Field of Science

Why Intelligent People Watch More TV

Satoshi Kanazawa is writing a bunch of blog posts using his pet hypothesis ("The Hypothesis") to explain human behavior:
I believe evolutionary psychology is key to uncovering the origin of individual preferences and values. The Savanna Principle states that the human brain has difficulty comprehending and dealing with entities and situations that did not exist in the ancestral environment. The theory of the evolution of general intelligence suggests that general intelligence evolved as a domain-specific psychological adaptation to solve evolutionarily novel problems. Their logical conjunction suggests a qualification of the Savanna Principle and leads to a new hypothesis about individual preferences and values.

If general intelligence evolved to deal with evolutionarily novel problems, then the human brain’s difficulty in comprehending and dealing with evolutionarily novel entities and situations (proposed in the Savanna Principle) should interact with general intelligence, such that the Savanna Principle holds stronger among less intelligent individuals than among more intelligent individuals. More intelligent individuals should be better able to comprehend and deal with evolutionarily novel (but not evolutionarily familiar) entities and situations than less intelligent individuals.
The more intelligent people are, the more likely they are to do evolutionary novel things. Like drinking alcohol, smoking cigarettes, and be vegetarian, have sexual affairs, commit less crime, go to bed later, be atheist, and, of course, be politically liberal.

Some of these observations I don't dispute, but (while I am persuaded that human behavior is influenced by our evolutionary past), I have some trouble believing that it is correct to invoke The Hypothesis so widely. If this principle is true, then why stop at the obvious stuff? (Though I do not think smoking is so obvious, and apparently the data doesn't reflect it, either.)

How about watching TV. Since spending large amounts of time passively watching other people go about their lives is evolutionarily novel, should we not hypothesize that more intelligent people watch more TV? Somehow I have trouble seeing the data coming out in favor of that prediction.

More intelligent people drive cars? There were no cars on the savannah 100,000 years ago. So, intelligent people do it more? Can't wait to see those data.

Shop in supermarkets? Conversely, are people who hunt less intelligent than people who people who don't?

I'll have to admit, though, that coming up with crazy examples - like I trust you'll agree the these are - is far harder than construing examples that are likely to fit with The Hypothesis. Like reading books, traveling, using computers, sitting in offices all day, and having fewer children, or practice BASE jumping. It also neatly fits with the observation that political leaders are so darn dumb.

But then, if you read Kanazawa's post on smoking cigarettes, it turns out not to conform to The Hypothesis. Particularly, in the UK the trend is opposite. The more intelligent people are before the age of 16, the less likely they are to be smokers later in life. Kanazawa plans to address this incongruence in a later post, but I worry that the answer might be some explaining away how smoking really isn't that novel in evolution. Determining what is evolutionarily novel is probably not as straightforward as I have suggested above. They could be lots of exceptions that would render The Hypothesis unusable, I would think, without at least taking lots of other factors into account.

Update 10/23: Kanazawa has a new post about why the correlation between intelligence and smoking is opposite in the US and the UK. Short answer is that he doesn't know. Longer answer is one that he doesn't believe in himself.


  1. Argh, it should be easy to do a sanity-check on the data, but the GSS appears to be down.

  2. The GSS does not have a really great way of correlating with intelligence, but the general standard is to use WORDSUM, which, simply, is the number of words the person got correct on a 10-item vocabulary test. It's not perfect, but it seems to do alright.

    Correlating WORDSUM with hours of TV watched per day produces a small but highly statistically significant negative correlation. Hypothesis FAIL!

    WORDSUM vs. whether or not the person or their spouse hunts produces a small but highly statistically significant positive correlation. Hypothesis FAIL!

    Driving is trickier, I haven't found a specific GSS variable that correlates with amount of time spent driving. "Have you purchased a car in the last five years?" and "What was the most important reason?" have no correlation to WORDSUM. There is a "Travel time to work" field but I cannot get it to cross-reference that with WORDSUM for some reason...?

    In any case, I was *not* controlling for other demographic factors, and in all but the TV-watching case, I would have expected demographics to push it in the opposite direction of The Hypothesis -- e.g. with whether a car was purchased in the last five years, I'd expect more intelligent people to have more income and therefore be more likely to have done so, but there was no statistically significant correlation (not even close, p was like .3 or something). So take this for what you will.

  3. Heh, so I misread it... it was actually, "Have you purchased a used car in the last 5 years?" When I control for family income, an almost statistically significant (p=.07) positive correlation appears. Not that this has much reflection on The Hypothesis... but it suggests to me that (maybe -- remember p was not quite under .05) intelligent people are more likely to buy a used car even if they can afford a new one...?

  4. So, higher scorers on WORDSUM hunt more?

    I'd really like to hear Kanazawa respond, but he doesn't allow comments to his posts, so it's hard to get to him. Email?

  5. If I were using my blog to describe Islam as a dominant allele, I wouldn't allow comments either.

  6. Kanazawa's "Hypothesis" is a fairly vapid catch-all in my opinion. To suggest that something we do today (drink, have sex, be atheist, etc.) is influenced by our evolutionary past isn't terribly specific or insightful, and I haven't found a very precise definition (put forth by Kanazawa) of the historical timespan that constitutes our "ancestral environment."

    The whole EEA concept is useful in a general sense (since we manifest an evolved phenotype that wasn't created de novo each generation), but breaks down when probed for specifics.

    Kanazawa's insights are pretty shallow and media-friendly but not much else.

  7. By evolutionarily novel, really what we're talking about is behaviours that are adaptive in one environment the one in which evolution took place) but not in another. So if you want to test the hypothesis, really you need to look at behaviours that could, plausibly have been done in both environments (not smoking or driving cars), and which provide different reproductive fitness according to the environment.

    I'm not sure if there are any behaviours that really fit the bill. Perhaps its a hypothesis that can't be tested.

    But maybe "Eat as much fat and salt as you can" is one. "Don't ever trust strangers" might be another.

  8. That is clearly not what Kanazawa means by evolutionarily novel, though. I'm with you that this definition makes good sense, but not sure that Kanazawa's doesn't, too, in terms of The Hypothesis...

  9. I agree with Bjorn, I think Kanazawa is trying to go past that. I would paraphrase what Tom said as "Intelligence is being able to formulate behaviors depending upon which is selectively advantageous in the present environment," which seems rather trivial to me. Kanazawa seems to want to say that "Intelligence is any behavior that deviates from evolved instinct."

    Personally I'm beginning to think that Kanazawa's Hypothesis is just his way to say, "Hey look, I smoke and drink because I'm smart, not because I'm weak!" heh... Though, there does seem to be a positive correlation between intelligence and drinking (WORDSUM vs. frequency of alcohol consumption is positive with p = ~0.04 in the GSS), but IMO that is more to do with the fact that intelligent people tend to be less happy... and my pet hypothesis (is it cooler if I capitalize it? my Hypothesis?) is that this correlation is down to a third variable, curiosity. People who are "born seekers" tend to be more intelligent, which makes sense... and if you are a seeker, then by definition it is harder to be content. QED.

    I can haz evoloooshunaree psychologie degree now?

  10. I should note that Kanazawa distinguishes between intelligent and smart, and for The Hypothesis, he is talking about the former, while smarts is something low-IQ also have. See this post.

    I can haz evoloooshunaree psychologie degree now?

    No degree is needed to be an expert in EP, don't you know?

  11. Just a quick note to let you know that your post is included in the latest Scientia Pro Publica. Thank you for sharing it.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS