doesnt fix washing machines, he doesnt know what a motor is, or something. In other words, he has no technical experience. There are lots of engineers in the world. There are lots of mechanically minded people. There are lots of people who are smarter than the reporter, say, in science, for example. It is, therefore, his duty to report the thing, whether he understands it or not, accurately and in the way its been given. The same goes in economics and other situations. The reporters appreciate the fact that they dont understand the complicated business about international trade, but they report, more or less, what somebody says, pretty closely. But when it comes to science, for some reason or another, they will pat me on the head and explain to dopey me that the dopey people arent going to understand it because he, dope, cant understand it. But I know that some people can understand it. Not everybody who reads the newspaper has to understand every article in the newspaper. Some people arent interested in science. Some are. At least they could find out what its all about instead of discovering that an atomic bullet was used that came out of a machine that weighed seven tons. I cant read the articles in the paper. I dont know what they mean. I dont know what kind of a machine it was just because it weighed seven tons. And there are now sixty-two kinds of particles, and I would like to know what atomic bullet he is referring to.
This whole business of statistical sampling and the determining of the properties of people by this manner is a very serious business altogether. Its coming into its own, but its used very often, and we have to be very, very careful with it. Its used for choice of personnelby giving examinations to peoplemarriage counseling, and things of this kind. Its used to determine whether people get into college, in a way that I am not in favor of, but I will leave my arguments on this. I will address them to the people who decide who gets into Caltech. And after I have had my arguments, I will come back and tell you something about it. But this has one serious feature, among others, aside from the difficulties of sampling. There is a tendency, then, to use only what can be measured as a criterion. That is, the spirit of the man, the way he feels toward things, may be difficult to measure. There is some tendency to have interviews and to try to correct this. So much the better. But its easier to have more examinations and not have to waste the time with the interviews, and the result is that only those things which can be measured, actually which they think they can measure, are what count, and a lot of good things are left out, a lot of good guys are missed. So its a dangerous business and has to be very carefully checked. The things like marriage questions, How are you getting along with your husband, and so on, that appear in magazines are all nonsense. They go something like this: This has been tested on a thousand couples. And then you can tell how they answered and how you answered and tell if you are happily married. What you do is the following. You make up a bunch of questions, like Do you give him breakfast in bed? and so on and so on. And then you give this questionnaire to a thousand people. And you have an independent way of telling whether they are happily married, like asking them, or something. But never mind. It doesnt make any difference what it is, even if the test is perfect. Thats not the part where the trouble is. Then you do the following. You see about all the ones who are happyhow did they answer about the breakfast in bed, how did they answer about this, how did they answer about that? You see its exactly the same as my rat race, with right and left. They have decided on the odds of the thing in terms of the one sample. What they ought to do to be honest is to take the same test that has now been designed, in which they know how to make the score. Theyve decided this gets five points, that gets ten points, in such a way that the thousand that they tried it on get marvelous scores if they are happy and lousy scores if theyre not. But now is the test of the test. They cannot use the sample which determined the scoring for them. Thats going backwards. They must take the test to another thousand people, independently, and run it out to see whether the happy ones are the ones that score high, or not. They do not do that, because its too much trouble, A, and the few times that they tried it, B, it showed that the test was no good.
Now, looking at the troubles that we have with all the unscientific and peculiar things in the world, there are a number of them which cannot be associated with difficulties in how to think, I think, but are just due to some lack of information. In particular, there are believers in astrology, of which, no
a thing would work.
But I cant disprove, of course, without investigating more carefully. Maybe one way would be always to ask them how do they know its true and to remember maybe that they are wrong. Just remember that much anyway, because you may keep yourself from sending in too much money