Demented LRH
Patron Meritorious
You did not specify how 700 correct answers and 300 incorrect answers were chosen.What I don't see is the value in random selection of 200 statements. Let us say you took 700 negative statements, and 300 positive statements, Then from them you selected your random 200. If you did that random sampling several times and averaged the results, you would find you had the same ratio as the original test sampling.
So, what is the point of the randomization step? It is a fudge factor that does not lead to accuracy of results.
What should then be the ideal ratio of positive to negative statements in your sample that would not peg the results at a particular IQ? How do you rule out bias by the person selecting the statements that are used to make up the 1000?
You are basing your IQ on a sample of 100 correct answers giving a 100 IQ and we can presume, if you found more that 100 correct in your sample of 200, there exists the possibility of an IQ rating of over 100. Is this rating system the same as is generally used in the testing community?
Now, you are arbitrarily taking statements and deciding which are true and which are not. In your example you stated engrams do not exist. An engram is a moment, or incident containing pain and unconsciousness. Saying that is wrong is an error on the evaluator's part. How then do you adjust for such errors? They would skew your results.
How do you deal with the "Unreliable Narrator" aspect of his writings? If your IQ test is solely based on right and wrong statements, how do you obtain accurate results from his writings, if he willingly ignores the "truth", since he doesn't consider it necessary to running a con?
How does a ratio of correct statements to incorrect statements equate to IQ? I feel you are omitting many factors in determining "actual" IQ.
Mimsey.
If you initial goal was to select 700 correct answers and 300 incorrect ones, then the method will not work. But if you begun selecting answers one by one and arrived at the result that you described, then the method will work, every statistician will tell you that.
Suppose, you want to determine average London income. You may use a table of random numbers to select 200 entries from a census data. But that will not work because some entries contain incomplete data -- the income fields are missing in them. You may replace incompelete entries with complete ones, but that woun't work, either, because this is not random selection by definition.
The right method would be to select complete 1000 entries in any way you like -- starting from the top of the list, starting from the botto, starting from the middle, etc -- and then use the table of random numbers or generator of random numbers to select 200 records from them