Thanks, these help a lot for understanding the TMorg pr-science. -Doug
http://groups.yahoo.com/group/FairfieldLife/message/123627 --- In [email protected], new.morning <[EMAIL PROTECTED]> wrote: > > --- In [email protected], Peter <drpetersutphen@> wrote: > > > > I'll give a more civil response. The methodology > > is all screwed-up. First there is no control. What > > functions as the control? Why do they even mention the > > word "controlled"? In fact this type of research can > > have no control. You would use a design that is called > > "interrupted time series analysis". Basically it's > > measure trends over unit of time; intervention in > > effect and measure trends over unit of time; then no > > intervention and measure trends over unit of time. You > > do this over and over again. Turning the intervention > > on and off, so to speak, and measuring your trends > > with the intervention present and then when it is > > absent. Then you crunch your numbers to see if there > > are statistically significant differences within and > > between these on and off cycles. Simply measuring > > trends and then applying the intervention and seeing > > the trends improve means absolutely nothing in social > > science research. > > Yes, the TMO's "imapct analysis" -- the essence of most ME studies, > is quite weak. As many, and you, have been pointing out, the dose, > intervention, independent variable(s) needs to vary in intensity and > time frame, and "location" to really assess strong correlation -- and > eventually causality. Impact studies can't even determine simple, > insignificant, correlation. One data point points to NO trend. > > > > Its like hiring 100 extra police > > officers for two months and then seeing the crime rate > > drop. It does not mean that the extra officers caused > > the crime rate to drop. The crime rate could have been > > dropping or increasing or staying the same regardless > > of the police officers. That is, what is causing the > > change in crime rate could have nothing to do with the > > police officers. The TM research on the ME, in this > > case, is the same as seeing a car crah after you've > > sneezed and then claiming your sneeze caused the car > > to crash. The only way to know if your sneeze causes > > car crashes is to count car crashes without sneezes, > > then count them while sneezing, then without and so > > forth. > > Yes. However, extending this, if the ME event corresponds to a > significant "rare event" -- as would be the case if the S&P500 > increases another 10-15% in the next 2-3 months, and other "predicted > events" continue to occur (like political transformation, low > hurricanes, etc.) -- that still does not establish causality. Or even > simple correlation. But its hell of an interesting set of > "coincidences". Feedstock and justificatin for real long-term > multivariate causality studies. > > And more than a timeseries analysis, a long term (aka often interupted > or varying independent variable(s)) multivariate analysis is needed -- > to fully control for socio-economic and other factors. > > > > Just so people know, I'd love for there to be a > > ME. It would be so absolutely cool. But outside of the > > one DC study which demonstrated a very, very slight ME > > (and even that is generous) these "studies" are the > > equivalent of a 4th grade science experiment. They are > > junk science and it is truly a shame. I think the > > reason they don't do a time series analysis is that > > they are quite concerned that it would show nothing. > > There are 30-40 studies on ME out there. If the TMO is sure there is > an ME, they should make all data sets available for independent > analysis. 1000 grad students would jump at that. Getting good data is > 70% of a research project. > > > If anyone on the course is reading this, please ask > > John Hagelin why an interupted time series analysis is > > not used to measure the ME. You can tell him that Dr. > > Sun and Fun wants to know. He'll know who I am with > > that comment and probably laugh. > > There is additional, different, "coherence events" through out TMO > history. Superimposition with financial markets can be seen. Some > interesting things. . See second graph down at > http://2006-course-effects.blogspot.com/ > > Again, if the TMO was serious about real ME research, they could > compile the precise data for such coherence events (SIMS wave > initiations, TTCs, citizent sidha courses, large Dome courses, etc.) > and make it available to any reasearcher via internet. I guarantee > 1000 studies would bloom via grad students. And profs looking to publish. >
