One of the few protected occupations in Soviet Russia was the airbrush department, which was tasked with removing non-persons from official photographs. Long before PhotoShop they diligently removed those who had fallen from grace. As the number of banished or murdered party members grew the skills required in this department rose to the highest order of visual deceit. In the end they just showed pictures of Stalin, and all was well with them, so long as the pictures were suitably flattering. A Russian Nomenklatura friend of my acquaintance told me that the official architects presented Stalin with their plans for a Moscow skyscraper. He took one look at the design, and without a word drew a big steeple on the top with his red pencil. The architects bowed, and left the room. Back at base they were in a state of terror: they knew that they had to add a large steeple to their building, but they could not work out whether it had to be red.
Something like that seems to be happening in social science research. Just when you would expect to see Trotsky staring out at you, instead you find a piece of old raincoat apparently belonging to a bystander.
For example, let us say that a very large sample of genomes and intelligence results are collected from people who live in very different parts of the world, such that they have different ancestries. The overall sample is analysed, and some overall effect is detected, and reported. The question then arises: what were the results of the comparison of one continental group with another on intelligence? Are there differences, and to what extent can these be explained from the genetics?
Another example, is when you compare two smaller racial groups on intelligence and genetic profiles, but standardise the results for each sample on each sample (show Z scores derived from the group’s own mean results) such that there is no apparent difference between them. Having done that you cannot easily identify genes which might account for the differences between the two groups, which would be a major finding, which might illuminate the development of intelligence.
So, I have decided to write to all researchers who leave out obvious bits of analysis, to ask them whether they intend to publish their findings at a later date or whether they can tell me right now about what they found when they carried out the obvious comparisons. If you find examples of papers in which group comparisons are air-brushed out of the record, can you please send them to me? It would be a great pity if scientists distorted their findings in response to political pressures. A pity, and utterly wrong, if science is still to be a search for truth.