On the Mega Test, Ron Hoeflin found that the number series questions had the best correlation to high-IQ. I wrote to say I was surprised at this, and he responded that he was at first himself. (This was over 20 years ago.) But number series aren't tainted by accidental factors as easily as vocabulary questions are. On the Mega, for example, physicians are likely to have an advantage because some of the puzzles rely on knowing the opposite of an obscure medical prefix (such as hyper- and hypo-). Easy to do online now, but difficult then unless you had those prefixes already learned or knew where in your medical references to find them.
Well, we don't usually have much opportunity to observe people discovering the rules behind number series, do we? So that's out in sizing up what their candlepower is. We can sometimes get a sense how at ease they are with numbers and scientific topics, and we can attend to the vocabulary they use. These can be deceiving if we don't attend to circumstance, however. People who have boned up on a topic just prior to your hearing may seem to have a facility somewhat beyond their actual knowledge - and reading from a prepared text can make them shine even brighter. It does take some intelligence to recognise the words, know their pronunciation, and deliver them naturally, but nowhere near what one would need to use the same words and concepts when caught off guard. Yet in the opposite direction, people who do have command of such may take pains to gear their speech to their estimate of the vocabulary of their hearers. This can be subtle, as people have usually been exposed to a richer vocabulary than they use themselves, and can follow the explanations of a person smarter or more knowledgeable than themselves.
There are also social cues that people mistake for intelligence markers. We have spoken of this before, but it bears remembering here. An offhand comment that one reads the New Yorker is an attempt to suggest one has a higher-than-average intelligence. That is likely true, but the magazine is understandable by people with an IQ of 95, so it could be an affectation. Moreover, the people who read it likely overestimate the IQ of what they believe is the average reader by 10 points.
Some folks delight in using obscure words or abstruse concepts as a way of showing off or bullying. More innocently, some people just like knowing words. Both take some intelligence, but can mislead the hearer. Buckley did both: he loved words, but would also use unfamiliar terms in debate to put his disputant on the defensive.
Knowing test scores removes the need for guessing at the intelligence - I use the term in the strict g-factor sense, without regard to more elusive concepts of common sense, insight, or clarity of expression - but many of these are rough, and we don't always have access to them. SAT scores (old scoring up to 1600) divided by 10 give a rough idea of IQ. Childhood scores are especially volatile - emotional and accidental factors can bring down the score of a bright child more easily than an adult. Not much raises your score, so if you have a childhood IQ score you may regard that as your probable minimum.
Knowing lots of people's scores gives one an intuitive feel for others whose scores one doesn't know. Where do they seem to fit - nearer Amy, Sarah, and Michael, or more like Jason, Jack, and Charlene? Not that there is usually any need to know, but it can be helpful in assessments such as oh, whether a presidential candidate is intelligent.
Such estimates circulate from time-to-time, usually for the purpose of making Republican presidents look stupid and Democrats smart. There are also wild estimates of Jefferson and other early presidents having astronomical IQs. Well, perhaps. But unlikely. People do not relate to leaders too far beyond them, and leadership skills are not well-correlated with IQ. There are interesting theories why each would be so, but frankly, we don't know.
The presidential nominees of my lifetime have all probably had IQ's between 120-130. As that is about the 90th to 97th percentiles of intelligence, that's plenty. Nixon is rated higher, for reasons I have not been able to discover, though I have a decades-old recollection that it was based on his performance at law school. Bush and Kerry were reliably estimated at 125 and 120, respectively, by Steve Sailer, who also puts McCain and Gore around 133, Nixon at 143, and Kennedy, Johnson, Carter, and Reagan around 120. He calls 125 a starting point for discussing Obama's IQ, as he fit in but did not stand out at his prestigious Hawaiian prep school, where he was unlikely to receive any affirmative action boost. I would partly agree, but would downgrade that. I think his verbal side is likely around that level, but his making the millions/billions mistake too many times, and his idiotic Curvature of Constitutional Space paper with Laurence Tribe suggests he does not understand the scientific concepts he pretends to. Also, the teleprompter jokes alerted me to pay closer attention to vocabulary when speaking informally. Not bad, but not exceptional.
Not that I'd knock him down too many points even with that, however. It does take considerable awareness of the concepts flying around you to even get them wrong, after all. So my starting point is 120, about 90th percentile, similar to many other presidents.