From time to time, I like to read old books to get a sense of what people were interested in many decades ago. Often this produces surprises, where you find out that people were already talking about the same thing you are thinking about now, but in different terms and no one seems to know about it. Other times, people are talking about stuff that at most results in a yawn when reading (as noted by Meehl). I’ve been meaning to read some more books by Eysenck (who wrote a lot). So this book is originally from 1954, a very long time ago, even before the Jensen 1969 events. In general I think it is definitely worth reading, and Eysenck’s down to earth style makes it a quick read as well. It has many tables of data from old US and UK elections, which are interesting in themselves. Some quotes and comments.

In discussion of pitfalls of surveys:

Understanding of words. A well known story deals with the surprise experienced by an American Government Agency when it was found that among Southern Negroes only a very small proportion voted in favour of levying tax on profits. An investigation on the spot indicated that these Negroes, whose only reading had been the Bible, could find no justification therein for taxing prophets!

A sort of specific instance of the curse of knowledge, vocabulary in this case.

On quantifying traits:

The second way in which psychologists differ from laymen in their approach is in their insistence on obtaining some numerical estimate of the degree to which people tend to manifest the same trait in different circumstances. As an example let us take the hypothetical trait of ‘persistence’. If we wanted to measure this trait we would first of all design a number of objective situations in which our subjects could demonstrate their persistence. Thus as our first test we might ask them to pick up a dumb-bell and hold it out sideways as long as they could, the time being a measure of their degree of persistence. As a second test we might give them a jigsaw puzzle from which some of the pieces had been removed and others substituted, so as to make the task impossible of solution; the length of time during which they continued with this activity would be taken as a measure of their persistence.

An interesting if not so practical approach to objective/ratio measurement of personality.

On confirmation bias:

There is considerable proof regarding the generalized nature of responses in the attitude field. We think of attitudes as determining our actions and our words. There is little doubt that they also determine the way in which we perceive things, the way in which we learn and remember things, and the manner of our reasoning. Thus, it has been shown, for instance, that when groups of pro- and anti-Communist students are made to listen to a pro-Communist argument, the pro-Communist students learn and remember the ideas contained in it much better than do the anti-Communist students. When the tables are turned, however, and both groups are made to listen to an anti-Communist argument, the anti-Communist students are superior in both learning and retention. Similarly, when students were made to learn paired associates like Stalin-Devil, and were later asked to reproduce the second word on being given the first, recall and learning were again found to be congruent with measured attitude. Thus, what we learn and what we remember depends in part at least on our pre-existing attitudes towards the material with which we are presented.

It’s hard to avoid confirmation bias consciously when it is literally easier to remember pro-information. Do these things replicate?

On intelligent people and politics:

We would suggest, therefore, that in general no conclusions can be drawn from the large mass of data which has been accumulated. Until studies of this type are carried out in different cultures and in different periods it will be impossible to generalize from data collected in one particular culture in one particular, very narrowly circumscribed, period of time. Certainly the data do not give any comfort to partisans in either camp; highly intelligent people have held beliefs located at any point of the radical-conservative continuum, and the very slight tendencies observed for the more intelligent to hold one view or the other can certainly not be generalized to argue in favour of the correctness of any particular political belief.

More or less the same as now.

On physiognomy:

An early experiment by Sir Charles Goring may be of interest in this connection. He was investigating the well-known theories of Lombroso, who believed that criminals could be recognized by the presence of certain physical characteristics which he called ‘stigmata of degeneration’. Goring thought that this belief was based merely on stereotyped thinking and set about to prove his point. He had an artist draw from memory portraits of inmates of a penal institution in London. Using a technique invented by Sir Francis Galton he made a composite photograph of these drawings and found that this approached very closely to Lombroso’s view. Then he took actual photographs of the same criminals and had another composite photograph made of these. This showed no evidence of Lombroso’s ‘criminal type’, and bore no resemblance at all to the one based upon the drawings. Lombroso and the artist clearly did not see what was physically there, but had their perception determined by previously acquired attitudes of a stero- typed character.

I am quite certain that machine learning will confirm these things however. Everybody has a sense of whether someone looks dodgy or shady, which is basically an intuitive neural network measure of likelihood of someone being trouble to interact with. There are a few studies on the topic (this one most famously), but needs more work! I would do it (using OKCupid data) if I wasn’t so busy with the race and IQ genomics work.

On nature and nature:

However, again it is necessary to bear in mind possible alternative formulations and hypotheses. We cannot rule out the possibility that a child’s extraversion is caused by authoritarian child rearing practices in the parents, but we must also consider the possibility that an extraverted pattern of behaviour is largely based on heredity, so that the true sequence of events would read (i) extraverted parents develop tough-minded attitudes; (2) tough- minded attitudes lead to authoritarian child rearing practices; (3) the children inherit the extraverted personality pattern, which in turn leads them to adopt tough-minded authoritarian attitudes. Thus, here again there would appear a correlation between child rearing practices and children’s attitudes which could not be regarded as directly causal but as purely fortuitous. The fact, then, that correlations are found between child rearing practices and children’s attitudes does not help us to decide between these different hypotheses; such a correlation could certainly not be accepted as proof of the environmentalist hypotheses.

Early discussion of this problem.

Polling versus simple methods:

Perhaps the most widely used proof in this connection is the agreement between poll prediction and voting behaviour. We have already noted in an earlier chapter the smallness of the error made by the British Institute of Public Opinion in predicting voting behaviour during three elections in this country. In America, Gallup has reported that the average error of prediction from 1935-47 was four percentage points; this average relates to over 300 election predictions in the United States. If we look only at presidential elections, Mosteller gives a table showing that the errors of prediction in percentage points for the 1936, 1940, 1944, and 1948 elections were 6-5, 30, 2-3, and 5-3 per cent, giving an average of 4-3. This may sound reasonable, but before we can estimate the success of forecasting which is implied by an average error of a given magnitude, we need a base line against which to compare the results. Such a base line is furnished by what is called persistence forecasting. This method, taken over from weather forecasting, is a simple routine method in which the forecast for the next occasion is simply made in terms of what happened last time. In weather forecasting one would simply predict that to-morrow will be exactly like to-day. In election forecasting, it would mean predicting that each state will have the same Democratic percentage of the major party vote that it had in the previous presidential election year. This ‘persistence’ method is quite mechanical, requires no new investigation, and may thus serve as a useful base line in terms of which we can estimate the accuracy of the polls. Mosteller presents a table of the errors in persistence forecasting as compared with errors in the forecasts made by Gallup and Crossley, and concludes that ‘taken as a whole, it cannot be said that the polling forecasts in the past four presidental elections have a very distinguished record compared to persistence forecasts, which were as good or better in three out of four elections. The implication here is not that polling is no better, or not much better, than persistence forecasting but rather that polling has not yet proved its superiority in election forecasting under the conditions obtaining during the last four presidential elections.’

Simple methods good!

On art preferences and politics:

No further work was done by the author in this connection, but the topic was taken up independently by Barron and Welsh, whose work lends considerable support to our general hypothesis. The first step taken by these writers was the construction of a scale for the measurement of complexity preferences. This scale consisted of 65 drawings of the kind illustrated in Figure 31. Very stable preferences were found for either the complex or the simple type of drawing and the whole scale had a very high reliability.22 It was used by the authors in a whole series of investigations, the results of some of which are of considerable interest here.

In the first place it was found that those who preferred the simple figures tended to prefer in their artistic preference judgments themes involving religion, authority, aristocracy and tradition, while those who preferred the complex figures preferred in their artistic judgments those pictures which were radically experimental, esoteric, primitive and naive. On the personality side simplicity was found related with masculinity and the rejection of soft, gentle and effeminate behaviour. This tendency, which was found both in questionnaires and ratings, fits in well with our hypothetical identification of simplicity and the attitudes in the tough-minded quadrant.

At the other end, the preference for complexity was found associated with originality, artistic expressions and excellence of aesthetic judgments. Again these correlations were found both in terms of ratings and of separate tests, and again these results are in line with our hypothesis.

The tendency of preference for complexity to be related to originality would lead one to hypothesize that preference for simplicity would tend to go with rigidity and this was indeed the case. Ratings of rigidity, which was defined as ‘inflexibility of thought and manner; stubborn, pedantic, unbending, firm’, correlated -35 with preference for simplicity and a similar correlation was observed when a questionnaire was substituted for the rating.

A trait related to both originality and rigidity is constriction, as opposed to impulsiveness. A comparatively strong tendency was found for preference for complexity to go with impulsiveness (50) and for constriction to go with preference for simplicity (42).

In line with the Jungian hypothesis we should expect a preference for tendency was found for preference for complexity to go with impulsiveness (50) and for constriction to go with preference for simplicity (42).

Among neurotic subjects such a tendency was in fact observed by Eysenck, and Barron, using questionnaires among university students, corroborated the finding. He showed that preference for simplicity correlated .30 with a hysteria scale, while complexity correlated .34 with a measure of anxiety. As Barron points out ‘preference for the complex in the psychic life makes for a wider consciousness of impulse while simplicity, when it is preferred, is maintained by a narrowing of that consciousness…. To tolerate complexity one must very often be able to tolerate anxiety as well.’

More directly relevant than these findings is the fact that preference for complexity was found to be negatively correlated with ethnocentrism. In addition to these results, which were based on questionnaire answers, staff ratings showed that preference for simplicity correlated 47 with conformity and -29 with submissiveness defined as ‘deference, willingness to be led, compliance, over- ready acceptance of authority’. These results directly and strongly support our original hypothesis.

In yet another study Barron and Welsh compared responses to certain attitude questions of a group of students having respectively very high and very low scores on complexity. The following questions were answered ‘true’ by high scorers on complexity: 1. The unfinished and imperfect often have greater appeal for me than the completed and the polished. 2. I could cut my moorings . . . quit my home, my parents and my friends . . . without suffering great regrets. 3. Politically I am probably something of a radical. 4. I think I take primarily an aesthetic view of experience. 5. I would enjoy the experience of living and working in a foreign country. 6. Many of my friends would probably be considered unconventional by other people. 7. Some of my friends think that my ideas are impractical if not a bit wild. 8. I enjoy discarding the old and accepting the new. 9. When someone talks against certain groups or nationalities I always speak up against such talk, even though it makes me unpopular.

In contrast, the following questions were answered ‘true’ by low scorers on complexity 1. I don’t like modern art. 2. Disobedience to the government is never justified. 3. Perfect balance is the essence of all good composition. 4. Straightforward reasoning appeals to me more than metaphors and the search for analogies. 5. It is a pretty callous person who does not feel love and gratitude towards his parents. 6. Things seem simpler as you learn more about them. 7. I much prefer symmetry to asymmetry. 8. Kindness and generosity are the most important qualities for a wife to have. 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. 10. It is a duty of the citizen to support his country, right or wrong. 11. Barring emergencies I have a pretty good idea what I will be doing for the next ten years. 12. I prefer team games to games in which one individual competes against another.

Barron sums up the main results of his work in the following words: ‘Preference for simplicity is associated with social conformity, respect for custom and ceremony, friendliness towards tradition, somewhat categorical moral judgments, and undeviating patriotism and suppression of. . . troublesome new forces. . . . This last item is almost prototypical of the simple person’s orientation towards repression as a psychic mechanism . . . complexity goes along with artistic interests, unconventionally, political radicalism, strong cathection of creativity as a value and a liking for change.’ ‘It seems evident that, at its best, preference for simplicity is associated with personal stability and balance, while at its worst it makes for categorical rejection of all that threatens disorder and disequilibrium. In its pathological aspect it produces stereotyped thinking, rigid and compulsive morality, and hatred of instinctual, aggressive and erotic forces which might upset the precariously maintained balance.’

So, Noah Carl’s funny finding is in fact quite old. Just forgotten!

Reasonable certainty:

Thus, we can never be certain that our predictions will be accurate within any given range. However, we can be reasonably certain of the degree of accuracy which we shall have obtained. This concept of reasonably certain requires more accurate statement, and it is usually taken to mean that our prediction would be borne out in 997 cases out of 1,000 and falsified only three times out of 1,000. In other words, the odds against would only be three in a 1,000.

Quite a lot different from the p<.05 standard, or even p<.10 in economics.

On stereotype accuracy:

Hofstätter reports a different study in which stereotypes were found to give an accurate picture of reality. He reports that there exists in Austria a stereotyped view of the intelligence of the inhabitants of the eight federal States of that country, the inhabitants of Vienna being at the top, those of Salzburg second, those of Lower Austria third, those of Upper Austria fourth, followed by those of the Tyrol, Styria, Carinthia, and Burgenland. During the war it was possible to compare intelligence test results for recruits coming from these various states. The results are shown in Figure 42 and it will be seen that while the differences are not very large, they agree exactly with the stereotyped order, with Vienna at the top and Burgenland at the bottom. Thus, stereotypes need not be incorrect but may have some factual basis.

Seems easy to do some modern replications of. I don’t think Jussim knows about this one either. It is not cited in the 2012 book. This might be the first such demographic stereotype accuracy study? There is no particular reference for this claim, but the following studies are given in reference section:

  • Hofstaetter, P. R. Die Psychologie der qffentlichen Meinung. Wien: Braumuller, 1949.
  • The actuality of questions. Int. J. Opin. Attit. Res., 1950, 4, 16-26.
  • Importance and actuality. Int. J. Opin. Attit. Res., 1951, 5, 31-52.
  • A factorial study of prejudice. J. Person., 1952, 21, 228-39.

On finding the right person to recruit:

Strong, although using an Inventory constructed on similar principles, altered the scoring system and the interpretation to such an extent that his work bears no real relation to that of the earlier students. Essentially, his argument runs as follows: People who are successfully employed in a given occupation tend to have a certain pattern of interests, i.e. if 500 of them fill in a questionnaire of this type they will tend to distribute their answers in accordance with a certain pattern which will differ from the pattern shown by a group of 500 people in some other occupation. A person’s likelihood to be successful and happy in a given occupation will depend, not on his saying that he likes that occupation very much, but on his giving a pattern of answers which resembles closely that given on the average by members of the profession to which he aspires. This was the a priori argument put forward by Strong, and there is overwhelming evidence to show that in its essential points his argument is correct. This is not the place to summarize the evidence; suffice it to say that for vocational guidance, no other test has been found more successful in predicting a person’s likelihood of success than the Strong Inventory. (We are excepting from this statement all test of ability, of course, as lack of ability is obviously something that cannot be compensated for in any way whatever. The statement applies to all efforts advising candidates possessing the requisite ability as to which of several professions they would be most likely to succeed in.)

This is essentially a sort of nearest neighbor approach to recruitment.

On theories in science:

A theory in science has two main functions. In the first place, it serves to organize and structure a variety of apparently unrelated facts, relating them all together in a system the properties of which can be deduced from some more general principle or law. In the second place, it serves to suggest deductions from such principles or laws which may lead to the discovery of new and hitherto unknown facts, and which may be used to support or disprove the original theory. Usually, there is little in the way of verifiable theory to be found in the early stages of a science, and certainly the psychological study of attitudes is no exception to this rule. There are thousands of empirical studies, but very few attempts to relate these together in a coherent scheme; thousands of isolated results, but no general agreement even on the concepts to be used in discussing and ordering these results. We have already seen in an earlier chapter how divergent are the definitions of such widely used terms as ‘attitude’, ‘opinion’, and ‘ideology’; definitions are almost as numerous as writers on the subject. Such a state of affairs is not compatible with a reasonable rate of advance in knowledge and understanding, and it is not to be wondered that this lack of theory has given rise to many pseudo-problems which befuddle thinking in this field.

Right on.

Tender minded and tough minded:

A better name for this dimension might perhaps be a set of terms taken from a book by W. James, where he refers to two opposed types of temperament leading to opposed philosophical beliefs as the ‘tender-minded’ and the ‘tough-minded’ respectively. As we shall make much use of this dichotomy, a brief quotation from James will make its meaning clearer. James starts his discussion on pragmatism by pointing out that philosophical systems are often influenced or determined by the temperament of their authors. He goes on to say that ‘the particular difference of temperament that I have in mind … is one that has counted in literature, art, government, and manners as well as in philosophy. In manners we find formalists and free-and- easy persons. In government, authoritarians and anarchists. In literature, purists or academicals, and realists. In art, classics and romantics. You recognize these contrasts as familiar; well, in philosophy we have a very similar contrast expressed in the pair of terms “rationalist” and “empiricist”, “empiricist” meaning your lover of facts in all their crude variety, “rationalist” meaning your devotee to abstract and eternal principles.’

On personality of communists and fascists:

Eysenck spends a lot of time discussing these, and their similarity. Most quotable probably is this very early use of a support vector machine in political science.

On mental rigidity:

Related to rigidity is the concept of intolerance of ambiguity intoduced by Frenkel-Brunswik. The term is almost self-explanatory. The rigid person attempts to gain security by grouping concepts into rigid categories of black and white, and refuses to admit intermediate shades of grey. The non-rigid, tolerant person can admit these finer grades and is less liable to form rigid dichotomies of good and bad, we and they, in-group and out-group.

Several measures of this tendency have been proposed such as, for instance, the Dog-Cat test used by Coulter. This test, reproduced in Figure 40 shows eight drawings of a dog turning slowly and by degrees into a cat. The hypothesis is that when these drawings are presented seriatim to the subject with the request that he should say what each drawing represents, then the rigid person would continue to cling to the original ‘dog’ concept long after this concept has objectively failed to account for all the observed details. According to this theory he would be forced into this rigid adherence to the original concept by his intolerance of the ambiguity introduced in the intervening pictures.


On aesthetic judgment:

Our first two methods of proof in support of the hypothesis that extraversion is a determining factor in the tough-minded attitude have been fairly direct and straightforward. Our third proof is somewhat more indirect and complex and will show how very wide are the ramifications of this temperamental variable. Historically we may begin with an observation made by the author during an intensive study of aesthetic judgments of a great variety of materials such as landscape paintings, sculptures, photographs, carpets, silverware, poems, odours, polygonal figures, and so forth. Rankings of the objects in each class by the subjects taking part in the experiment were intercorrelated, and it was found that there was considerable agreement between judges on the aesthetic value of the objects judged. It was also found that people who were good judges in one field also tended to be good judges in another field. By the term ‘good judges’ we mean in this connection nothing more than that they agreed particularly strongly with the average judgments. This use of the term will probably be disputed by aestheticians; however, there is ample evidence to justify its use, even from their own standpoint. (This evidence has been discussed in Structure of Human Personality and as it is not relevant to our present purpose we shall not dwell on it here.)

Very much in line with a sort of objective/intersubjective account of aesthetic judgment, even with a general factor of skill!


Leave a Reply