Thanz
Fuzzy Thinker
- Joined
- Jul 24, 2002
- Messages
- 3,895
You still do not understand the comparison, do you? I do not make the comparison to show that the data IS flawed, I make the comparison to show the effect of that flaw on the calculation. Don't you see the difference? I am not pointing to the two results and saying that his is wrong because it is different. I am pointing to the two results to show that the error does make a difference, and it does have an effect. And that comparison is perfectly valid.CFLarsen said:You cannot use the comparison of two calculations to show that the data used in one of the calculations is flawed. You have to show that the data itself is flawed.
Further, I have shown that the data itself is flawed, several times. I used simple logic to show that counting "J, like John or Jim" as three independent J guesses makes no logical sense. Neither Mr. Hoyt nor yourself have provided a logical argument to the contrary.
Once I showed that Hoyt's method overcounted everything, Hoyt asked me to show why it mattered. His claim was that if he overcounted everything, both the numerator and the denominator would rise and this would wash out any impact of the overcounting. The comparison was required to show that this is not the case. The overcounting caused the results to go from non-significant to significant. There is no way to show this without comparing the two data sets.
You really don't know what you are talking about, do you? Once again, I do not use the comparison to show that Hoyts data IS flawed - I only use it to show the EFFECT of the flaws on the analysis. I use simple logic to show that the data itself is flawed, and that logic has yet to be refuted. Ergo, you still don't have a point.Ergo, your comparison is worthless. Ergo, your point about Hoyt's data is invalid.