Thursday, 2 October 2014

Are lefties clever, or just grasping?

In my life, based on personal observation, the brightest people always seemed to incline to the political left, which in the UK meant the Labour Party. I did not conduct any proper sampling, and merely deduced opinions from the nature of their conversation, since outright enquiry about politics was considered unseemly in a university setting. Despite the very weak methodology, this was my general, anecdotal impression.

I believed all this because it was true to my experience. Kahneman and Tversky had yet to sail into my consciousness. I failed to notice that I had not sampled business leaders, lawyers, engineers and City of London entrepreneurs and investors, simply because I did not know many. The availability heuristic was alive and well, and distorted my perceptions.

Later life experiences suggested that, after graduating, bright leftists turned left into education and public service, bright rightists turned right into the professions and business. I turned left with the former, and assumed they represented most of the world. I noticed much later that the Right had thinkers as well as the Left.

More recent work has shown that social scientists are generally of leftist persuasion, whichever way that is labelled in each nation. Can we say whether, as a general rule, the brightest tend towards the Left?

Idan Solon  How intelligence mediates liberalism and pro-sociality. Intelligence 47 (2014) 44–53

Solon has looked at this question largely in the US context, in which the Left are Liberals (Democrats) and the Right are Conservatives (Republicans). Solon does include data from the Netherlands 2004, South Africa 1990, Brazil 2012, Germany 2012, so the phenomenon is not restricted to the USA. His analysis is that the relationship between intelligence and leftism is U shaped: both the brightest and the dullest are Leftists, the middling intellects Rightist. If so, it seems I had noticed only that the brightest were indeed Left, but ignored the fact that the least clever were also Left. That is, if the finding generalises widely.

Individuals of most or least intelligence consistently orient further left politically than those of middle intelligence, producing a U-shaped curve.  The lower intelligence group orients toward the left because it is populated disproportionately by leftist beneficiaries (e.g., economic and racial minorities). The leftists vote leftist strictly on issues on which they are beneficiaries. Unlike the lower intelligence extreme, the higher intelligence extreme is broadly leftist, self-identifies as liberal, and is strongly pro-social overall.

Solon describes leftists as: characterized by a tendency to demonstrate consideration toward the less represented, whether the less represented comprise minority demographic segments (e.g., homosexuals on gay marriage, pregnant women on abortion, convicted criminals on capital punishment, the poor on fiscal policies and health care, immigrants on immigration, citizens of other countries on foreign policy issues) or less represented alternatives to religious, governmental, or cultural orthodoxy.

In the United States, Democrats draw support partly from the well educated and affluent liberals, who orient consistently toward the left across all issues, and partly from “conservative and disadvantaged Democrats,” who are disproportionately populated by racial and economic minorities, are characterized by high religiosity and low education, and orient toward economic liberalism (i.e., in support of government interventions to assist the poor) but social conservatism (Pew Research Center, 2005). These latter Democrats vote for a leftist party (and are therefore considered leftists) but tend not to identify as liberals, who, by contrast, are broadly leftist.

The U-shaped curve explored below relates intelligence and leftism, with the lower intelligence extreme predominately leftist but not liberal, and the higher intelligence extreme predominately liberal (i.e., consistently leftist across a range of issues). An analogous U-shaped curve appears to characterize public opinion in numerous countries, though the names denoting leftist parties and ideologies are variable.

Openness/Intellect is the only Big Five personality trait that correlates consistently with general intelligence (DeYoung, 2011) and is also the trait that correlates most strongly, by far, with both social and economic liberalism (Gerber, Huber, Doherty, Dowling, & Ha, 2010).

Liberalism in academia: College professors have a very high average IQ (Dutton & Lynn, 2014) and orient toward the left across a range of academic disciplines. Left–right ratios of roughly 4 to 1 have been documented for professors overall, with those in the humanities consistently the most liberal and those in economics and business consistently the least liberal (Dutton & Lynn, 2014; Gross, 2013; Rothman & Lichter, 2009). In comprehensive inquiries into professors' liberalism, Gross (2013) and Gross and Fosse (2012) considered six commonly advanced hypotheses, including the argument that those who exhibit intellectualism lean toward the left, before offering one of their own: jobs in academia have become politically typed, thus discouraging conservatives from applying for them. They acknowledged offering no direct empirical support for their hypothesis.

Here is a scary table, the only one in the paper. Some famous, highly prized US academics have made public donations to political parties, and these donations are tabulated according to where the recipients fall on the Left-Right political dimension.


The Right gets barely a penny, the exclusive Left the overwhelming majority of the donations. Prized minds incline leftwards. (The MacArthur grant might be run by a Soviet of socialists, but surely it cannot be the case that all of these august bodies are covert fellow-travellers of Marxism-Leninism? Are not even Scrabble and Bridge sacrosanct?)

However, the Solon paper is a bit thin on actual intelligence measures. Ability is calculated on the basis of achievement in terms of academic prizes like Nobels (a good measure) years of education (a reasonable but imprecise measure) and social and occupational status (fairly reasonable but more imprecise). For example, the Benbow and Lubinski papers on mathematically precocious youth would be a prospective sample worth studying, though I cannot find anything on political orientation at the moment. David Lubinski might be able to supply something, in which case I will add it in.

So, this thesis proposes that bright people are leftist by conviction and dull people leftist by expediency. Eysenck argued as much in 1954, saying that the working class were torn between: the material advantages of Leftism in the shape of the Labour Party; and the ideological advantages of tough minded Rightism in the shape of the hang-them-and-flog-them Conservatives, to which they naturally inclined. Women, being more anxious and conservative usually swung the vote for the Conservatives.

A key point: if most of social science is produced by Leftists then, although it should not distort the truths revealed by nature, it might do so in precisely in the same way same some enthusiastic Leftists level accusations against their opponents: telling them that their innermost Rightist preferences reveal them to be biased.

In Prof Tony Flew’s (1975) “Thinking about thinking” he describes (4.21) these arguments as “You would believe that, because you are a Christian or Marxist or Conservative….” That may be true as a general observation about a cluster of opinions, but it is not a refutation of a particular argument. A Marxist might believe that particular socialised farming practices would benefit humanity. They might be right or wrong, but the proposition cannot be refuted because of their political inclinations.  Stephen Jay Gould was not proved wrong because he was a Marxist, but because his insinuations about faked cranial volumes were shown to be false.

So, all of social science could be Leftist but right. That would depend on all arguments being tested to destruction. As Socrates observed, we must follow the argument where ever it leads.


  1. The left engages in a sophisticated form of magic, and requires some intelligence due to their dependence on abstraction. It doesn't take much intelligence to think marriage is what it is, and the state can't really change the definition, for instance. It takes more intellectual effort to believe the deceptions. The left gets a subsection of the low IQ voters by buying them, not by convincing them of anything.

  2. I've long noticed the same things, James. Of course, what these findings ultimately imply is that if we shift our axes 45 degrees to separate social and economic issues, social leftism should show a positive linear correlation with IQ. More tentatively, and probably more weakly, we may expect economic leftism to show a negative linear correlation with IQ.

    But perhaps a better question might be whether economic and social attitudes really do cluster together meaningfully. Put another way, is there really a libertarian/authoritarian axes lying orthogonal to the left/right axis? Libertarians always insist that there is, but libertarianism seems primarily an American phenomenon, and may not have any objective existence elsewhere. Definitely Eysenck found no evidence for it in his own investigations into the psychology of politics.

  3. The smartest people, the ones who really get at a deep level occums razor and selection, are more likely to be on the right or libertarian side. The people one step below them, who like complexity but haven't grasped the beauty of simplicity and coherence are the most likely to follow the complex rules taught in colleges, which are universally left wing.

  4. "Openness/Intellect is the only Big Five personality trait that correlates consistently with general intelligence (DeYoung, 2011)"

    Staffan made a detailed argument on that: the reason this is so is due to the way Openness is defined in the Big Five – it actually includes intellect. The trait is measuring IQ:

    Openness to Experience – That Liberal Je Ne Sais Quoi | Staffan's Personality Blog

    This was removed – rightly – in the HEXACO. For that and other reasons, people should stop conducting studies using the Big Five and use the HEXACO instead.

    "Solon does include data from the Netherlands 2004, South Africa 1990, Brazil 2012, Germany 2012"

    Well, here's the fundamental problem: much of this data (and indeed, most psychology data in general) comes from Northwestern European populations (and their offshoots). HBD Chick's theory would posit that much of modern liberal thought is the hallmark of outbred, non-clannish peoples. By contrast, conservatism is the purview of clannish, inbred peoples, and is pretty much the prevailing state across much of the world (if the World Values Survey is any guide).

    "The lower intelligence group orients toward the left because it is populated disproportionately by leftist beneficiaries (e.g., economic and racial minorities)."

    In general, clannish people act in self-interest: if they stand to lose from redistributionist policies, they oppose them. If they stand to gain, they support them. Most of the non-Whites in America, for example, are Democrats because they are typically clannish peoples who stand to benefit from redistribution.

    That said, it would not be surprising that if you were to look in any population across the world, you would find that the more outbred individuals tend to be more liberal. Most societies have given up cousin marriage long ago – the Islamic world excepted. And even in the Muslim world, some individuals appear to have practiced less close marriage. Those individuals appear to be more liberal (if Egypt is any example).

    HBD Chick has noted that, over all, clannish populations are less intelligent than non-clannish populations (China being the largest outlier). This is partly because clannish societies support a level of dysgenic breeding with their own forms of in-clan welfare. Hence, less clannish individuals may be more intelligent overall. More research is needed.

    See also, which seems relevant:

    Predictions on the Worldwide Distribution of Personality | JayMan's Blog

  5. Stupid people believe X, etcetera. If Dr Cochran is reading this he can complete the quote from himself.

  6. Dear Jayman, Thanks for your thoughtful observations. Staffan post very good. I need to look at HEXACO again. The point about self interest of clanish and not very bright group makes sense to me. I dropped the distinction Solon drew between self interest and experience, partly for brevity partly because I was not convinced, and thought I would go back to it sometime. Solon has now taken me up on that: "With regard to the notion that dull people are leftist by expediency: I make the distinction between self interest influence and personal experience influence and it appears that the least intelligent people orient toward the left not out of selfishness but because their personal experience has led them to perceive the segment in question with an external locus ideology, empathy and trust, which leads them to a leftist vote strictly on issues relating to that segment". I will need to work that one out after looking at the paper again

  7. "This was removed – rightly – in the HEXACO. For that and other reasons, people should stop conducting studies using the Big Five and use the HEXACO instead."

    Not to my knowledge; HEXACO Openness is filled with questions relating to intelligence, asking about things like scientific interests and nonconformity. HEXACO Openness did remove contamination from Extraversion, and hence doesn't give rise to that ridiculous GFP trait theorists talk about. But it's rotated to be roughly orthogonal to Schizotypy, not IQ. Although the Big Five is more primitive and I think should ultimately be superseded by the HEXACO model, researchers have noted that BF Openness can be split between two components, one ("Intellect") which is rotated towards intelligence, and the other ("Openness") which is best measured by aesthetic and imaginative items and is independent of IQ, but correlates with positive schizotypy. Although there are many more papers you'd have to look at to get a full sense of what I mean, these two can at least start you along:

    Ashton, M. C., & Lee, K. (2012). Oddity, schizotypy/dissociation, and personality. Journal of personality, 80(1), 113-134.


    DeYoung, C. G., Grazioplene, R. G., & Peterson, J. B. (2012). From madness to genius: The Openness/Intellect trait domain as a paradoxical simplex. Journal of Research in Personality, 46(1), 63-78.

    1. Thanks. Why is GFP ridiculous? General factors don't cover everything that specific factors do, but that is the whole point.

    2. As I understand it, the whole point of exploratory factor analysis in trait psychology - the kind of analysis that established the existence of five factors, and, later, six - is to map out the vector space of personality factors in an attempt to uniquely position individuals in terms of x, y, z, and however many other coordinates are necessary. That measured Openness correlated with Extraversion in classic Big Five studies indicates that neither test had sufficient discriminant validity. The fact that O and E correlated with one another at all simply indicated contamination which should have been removed; in other words, the vector space should have been properly orthogonalized.

      Your perspective may be different from mine, as I'm not a psychologist. But discriminant validity has always been a major consideration in all the personality and attitude inventories that I've designed, and when the following study finally appeared, I can tell you without exaggeration that I received it as a great big breath of fresh air:

      de Vries, R. E. (2011). No evidence for a General Factor of Personality in the HEXACO Personality Inventory. Journal of Research in Personality, 45(2), 229-232.

    3. Dear Mark, Thanks. de Vries and quite a few others includind John Rust are in favour of 5 or so factor models, and this is probably the best fit with the data derived from questionnaires. However, general factor theorists have argued that personality must be for something, in that it must be selected for. They try to link the scores with real life outcomes, not just questionnaires. That is what I like about their approach. However, de Vries argues that personalities arise from different circumstances, so the argument goes round. I will now look up the link to a previous post I wrote.

    4. Intelligence, Personality and Self Knowledge. Try searching for that post written about a year ago. My search facility is down at the moment, so I cannot send you a link.

    5. I just reread that post. Interesting.

      The use of peer ratings is a strength. Behavioral genetic studies show that peer ratings increase the heritability of personality traits (and decrease the shared environment when there is one). These suggest that for most traits, the heritability can used as a rough gauge of the noise in the measurement.

      The thing I don't like with the general factor of personality (GFP) is this: why would extraversion correlation with the other "desirable" traits. Indeed, around the world and within societies, higher IQ people tend to be more introverted.

      To confirm the existence of the GFP, we need large, genetically informed measurements – using peer ratings* – from around the world. Then we'd get a better idea if this was a real thing or not, and then, we'd have a better idea on why.

      *for the record, I'm not sure if peer ratings increase reliability on the Honesty-humility dimension of the HEXACO. Psychopathic people have reasons to hide such.

    6. James,

      I read your post; is this it?

      Whether or not people with higher general intelligence are also higher in the GFP (I assume this is what you mean - I can't follow the link provided in your post) doesn't change the fact that they could also be described as more agreeable and conscientious. Ultimately, whether or not personality is "for" something doesn't alter the fact that we could select any rotation or any set of vectors, orthogonal or oblique, to describe a space. Given the methodology of psychological science, which leans heavily on correlations to show effects, using a non-orthogonal set leads to false positives because of poor discriminant validity.

      Consider the issue of the post we're commenting in: Numerous studies found liberals more Open than conservatives, and others found liberals more intelligent than conservatives. But it was not possible to determine which was actually the main effect, since O and IQ are positively correlated. Single studies seldom reported effects from both personality and intelligence batteries, and if they did, the authors never attempted to control one for another. Only now that psychologists have finally developed a measurement of imaginative/creative/sensitive personality which is independent of intelligence can we see clearly that the main effect is (surprisingly to me) on intelligence. Social conservatives evidently can be imaginative and aesthetically sensitive, but tend to have a significantly less intellectual outlook. This clarity was impossible without orthogonality in the vector space.

      In fact, there are other areas outside of psychology where an orthogonal set of vectors should also be preferred for describing a space. Consider for instance my discussion on color, here:

      Orthogonality isn't necessary, but it is very, very nice.

    7. @Mark Graybill:

      "Consider the issue of the post we're commenting in: Numerous studies found liberals more Open than conservatives, and others found liberals more intelligent than conservatives. But it was not possible to determine which was actually the main effect, since O and IQ are positively correlated."

      Were those on homogenous populations? And then, was that with the Big Five or the HEXACO?

      "Social conservatives evidently can be imaginative and aesthetically sensitive, but tend to have a significantly less intellectual outlook. This clarity was impossible without orthogonality in the vector space."

      I'd like to see this data, with the above qualifications.

    8. Mm. Perhaps I'll discuss it in a future post.

    9. I enjoyed your essay, and found it very helpful.

  8. My reading of the literature is somewhat different to Solon's. I am currently working on a critical comment piece, which I intend to submit as a response.

    1. Great. Let me know when it is available so I can read and comment.

  9. Steve Sailer mentions leftist attitudes as a social status marker among upper middle class whites. Something they can lord over lower social status, non-leftist whites. This idea seems plausible. Alienated intellectuals (many of whom in 19th and early 20th century society may have been insecure about their social standing) may have increasingly developed leftist attitudes as part of some sort of social class identity.

    1. Steve could be right. Solon thinks it is not the case. I did not cover that, simply for brevity.

  10. Stupid parlour game, I know. But here goes. Try and rank the following in ascending order of IQ, based on any evidence you can think of:

    Fidel Castro
    Hugo Chávez
    Winston Churchill
    Benjamin Disraeli
    Osama Bin Laden
    Richard Nixon
    Enoch Powell
    Pol Pot
    Maximilien de Robespierre
    Leon Trotsky

  11. Thompson I failed to notice that I had not sampled business leaders, lawyers, engineers and City of London entrepreneurs and investors, simply because I did not know many.

    Regarding business specifically, wanting to make a great deal more money than you need to live, as an average, affluently middle class person, is actually a fairly unusual trait. Same as being willing to live with a great deal less money than the average person.

    Most people talk about being rich in the way that is a fantasy, but very few people are really very interested in becoming very rich, vs having more leisure, time with the family, etc. People in general are quite happy without expensive wines, wagyu steak, big houses, so long as they are part of the comfortable middle.

    Smart people can live fairly well, and get intellectual stimulation, outside of business.

    So to choose to go into business, with all its grubby competition* and big money, in a big way, rather than art or science or government, you probably have to be in the upper 30% of avarice, at least (to become a billionaire, the top 1% of avarice, in additional to intelligence thresholds, etc.)- you need a certain level of greed to push you beyond simply going for the average middle class, comfortable lifestyle.

    By imposing this threshold for avarice that is far from the median (above a particularly high threshold or below a particularly low one), we impose a limit on how many smart people will fall into that category.

    So to me, presuming being in business leads a smart person to become right wing, it makes a lot of sense that most very smart people are not into business and hence, are not particularly right wing (thus are relatively left wing), just as very few are ridiculously saintly.

    if you're comfortable and not particularly concerned with being allowed to carry out activities with which you can get rich, why would you try to structure the economy to further benefit yourself (i.e. be economically right wing in a functionally pro-corporate, pro-banking, pro-shareholder way)?

    *most everyone who does business, who is not some sort of mild psycho, generally feels slightly soiled by the commercial and bourgeois. this is true throughout our history (every admirable civilized culture of note has taken this attitude, across history).

  12. In other posts I have discussed the fact that students identified as bright at 13 go on to become very much wealthier than average by mid-career. Benbow and Lubinksi have excellent data on this.