Article by Jim Lindgren in "reason":
The fracas over Don Lemon of CNN laughing at his panel's insults
about the intelligence of Trump supporters raises a larger issue: the
ignorant belief that Trump supporters are much dumber than the general
public and much dumber than those who supported Clinton in 2016. Don
Lemon and his guests specifically ridiculed Trump supporters for
supposed problems with "readin'" and "geography" (e.g., picking out
Ukraine on a map).
Even without looking at the
data, it would be surprising if there were any VERY LARGE differences in
intelligence between the average Trump supporter and the rest of the
general public.
INTELLIGENCE OF TRUMP SUPPORTERS
We
don't have great data on the intelligence of Trump supporters, but the
best available is in the 2018 General Social Survey. For those
unfamiliar with the GSS, it is usually regarded as the leading omnibus
academic survey in the US; it usually achieves response rates about 10
to 20 times higher than the typical public opinion poll.
In
1974, the GSS adopted a 10-question vocabulary test (WORDSUM) that was
extracted from a standard, widely used IQ test. The National Science
Foundation (NSF), in its 2018 report on science knowledge, refers to this battery of GSS items as a "verbal ability" test.
In
the 2018 GSS, respondents were asked for whom they voted in 2016
(PRES16) or for whom they would have voted if they had voted (IF16WHO):
Clinton, Trump, someone else, or no one.
On the
verbal ability test (WORDSUM), not surprisingly the median number of
vocabulary questions correct was the same for both Clinton and Trump
supporters: 6 out of 10 words correct. The mean verbal ability
score for Trump supporters was 6.15 words correct, while the mean verbal
ability score for Clinton supporters was 5.69 correct, a difference of
nearly a half a question on a 10-question test. This moderate difference is statistically significant at p<.0005.
Further,
Trump supporters score significantly higher on verbal ability (6.15
correct) than the rest of the public combined (5.70 correct), whereas
Clinton supporters score significantly lower on verbal ability (5.69
correct) than the rest of the public combined (5.98 correct).
This should not be too surprising. On
the 22 General Social Surveys using the verbal ability scale since
1974, for every single one, conservative Republicans score significantly
higher than the rest of the public combined. As for Republicans
overall, they score significantly higher in verbal ability than
Democrats in all five decades, including for the 2010s combined.
But
the Trump era is helping Democrats to catch up: the Republican
advantage dropped to insignificance in 2016, and in 2018 Democrats (6.03
correct) actually scored slightly (but insignificantly) higher than
Republicans (5.98 correct).
In 199[4], the GSS
employed another module lifted from a standard IQ test, one testing
analogical reasoning. Again, Republicans and conservative Republicans
in 199[4] performed significantly better on analogical reasoning than
the rest of the public and significantly better than Democrats.
TRUMP SUPPORTERS' KNOWLEDGE OF SCIENCE
These
results on verbal ability are also consistent with the results of most
(but not all) of the National Science Foundation's science knowledge
questions on the GSS.
Testing the hypothesis
that Trump supporters have greater science knowledge than those who
supported Clinton in 2016, on six questions Trump supporters offer the
correct answer significantly more often than Clinton supporters: those
about lasers, radioactivity, viruses, the father's contribution to the
biological sex of the child (BOYORGRL), whether "according to
astronomers" the universe began with a huge explosion (BIGBANG1), and
that the earth goes around the sun and that it takes a year to do so
(combined EARTHSUN and SOLARREV).
On one science
knowledge question—whether the center of the earth is hot (HOTCORE)—the
superior performance of Trump supporters over Clinton supporters is
borderline significant (1-sided Fisher's Exact Test p=.05-.10).
On
two questions, the structure of atoms (ELECTRON) and continental drift
(CONDRIFT), Trump supporters score slightly, but insignificantly, better
than Clinton supporters. On none of these nine science questions do
Trump supporters score worse than Clinton supporters.
When
one compares Clinton supporters to the rest of the public combined,
Clinton supporters perform significantly worse than the rest of the
public on the same six science questions on which Trump supporters
perform better than Clinton supporters.
Indeed, less
than half of 2016 Clinton supporters (49.6%) are able to answer
correctly both of two related questions: whether the earth goes around
the sun or the sun goes around the earth (EARTHSUN) and whether that
takes a day, a month, or a year (SOLARREV). Remember these two
questions are multiple choice! You would have a 50-50 chance of guessing
correctly on the first part: whether the earth goes around the sun or
vice versa. Sadly, the general public didn't do hugely
better than Clinton supporters, with only 57.1% (compared to 49.6%)
knowing that the earth goes around the sun and that it takes a year to
do so.
When one compares Trump
supporters to all the rest of the public combined (rather than just to
Clinton supporters), the pattern for these nine science questions is
roughly similar (though weaker).
Overall, on
most science knowledge questions Trump supporters score significantly
higher than Clinton supporters and significantly higher than the
combined non-Trump supporting public. If, however, you asked about
beliefs, rather than knowledge, on evolution and the origins of the
universe you would get substantially better answers on individual
science questions from Clinton supporters than Trump supporters.
MAP-READING
As
for reading maps and picking out countries, which the CNN segment
raised, I searched quickly and found two Pew surveys from 2013 that
asked respondents to pick out Egypt or Syria on a map of the Middle
East. Testing the hypothesis that Republicans were significantly
better at finding an unlabeled country on a map than Democrats, one
2013 Pew study supported that hypothesis (Republicans were indeed
significantly more likely to pick out Syria on a map), while the other
2013 Pew study reported that Democrats were insignificantly better at
picking out Egypt on a map.
Thus,
neither of these two studies supports the CNN's panel's ridicule of
right-wing map reading, and there is some weak evidence pointing in the
other direction. Of course, this was a test of Republicans, not
Trump supporters, but Trump supporters did better on the 2018 GSS
verbal ability test and on 2018 science knowledge questions, so there is
no strong reason to suppose that the results would be radically
different if one were to test Trump supporters today rather than
Republicans in 2013. In 2013 the differences were not large either way,
and it's unwarranted to suppose that (in a study of the quality of the
GSS) any differences in map-reading would be large today.
IGNORANCE LEADS TO BIGOTRY
Don Lemon laughed uncontrollably at his guests insulting the intelligence and knowledge of Trump supporters. The
best evidence we have suggests that, compared to the general public,
Trump supporters score significantly better than the rest of the
public—and Clinton supporters score significantly worse—on a standard
verbal ability test. Likewise, Trump supporters score significantly
better on most science knowledge questions than Clinton supporters or
the general public.
In
this essay, I analyzed the results of over 30 questions from 22
different representative national surveys, involving over 20,000
respondents. Not one of the questions I examined here supports the idea
that Trump supporters are significantly less knowledgeable than Clinton
supporters, and some of them point to small or moderate differences in
the opposite direction. The idea that there are very large differences in intelligence or knowledge here is implausible without strong evidence.
In short, Don Lemon is a bigot—and like most bigots, he's an ignorant one as well.
[Research Note: General Social Survey data were downloaded from
NORC. GSS data are weighted by WTSALL. On science questions, I coded the
correct answers v. those who gave wrong answers, said they don't know,
or failed to answer. The Pew data were downloaded from the IPOLL
database at the Roper Center, and the WEIGHT variable was used. For 2x2
tables, significance was determined by 1-tailed Fisher Exact tests. For
differences of means, 1-sided independent T-Tests were used without
assuming equal variances.]