Gottfried, I make comments below: "Gottfried Helms" <[EMAIL PROTECTED]> wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > [EMAIL PROTECTED] schrieb: > > > > Ggottfried, > > > > Are these D, rde, CC or ? > > > > Given the obvious meaning of simulations, why do you think these values are > > possible? > > What value should any answer to this have? > > > Gottfried Helms
Gottfried, the answer should be the truth. 1. Do you see or am I wrong that a negative D can be obtained by subtracting a large positive rde from a small positive rde? Do you see a problem with this. Did we not discuss the four possibilities, years ago: ++, -+, +- and --? 2. Do you agree that a negative rde implies the obvious polarization that comes from combining x1 and x2? 3. Can you see why defining the partions at more extreme levels for the extreme group and at more moderate levels for the midrange group, would increase the polarization effect? Do you see that normally distributed data is like using normally distributed factor cell sizes in an anova? 4. Can you think of any causal variables that might be confounded with mileage? 5. If you could determine the causal relationship between two variables, would you rather do this by analyzing the two variables or dozens of variables in a system? 6. Are the variables in a SEM model any less likely to be confounded than those in a CR model? Is not the answer to confounding the use of better measures? 7. Would not better measures require to pursuit of simple structure and thus more heterogeneous sampling strategies? 8. Do you believe me when I say my article was accepted by SEM for three years and then rejected? If so does this action seem unethical to you? What if it happened to you? Did it give TETRAD a commercial advantage that was translated into an academic advantage? One of your SEMNET colleagues recently told me there is not requirement in science to recognize other peoples work unless you actually use that work. So people can claim they discover something, without referencing those who came before them. Is this what you call science? 9. What does the Tetrad program say about the following model: y1=x1+x2, y2=x2+x3, y3=x3+x4. If we only measure y1, y2 and y3. then we have an attenuated unshielded collider. y1 and y3 are uncorrelated. y2 is correlated at .5 with both y1 and y3. In the collider, y2 would be correlated .7 with y1 and y3. Tetrad does not conclude that y2 is the effect y2 and y3, unless the correlation is .7. CR says y2 is the effect y1 and y3 (y2 is a composite of y1 and y3), when y2 correlates at ,5 and at other levels with y1 and y3. What do you make of this? 10. Do scientists have any responsibility to seek the truth? Bill > > ------------------------------ > > William Chambers wrote: > (presented (in a previous post) some arguments already): > > > a conceptional argument: > > > (...) Gottfried, like his colleagues on SEMNET (*) > > refuse to consider the uniform distributed cell sizes used in attribute > > based ANOVA designs. > > > > > a practical/ethical argument: > > > This is because they make their money promising to > > make sense of convenience samples, that are sparesly sampled in the > > extremes. (...) > > > > > > Hey, and I learned, better not to take more than two variables at a time... > so to avoid matrices at all: > > > Take the variables two at a time. You would then have something to say, > > which is a lot better than having nothing to say. Furthermore, it can be > > > > > I also don't know the actual enhancements: > > > Gottfried has avoided me and CR for four years and does not know what > > changes have occured. He and his friends on semnet have refused to read my > > > > And, well, I don't think being able to do a fruitful discussion with you at all: > > > Gottfried, I do believe that you really are stupid. CR allows us to say when > > > > > So - > > for what purpose do you try do get any words of this Gottfried? > > > ---- > (*) one of the many characteristics of SEMNET is given as: > > > (...) When I first signed onto SEMNET I was > > attacked for suggesting that even this was possible. It took a great deal of > > demonstration and tolerance of personal abuse to get even a few people to > > So, ordering me to "my collegues" of SEMNET indicates already, that you're ex- > pecting to need a great deal of tolerance of personal . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
