Student of Trinity
Silver Meritorious Patron
The control group were asked to give their own statements about topics such as math, physics, etc. What they thought about Hubbard's ideas is irrelevant, so it does not matter at what time they lived.
All IQ tests, including the one that my cousin designed, are normalized in such way that the average score is 100. The standard deviations are different for these tests, it is impossible to normalize them in such way that they have standard deviation of 15. (standard deviation is non-normalizable).
How large was the control group? How was it chosen? How were the control group IQs determined, independently of your procedure? These are things that someone who understood statistics, let alone psychometrics, would normally have mentioned immediately, because they make all the difference between solid evidence and no evidence at all.
The time period is important because common opinions about things like the link between tobacco and lung diseases have changed. To believe as Hubbard did in his day was not uncommon, whereas today it would be unusual ignorance.
Standard deviation is trivially easy to normalize. Simply multiply the difference between your raw score and its mean by a constant factor. If your raw scores have a standard deviation of 30, for example, then you simply need to rescale all your results (relative to the mean) by a factor of two, so that a raw score of 160 on your test becomes 130. If you don't understand why this works, just do the arithmetic, on any sample you wish. It's not complicated. So I don't know where you got the idea that SD can't be normalized. I'm afraid I'm starting to suspect that you don't understand statistics very well, which would be a serious handicap in psychometric work.
Less trivial ways of normalizing a typical IQ test would include weighting the questions differentially, or choosing a different set of questions. Since lack of control over the questions is one of your method's basic flaws, though, you can't really do those things; but the rescaling will always work. However you normalize your test for standard deviation, it is important. If your standard deviation is not 15, you are not measuring IQ.
Last edited: