each item the sum of D.P.F+D.P.AS is obtained. The final rank of an item is the rank of this sum in the distribution of sums for the whole scale.
The average D.P.F, 1.80, is considerably below that found in the case of the A-S or E scales. Yet it indicates that, in general, the items yield statistically significant differences between the high and the low quartiles. Sixteen D.P.'s are above 2, 18 fall in the range 1–2, and only 4 are below 1. The means are, in general, fairly satisfactory; they average 3.71, which is near the neutral point of 4.0, and only 9 means are definitely too extreme, i.e., above 5.0 or below 3.0. As is to be expected, only 2 of the items with extreme means yield D.P.'s as great as 2.0.
The D.P.'s in terms of A-S are, of course, much lower; yet there are 17 items which appear to be significantly related to A-S, i.e., have a D.P.A-S greater than 1.0. Since it is the total F pattern that we expect to correlate with A-S and E, it is not necessary that each single F item by itself be significantly related to the latter. In general, items which are most discriminating in terms of F tend to discriminate best in terms of A-S, though there are some striking exceptions. In deciding whether to retain an item for use in a revised scale most weight was given to the D.P.F and to the general principles guiding our scale construction; these things being equal, the greater an item's D.P.A-S, the greater its chances of being included in the revised scale.
We may now inquire what it is that distinguishes the items which turned out well statistically from those that turned out poorly. Can any general statements be made about each of these two groups of items that can serve as guides in the formulation of new items? The first question concerns the nine groups of items chosen to represent the variables that entered into the conceptualization of F. Do most of the items with high D.P.'s pertain to a few of the variables? Are there some variables which simply do not belong to the pattern we are considering? Three of the clusters, Sex, Authoritarian Aggression, and Authoritarian Submission, had mean D.P.'s above 2.0, the remaining clusters having mean D.P.'s m the range 1.26–1.80. Projectivity (1.70), Destructiveness and Cynicism (1.56), and Conventionalism (1.26) were the least satisfactory. However, it is to be noted that every cluster has within it at least one item with a D.P. above 2.0. At this stage, therefore, it seemed best not to eliminate any of the variables but to give attention to improving or replacing the poorer items found in each cluster.
Turning to a consideration of items which proved to be outstandingly good in the statistical sense, we note that Item 75 (Sex crimes) leads all the rest, i.e., has the highest sum of D.P.F plus D.P.A-S. This item represents rather well the ideal to which we aspired in formulating items for the F scale. Not only is there a wide distribution of responses, with a mean fairly near the neutral point, but the item combines, apparently in a very effective way, several ideas which according to theory have crucial roles in prejudice: the underlying interest in the more primitive aspects of sex, the readiness for all-out physical aggressiveness, the justification of aggression by an appeal to moralistic values. More than this, the item seems to be sufficiently free of any logical or automatic connection with overt prejudice. That the next best item, 50 (Obedience and respect), should be outstandingly differentiating is not surprising since this kind of authoritarianism is a well-known aspect of the fascist outlook. The device of putting the authoritarianism in a context of child training seems to remove it from the surface of ethnocentrism; but whether or not this is true, the item pertains to an aspect of the fascist philosophy that could in no case be left out of account.
Third in the rank order of goodness is Item 59, »Human nature being what it is, there will always be war and conflict.« This item, from the Destructiveness and Cynicism cluster, expresses several ideas which are particularly important in the F syndrome. In addition to an element of overt antipacifist opinion, there is contempt for men and acceptance of the »survival of the fittest« idea as a rationalization for aggressiveness. The next item, 39 (Supernatural force), seems to express very well the tendency to shift responsibility to outside forces beyond one's own control. This is a manifestation of what has been termed ego weakness; the item has also been placed in the Authoritarian Submission cluster on the ground that faith in a supernatural force is related to faith in ingroup authorities. It was not expected that the presence of religious feeling and belief would by itself be significant for prejudice; the aim in devising the present item was to compose a statement which was so extreme that not too many subjects would agree with it and which placed enough emphasis upon »total allegiance« and obedience »without question« so that the uncritically submissive person could distinguish himself. The mean of 3.97 and the D.P.F of 2.54 indicate that this aim was largely realized. Item 23 (Undying love), which ranked fifth in order of goodness, expresses extreme moral conventionality and ingroup feeling related to the family. It has a place in both the Authoritarian Submission and the Authoritarian Aggression clusters, because it includes both allegiance to the ingroup and a punishing attitude (»He is indeed contemptible«) toward those who violate this value. The statement is so exaggerated, so expressive, as it seems, of the »protesting too much« attitude that we may wonder if strong agreement with it does not mask underlying but inhibited rebellious hostility against parents and parent figures.
Concerning all five of these items it may be said that they are highly diverse in their surface content, that they pertain to various aspects of the underlying theory – superego, ego, and id are expressed – and that with the possible exception of Item 50 (Obedience and respect) they are highly indirect in the present sense of the term. Indeed, as one examines further the ranking of the items in terms of their Discriminatory Powers – (Sciences like chemistry), (Honor), (Pearl Harbor Day), (No sane person), (Familiarity), (Leisure), (Infection and disease) – he may note that, in general, items which are best in the statistical sense are those which seem best in their formulation and in terms of our over-all theory and method of approach.
Items which turned out to be poor in the statistical sense are, in retrospect, easy to criticize. In some instances there was a failure in formulation: the statement was so unclear or ambiguous that many of the subjects, apparently, drew from it different implications from those intended. This would seem to have been true particularly of Items 12 (Modern church) and 20 (Progressive education). In other instances, e.g., Items 24 (Things unstable), 74 (Tireless leaders), and 58 (Crime wave), the statements contained too large an element of truth or rational justification and so appealed, as shown by the high mean scores, to both high and low scorers. Again, some of the items were too crude or openly aggressive, so that most of the high scorers as well as the low scorers tended to disagree. For example, Items 42 (For one reason) and 35 (Law in own hands) have both low means and low D.P.'s.
3. The Second F Scale: Form 60
In preparing the new form of the questionnaire, the E and PEC scales were shortened so drastically that a comparatively long F scale (34 items) could still be used without allowing the total number of scale items to go above 60. Since we were still faced with the task of producing a reliable F scale, without sacrificing breadth or meaningfulness, it seemed the better part of wisdom not to undertake much shortening of it at this stage.
Table 4 (VII)
The F Scale: Form 60
a. Conventionalism: Rigid adherence to conventional, middle-class values.
1. Obedience and respect for authority are the most important virtues children should learn.
6. A person who has bad manners, habits, and breeding can hardly expect to be liked and accepted by decent people.
53. One main trouble today is that people talk too much and work too little.
b. Authoritarian Submission: Submissive, uncritical attitude toward idealized moral authorities of the ingroup.
1. Obedience and respect for authority are the most important virtues children should learn.
7. Science has carried man very far, but there are many important things that can never possibly be understood by the human mind.
11. Every person should have complete faith in some supernatural power whose decisions he obeys without question.
23. Young people sometimes get rebellious ideas, but as they grow up they ought to get over them and settle down.
29. No sane, normal, decent person could ever think of hurting a close friend or relative.
32. What this country needs most, more than laws and political programs, is a few courageous, tireless, devoted leaders in whom the people can put their faith.
44. In order for us to do good work, it is necessary that our bosses outline carefully what is to be done and exactly how to go about it.
c. Authoritarian Aggression: Tendency to be on the lookout for, and to condemn, reject, and punish people who violate conventional values.
6. A person who has bad manners, habits, and breeding can hardly expect to be liked and accepted by decent people.
17. An insult to our honor should always be punished.
19. What the youth needs most is strict discipline, rugged determination, and the will to work and fight for family and country.
22. The trouble with letting everybody have a say in running the government is that so many people are just naturally stupid or full of wild ideas.
34. Sex crimes, such as rape and attacks on children, deserve more than mere imprisonment; such criminals ought to be publicly whipped, or worse.
39. There is hardly anything lower than a person who does not feel a great love, gratitude, and respect for his parents.
49. Most of our social problems would be solved if we could somehow get rid of the immoral, crooked, and feebleminded people.
53. One main trouble today is that people talk too much and work too little.
58. Homosexuals are nothing but degenerates and ought to be severely punished.
d. Anti-intraception: Opposition to the subjective, the imaginative, the tender-minded.
16. When a person has a problem or worry, it is best for him not to think about it, but to keep busy with more cheerful things.
30. Some cases of feeblemindedness are caused by overstudy.
45. Nowadays more and more people are prying into matters that should remain personal and private.
53. One main trouble today is that people talk too much and work too little.
e. Superstition and Stereotypy: The belief in mystical determinants of the individual's fate; the disposition to think in rigid categories.
7. Science has carried man very far, but there are many important things that can never possibly be understood by the human mind.
11. Every person should have complete faith in some supernatural power whose decisions he obeys without question.
18. It is more than just chance that Japan had an earthquake on Pearl Harbor Day, December 7, 1944.
24. Some people are born with the urge to jump from high places.
30. Some cases of feeblemindedness are caused by overstudy.
35. People can be divided into two distinct classes: the weak and the strong.
40. Some day it will probably be shown that astrology can explain a lot of things.
46. It is possible that wars and social troubles will be ended once and for all by an earthquake or flood that will destroy the whole world.
50. It's a mistake to trust anybody who doesn't look you straight in the eye.
f. Power and »Toughness«: Preoccupation with the dominance-submission, strong-weak, leader-follower dimension; identification with power figures; overemphasis upon the conventionalized attributes of the ego; exaggerated assertion of strength and toughness.
2. No weakness or difficulty can hold us back if we have enough will power.
5. Any red-blooded American will fight to defend his property.
17. An insult to our honor should always be punished.
19. What the youth needs most is strict discipline, rugged determination, and the will to work and fight for family and country.
32. What this country needs most, more than laws and political programs, is a few courageous, tireless, devoted leaders in whom the people can put their faith.
35. People can be divided into two distinct classes: the weak and the strong.
57. Most people don't realize how much our lives are controlled by plots hatched in secret by politicians.
g. Destructiveness and Cynicism: Generalized hostility, vilification of the human.
10. Human nature being what it is, there will always be war and conflict.
25. Familiarity breeds contempt.
41. The true American way of life is disappearing so fast that force may be necessary to preserve it.
h. Projectivity: The disposition to believe that wild and dangerous things go on in the world; the projection outward of unconscious emotional impulses.
36. Nowadays when so many different kinds of people move around so much and mix together so freely, a person has to be especially careful to protect himself against infection and disease.
45. Nowadays more and more people are prying into matters that should remain personal and private.
46. It is possible that wars and social troubles will be ended once and for all by an earthquake or flood that will destroy the whole world.
52. The wild sex life of the old Greeks and Romans was tame compared to some of the goings-on in this country, even in places where people might least expect it.
57. Most people don't realize how much our lives are controlled by plots hatched in secret by politicians.
i. Sex: Exaggerated concern with sexual »goings-on.«
34. Sex crimes, such as rape and attacks on children, deserve more than mere imprisonment; such criminals ought to be publicly whipped or worse.
52. The wild sex life of the old Greeks and Romans was tame compared to some of the goings-on in this country, even in places where people might least expect it.
58. Homosexuals are nothing but degenerates and ought to be severely punished.
The 19 items from the F scale (Form 78) that ranked highest in order of goodness were retained, in the same or slightly revised form, in the new scale. Thus, statistical differentiating power of the item was the main basis of selection. As stated above, however, the items which came out best statistically were, in general, those which seemed best from the point of view of theory, so that retaining them required no compromise with the original purpose of the scale. Of these items, 5 were changed in no way; revision of the others involved change in wording but not in essential meaning, the aim being to avoid too much uniformity of agreement or disagreement and, hence, to produce mean scores as close as possible to the neutral point.
Given 19 items of known dependability, the task was to formulate 15 additional ones which, singly, met the requirements of good items and which, taken together, covered the ground mapped out according to our theory. Here, criteria other than statistical ones played an important role. In attempting to achieve a maximum of indirectness we not only eliminated items which were too openly aggressive (they had low D.P.'s anyway) but retained, in a slightly revised form, Item 65 (World catastrophe) despite its relatively low D.P. (R.O. 23.5), because it expressed a theoretically important idea and appeared on the surface to be almost completely removed from »race« prejudice and fascism. In the name of breadth, Item 67 (Eye to profit), whose D.P. was not low (R.O. 21), was eliminated because of its too great similarity to the highly discriminating Item 59 (Always war). To cover a great variety of ideas as efficiently as possible, two or more of them were combined in the same statement, e.g., »Any red-blooded American will fight to defend his property« or »... people think too much and work too little.« With attention to these criteria, and to meaningfulness, contribution to the structural unity of the scale, and proper degree of rational justification, 4 items from the F scale (Form 78) whose D.P. rank orders were lower than 19, were revised and 11 new items were formulated to complete the new form. The 34 items, grouped according to the variables which they were supposed to represent, are shown in table 4 (VII).
Table 5 (VII)
Reliability of the F Scale (Form 60)a
Property Group Over-allb
I II III IV V
Reliability .86 .91 .89 .87 .81 .87
Mean (total) 3.32 3.39 3.82 3.74 3.25 3.50
Mean (odd half) 3.41 3.42 4.09 3.78 3.19 3.58
Mean (even half) 3.24 3.36 3.56 3.73 3.28 3.43
S.D. (total) .86 .96 .93 .81 .71 .85
S.D. (odd half) .97 1.03 .99 .77 .83 .92
S.D. (even half) .75 .96 .97 .93 .76 .87
N 47 54 57 68 60 286
Range 1.00–5.50 1.24–5.50 1.82–4.38 2.24–5.62 1.97–5.35 1.82–5.62
a The five groups on which these data are based are:
Group I: University of Oregon Student Women.
Group II: University of Oregon and University of California Student Women.
Group III: University of Oregon and University of California Student Men.
Group IV: Oregon Service Club Men.
Group V: Oregon Service Club Men (A Part only).
b In obtaining the over-all means, the individual group means were not weighted by N.
Reliability of the scale, mean score per item, S.D., and the range of scores for each of the five groups to whom the F scale (Form 60) was given are shown in Table 5 (VII). The reliability of the scale is a considerable improvement over that obtained with Form 78 (.87 as compared with .74); it is as high as that of the shortened E scale (.87 as compared with .86) and much better than the reliability of .70 for the shortened PEC scale. The mean scores are not quite so close to the neutral point as was the case with Form 78 (over-all mean of 3.5 as compared with 3.7); the range and the variability, however, are somewhat greater.9
Inspection of the Discriminatory Powers of the items, as shown in Table 6 (VII), shows once again considerable improvement over Form 78. The mean D.P.F is now 2.15 as compared with 1.80 for Form 78. Three D.P.'s are above 3.0, 18 fall in the range 2.0–3.0, 12 are in the range 1.0–2.0, and only 1 is below 1.0. The mean D.P. in terms of E, 1.53, is notably greater than the mean D.P.A-S, .89, found with Form 78. There are 28 items with a mean D.P.E greater than 1.0; these F items are significantly related to ethnocentrism at the 5 per cent level of confidence or better. Each of the variables that entered into the F scale – Conventionalism, Superstition, etc. – is represented by items that are satisfactorily differentiating.
The correlation between the F scale (Form 60) and E is, on the average, .69. This is a considerable improvement over the results obtained with Form 78, where F correlated .53 with A-S and .65 with E, though it is still not quite as high as its intended functions require.
Table 6 (VII)
Means and Discriminatory Powers of the F-Scale Items (Form 60)a
Item Mean D.P.Fb Rank D.P.Ec Rank Final
D.P.F D.P.E Rankd
(D.P.F
+ D.P.E)
1. (Obedience & respect) 4.86 2.39 (14) 1.52 (17) (13)
2. (Will power) 4.44 2.50 (11) 1.46 (19) (12)
5. (Red-blooded American) 5.49 1.46 (29.5) 1.18 (25.5) (27)
6. (Bad manners) 5.30 1.80 (23) 1.56 (13.5) (22)
7. (Science) 4.98 1.71 (24) 1.32 (23) (25)
10. (War & conflict) 4.46 1.67 (26) 1.70 (10) (21)
11. (Supernatural power) 3.60 2.91 (4) 1.38 (21) (10)
12. (Germans & Japs) 3.71 3.16 (3) 2.83 (1) (1)
16. (Cheerful things) 3.15 2.08 (20.5) 1.18 (25.5) (23)
17. (Honor) 3.14 2.46 (12) 2.34 (4) (7)
18. (Pearl Harbor Day) 2.19 2.51 (10) 1.83 (9) (9)
19. (Discipline & determination) 3.68 3.17 (2) 2.28 (6.5) (3)
22. (Not everybody in gov't.) 2.74 1.46 (29.5) 1.17 (27) (28)
23. (Rebellious ideas) 4.30 2.70 (7) 2.29 (5) (5)
24. (Born with urge) 2.87 2.60 (8) 2.28 (6.5) (6)
25. (Familiarity) 3.30 2.08 (20.5) 1.33 (22) (20)
29. (No sane person) 3.55 2.82 (6) 1.95 (8) (8)
30. (Feebleminded) 1.84 1.43 (32.5) 0.91 (30) (30)
32. (Devoted leaders) 4.49 2.42 (13) 1.43 (20) (15)
34. (Sex crime) 3.43 2.83 (5) 2.52 (3) (4)
35. (Two classes) 1.44 0.73 (34) 0.38 (34) (34)
36. (Infection & disease) 4.80 1.68 (25) 1.03 (28) (26)
39. (Love for parents) 3.16 3.28 (1) 2.56 (2) (2)
40. (Astrology) 2.56 2.15 (17) 1.66 (11) (16)
41. (Force to preserve) 2.48 2.31 (15) 1.56 (13.5) (14)
44. (Bosses outline) 2.46 1.60 (27) 0.50 (33) (33)
45. (Prying) 3.48 2.52 (9) 1.56 (13.5) (11)
46. (Flood) 2.15 1.43 (32.5) 0.94 (29) (29)
49. (Rid of immoral people) 2.74 2.12 (19) 1.56 (13.5) (18)
50. (Mistake to trust) 2.12 1.45 (31) 0.84 (31) (31)
52. (Sex life) 3.18 2.13 (18) 1.50 (18) (19)
53. (Talk too much) 3.87 1.83 (22) 1.24 (24) (24)
57. (Plots) 4.24 1.55 (28) 0.63 (32) (32)
58. (Homosexuals) 2.29 2.20 (16) 1.54 (16) (17)
Mean/Person/Item 3.42 2.15 1.53
a The four groups on which these data are based are:
Group I: University of Oregon Student Women (N = 47)
Group II: University of Oregon and University of California Student Women (N = 54)
Group III: University of Oregon and University of California Student Men (N = 57)
Group IV: Oregon Service Club Men (N = 68)
In obtaining the over-all means, the individual group means were not weighted according to N.
b D.P.F is based on the difference between the high quartile and the low quartile of the F scale distribution.
c D.P.E is based on the difference between the high quartile and the low quartile of the E scale distribution e.g., the D.P.E of 1.52 on Item 1 indicates that the mean of the low quartile on E was 1.52 points lower than the mean of the high quartile on E.
d For each item the sum of D.P.F + D.P.E is obtained. The final rank of an item is the rank of this sum in the distribution of sums for the whole scale.
4. The Third F Scale: Forms 45 and 40
Although the F scale (Form 60) might be described as a fairly adequate instrument, it still had some obvious shortcomings, and it was hoped that these might be removed before the scale was used with numerous groups of subjects. It still contained a number of items so poor statistically that they contributed almost nothing to the purpose of the scale. Also, there were two items (numbers 12 and 18) which, despite their ranking 1 and 9 in order of goodness, had to be dropped in the early fall of 1945 because they had lost their war-period timeliness. More than this, experience had shown that when it came to working with highly diverse groups outside the University, a questionnaire shorter than Form 60 was highly desirable. The other scales having already been reduced to an absolute minimum, the F scale had now to bear some of the pruning. Yet, such pruning had to be extremely judicious if the scale was not to lose in comprehensiveness and if, as was hoped, its reliability and its correlation with E were to be raised. Thirty was the number of items decided upon.
Revision consisted mainly in discarding seven items and adding three new ones. Two excellent items were discarded because, as mentioned above, they were no longer timely. Five items were taken out because they had relatively very low D.P.'s and because other items seemed to fulfill the same functions better: 44 (Bosses outline, R.O. 33), 50 (Mistake to trust, R.O. 31), 30 (Feebleminded, R.O. 30), 22 (Not everybody in government, R.O. 28), and 5 (Red-blooded American, R.O. 27). Several items which ranked low in terms of D.P. were thoroughly revised and kept in the new scale: 35 (Two classes, R.O. 34), 57 (Plots, R.O. 32), and 46 (Flood, R.O. 29).
One of the three new items was the following: »The businessman and the manufacturer are much more important to society than the artist and the professor.« This is a slightly revised version of an item which appeared in the earlier forms of the PEC scale, where it had proved to be discriminatory both in terms of PEC and in terms of E. Since this item expresses rather well both conventionalism and anti-intraception, it appeared to be a promising addition to the F scale.
A second new item was, »It is best to use some prewar authorities in Germany to keep order and prevent chaos.« At the time this item was formulated – late summer 1945 – General Patton had been accused of using German Nazis and extremely conservative bigwigs to help carry out certain phases of the occupation. His critics argued that this was a poor way to secure the cooperation of democratic elements in Germany; his defenders pointed to the necessity for promoting efficiency and preventing chaos. The issue was a lively one; and it appeared that an item bearing upon it might give the high scorer a chance to express his admiration for military authority, his means-centeredness, his preoccupation with order vs. chaos. (That the item did not work very well is probably due to inadequate wording: we had not dared to be sufficiently explicit about which prewar authorities were to be used, so that »prewar« could be taken to mean »pre-Nazi.«)
»Nobody ever learned anything really important except through suffering,« was the third of the new items. Here the temptation – constantly present during each revision of the F scale – to test a new hypothesis, or better, to obtain quantitative data bearing upon a phenomenon which in clinical study had appeared in relation to the general pattern of potential fascism, became too strong. The item was taken from an editorial in a prominent picture magazine, where it had appeared in a context of political reaction. It seemed well adapted to bring out the sado-masochistic theme believed to be prominent in the personality of the high scorer: he believes that he has suffered and, therefore, knows the important things and that those who have not succeeded in raising their status, i.e., the underprivileged, should suffer more if they hope to improve their lot. The item did not work very well, its rank in order of goodness for men being 29. (Its D.P., 1.70, is still significant at the 5 per cent level, however.) It seems that this was partly because many subjects thought it unreasonable (the mean was 2.54), and partly because, where it was agreed with, it probably appealed to different subjects for different reasons: if it tapped the deep-lying sado-masochistic structures in some high scorers, it also appealed to the surface masochism, and perhaps to the intraceptiveness, of some low scorers.
The final F items, grouped according to the variables to which they pertain, are presented in Table 7 (VII).
Table 7 (VII)
F-Scale Clusters: Forms 45 and 40
a. Conventionalism: Rigid adherence to conventional, middle-class values.
1. Obedience and respect for authority are the most important virtues children should learn.
12. A person who has bad manners, habits, and breeding can hardly expect to get along with decent people.
37. If people would talk less and work more, everybody would be better off.
41. The business man and the manufacturer are much more important to society than the artist and the professor.
b. Authoritarian Submission: Submissive, uncritical attitude toward idealized moral authorities of the ingroup.
1. Obedience and respect for authority are the most important virtues children should learn.
4. Science has its place, but there are many important things that can never possibly be understood by the human mind.
8. Every person should have complete faith in some supernatural power whose decisions he obeys without question.
21. Young people sometimes get rebellious ideas, but as they grow up they ought to get over them and settle down.
23. What this country needs most, more than laws and political programs, is a few courageous, tireless, devoted leaders in whom the people can put their faith.
42. No sane, normal, decent person could ever think of hurting a close friend or relative.
44. Nobody ever learned anything really important except through suffering.
c. Authoritarian Aggression: Tendency to be on the lookout for, and to condemn, reject, and punish people who violate conventional values.
12. A person who has bad manners, habits, and breeding can hardly expect to get along with decent people.
13. What the youth needs most is strict discipline, rugged determination, and the will to work and fight for family and country.
19. An insult to our honor should always be punished.
25. Sex crimes, such as rape and attacks on children, deserve more than mere imprisonment; such criminals ought to be publicly whipped, or worse.
27. There is hardly anything lower than a person who does not feel a great love, gratitude, and respect for his parents.
34. Most of our social problems would be solved if we could somehow get rid of the immoral, crooked, and feebleminded people.
37. If people would talk less and work more, everybody would be better off.
39. Homosexuals are hardly better than criminals and ought to be severely punished.
d. Anti-intraception: Opposition to the subjective, the imaginative, the tender-minded.
9. When a person has a problem or worry, it is best for him not to think about it, but to keep busy with more cheerful things.
31. Nowadays more and more people are prying into matters that should remain personal and private.
37. If people would talk less and work more, everybody would be better off.
41. The businessman and the manufacturer are much more important to society than the artist and the professor.
e. Superstition and Stereotypy: The belief in mystical determinants of the individual's fate; the disposition to think in rigid categories.
4. Science has its place, but there are many important things that can never possibly be understood by the human mind.
8. Every person should have complete faith in some supernatural power whose decisions he obeys without question.
16. Some people are born with an urge to jump from high places.
26. People can be divided into two distinct classes: the weak and the strong.
29. Some day it will probably be shown that astrology can explain a lot of things.
33. Wars and social troubles may someday be ended by an earth-quake or flood that will destroy the whole world.
f. Power and »Toughness«: Preoccupation with the dominance-submission, strong-weak, leader-follower dimension; identification with power figures; overemphasis upon the conventionalized attributes of the ego; exaggerated assertion of strength and toughness.
2. No weakness or difficulty can hold us back if we have enough will power.
13. What the youth needs most is strict discipline, rugged determination, and the will to work and fight for family and country.
19. An insult to our honor should always be punished.
22. It is best to use some prewar authorities in Germany to keep order and prevent chaos.
23. What this country needs most, more than laws and political programs, is a few courageous, tireless, devoted leaders in whom the people can put their faith.
26. People can be divided into two distinct classes: the weak and the strong.
38. Most people don't realize how much our lives are controlled by plots hatched in secret places.
g. Destructiveness and Cynicism: Generalized hostility, vilification of the human.
6. Human nature being what it is, there will always be war and conflict.
43. Familiarity breeds contempt.
h. Projectivity: The disposition to believe that wild and dangerous things go on in the world; the projection outwards of unconscious emotional impulses.
18. Nowadays when so many different kinds of people move around and mix together so much, a person has to protect himself especially carefully against catching an infection or disease from them.
31. Nowadays more and more people are prying into matters that should remain personal and private.
33. Wars and social troubles may someday be ended by an earth-quake or flood that will destroy the whole world.
35. The wild sex life of the old Greeks and Romans was tame compared to some of the goings-on in this country, even in places where people might least expect it.
38. Most people don't realize how much our lives are controlled by plots hatched in secret places.
i. Sex: Exaggerated concern with sexual »goings-on.«
25. Sex crimes, such as rape and attacks on children, deserve more than mere imprisonment; such criminals ought to be publicly whipped, or worse.
35. The wild sex life of the old Greeks and Romans was tame compared to some of the goings-on in this country, even in places where people might least expect it.
39. Homosexuals are hardly better than criminals and ought to be severely punished.
Reliability of the scale, mean score per item, S.D., and range for each of the fourteen groups (total N = 1518) taking Form 40 and/or 45 are given in Table 8 (VII). The average of the reliability coefficients is .90, their range .81 to .97. Not only is there a slight improvement in reliability over Form 60 (av. r = .87) and a very marked improvement over the original Form 78 (av. r = .74), but the scale has now been developed to a point where it meets rigorous statistical requirements. A reliability of .90 may be interpreted to mean that the scale can place individuals along a dimension – in this case a broad and complex dimension – with a small margin of error. In other words, the score attained by an individual can be relied upon in the sense that chance errors of measurement have been minimized, so that in a repetition of the scale, at a time when political-socioeconomic conditions were generally the same as before, his new score would either be the same as his first or fall within narrow limits above or below it. The degree of reliability attained here is within the range which characterizes acceptable intelligence tests.
Table 8 (VII)
Reliability of the F Scale (Forms 40 and 45)
Group N Reliability Mean S.D. Range
Form 40:
George Washington
Univ. Women 132 .84 3.51 .90 1.2–5.4
California Service
Club Men 63 .94 4.08 1.03 1.8–7.0
Middle-Class Men 69 .92 3.69 1.22 1.3–6.7
Middle-Class Women 154 .93 3.62 1.26 1.1–6.7
Working-Class Men 61 .88 4.19 1.18 1.8–6.9
Working-Class Women 53 .97 3.86 1.67 1.3–6.6
Los Angeles Men 117 .92 3.68 1.17 1.1–6.0
Los Angeles Women 130 .91 3.49 1.13 1.2–5.8
Meana 779 .91 3.76 1.20 1.3–6.4
Form 45:
Testing Class Women 59 .89 3.62 .99 1.3–5.9
San Quentin Men Prisoners 110 .87 4.73 .86 2.0–6.8
Psychiatric Clinic Womenb 71 .94 3.69 1.30 1.0–6.3
Psychiatric Clinic Menb 50 .89 3.82 1.01 1.7–5.9
Mean 290 .90 3.96 1.04 1.5–6.2
Form 40 and Form 45:
Employment Service
Men Veterans 106 .89 3.74 1.04 1.2–5.8
Maritime School Men 343 .81 4.06 .77 1.6–6.1
Meana 449 .85 3.90 .90 1.4–5.9
Over-all mean 1518 .90 3.84 1.10 1.4–6.3
a In obtaining the combined group means, the individual group means were not weighted by N.
b Due to a substitution of forms, the F scale for the Psychiatric Clinic subjects contained only 28 items.
The means, though they vary from one group to another (a matter to be discussed later), are fairly close, on the whole, to the neutral point. As is to be expected from administration of the scale to a great variety of subjects, the range and the S.D. are greater than in previous forms. While no distribution curves have actually been made, the scatter diagrams indicate that they would be fairly normal in form (symmetrical but slightly platykurtic).
a. INTERNAL CONSISTENCY. The Discriminatory Powers of the scale items, as shown in Table 9 (VII), are considerably higher on the average (2.85) than in the case of Form 60 (2.15). All of the items differentiate significantly between the high and the low quartiles. It is to be noted that numerous items taken over without change from Form 60 work much better here than in that instance. This is probably due in part to the fact that the diverse groups given Form 45–40 included more extreme scorers and in part to improvement of the scale as a whole: a good item differentiates the more sharply between the upper and lower quartiles the more successfully the total scale distinguishes individuals who are actually extreme with respect to the trends being measured.
The fact that the D.P.'s are somewhat higher, on the average, for women than for men is deserving of some comment. This phenomenon would seem to be connected with the fact that there were three groups of men – Maritime School, San Quentin Inmates, and Working-Class Men – in whose cases the reliability of the scale was relatively low (.81-.88). Since these groups of men were less educated than most of our subjects, there is considerable likelihood that they failed to understand some of the scale items, a circumstance that would work against high D.P.'s as well as against reliability. Moreover, these are the three groups who, of all those studied, obtained the highest mean scores. It can be inferred from this that there was too much general agreement with some of the items, something which, as we have seen, tends to lower the D.P. This raises the question of whether we did not encounter in these groups not only more extreme manifestations of potential fascism than had been anticipated but also patterns of prefascist personality trends that the F scale did not adequately cover. Most of the work that went into the construction and revision of the scale was performed with groups of subjects in which the high scorers were, in the main, highly conventional. The procedure of retaining items which differentiated best within these groups was probably not the best one for constructing an instrument which would work with maximum efficiency in groups where tendencies to psychopathy and delinquency were much more pronounced. This is a matter to be discussed in more detail later.
Despite the absolute differences in the D.P.'s between men and women, items which work well for one sex tend, in general, to work well for the other. The correlation between the D.P. rank orders for the men and those for the women is .84. This is sufficient justification for averaging the D.P.'s of the two groups to obtain an over-all »order of goodness« for each item. Since the differences between men and women, in the present context, are probably as great as the differences between any two groups of the same sex in the present sample, it is highly probable that a correlation between the D.P. rankings of any two such groups would be in the neighborhood of .84. There appear to be no general or systematic differences between the items which work better for men and those which work better for women.
Mean scores for the men's groups are somewhat higher on the average than mean scores for the women's groups. This phenomenon would seem to be due primarily to the three male groups discussed above whose scores are particularly high. If men and women of the same socioeconomic class are compared, the means are not significantly different. Moreover, items which appeal most strongly to the men are much the same as those which appeal most strongly to the women, the rank-order correlation between the means for men and those for women being .95.
b. CORRELATIONAL ANALYSIS. As a part of an independent investigation, the E, PEC, and F scales (from Forms 40 and 45) were administered to 900 students in an Elementary Psychology Class at the University of California. It was decided not to include the data from this new college group among the general results of the present study because the total sample of subjects was already weighted too heavily on the side of young and relatively well-educated people. However, the 517 women from this psychology class constitute the only group in whose case the scales were subjected to an item-by-item correlational analysis.10 The results of this analysis will be summarized here.
Table 9 (VII)
Means and Discriminatory Powers of the F-Scale Items (Forms 40 And 45)a
a These data are based on all fourteen groups taking Forms 40 and 45 (see Table 8 (VII)).
Each item of the F scale was correlated with every other item. The average of the 435 coefficients was .13, the range – .05 to .44.11 In addition, each item was correlated with the remainder of the scale, the mean r here being .33, the range .15 to .52. In the case of the E scale the mean interitem r was .42, and the mean item-total score r, .59. Whereas the E scale has about the same degree of unidimensionality as do acceptable intelligence tests (in the case of the 1937 Stanford-Binet Revision the average interitem r is about .38, the average item-total score r, .61), the F scale rates considerably lower in this regard. Despite the scale's relative lack of surface homogeneity, however, we are justified in speaking of an F pattern or syndrome, for the items do »hang together« in the sense that each is significantly correlated with the scale as a whole. It will be recalled in this connection that in constructing the F scale two purposes were held in mind: (a) to seek over a wide area for diverse responses that belonged to a single syndrome, and (b) to construct an instrument which would yield a reliable prediction of scores on E. It is clear that the first purpose has been in large part realized, although the search for additional items that would help characterize the F syndrome could be continued with profit. The fact that the individual F items correlate .25 on the average with the total E scale augurs well for the fulfillment of the second purpose – a matter to which we shall turn in a moment.
Proof that the variables or groups of items used in thinking about the F scale are not clusters in the statistical sense, is contained in the data from the present group of 517 women. Although the items within each of the Form 45 F-clusters tend to intercorrelate (.11 to .24), the items in any one cluster correlate with one another no better than they do with numerous items from other clusters. We are justified in using these clusters, therefore, only as a priori aids to discussion.