No.
There is no correlation whatsover. You can be a great parent or a terrible parent it's not going to have ANY bearing on whether a child is born with a form of Autism.
Just curious... Do you guys (in the US) spell Aspergers differently? I noticed you and Docbennett both refer to it as 'Asbergers'. No biggie, just curious.
Hans... great pic!
Modified SE Bernie Marsden, Fender Strat
Laney Lionheart L5T-112, Fender Mustang 1
Wishing for a Blue Bernie!
Click here for SE Bernie Marsden demo!
Lessons, covers, backing tracks, etc...www.youtube.com/mikegarveyblues
As my statistics professor taught me many many years ago...it does, when the correlation is 1.0. A correlation of 1.0 = "de facto causation"
You can never infer causation from correlation analysis...but, as the .r approaches 1.0 and exceeds .9....you have virtual causation, although you can never actually state that Variable A results in Effect B.
Last edited by docbennett; 12-19-2012 at 01:54 PM.
I spend much of my day performing distribution analytics, dynamic correlation analysis, hedging and tail risk truncation functions. Correlations are limited by a wide degree of factors the most common include the impact of outliers, the potential for spurious correlation, the confusion of exogenous and endogenous variables, and the limitations of correlation to measure relationships beyond the linear. To avoid some of the problems with simple linear regression you can attempt to use multiple regression in which lagged terms, control variables, and nonlinear terms can all be included as independent variables to better specify the relationship...however I would be extremely careful with the notion of assuming de facto causation. Assumption in my world is the mother of all ****ups.
In addition have I mentioned how much I hate the tendancy of gaussian based functions to creep up everywhere? especially where they don't belong?
Last edited by Dirty Bob; 12-19-2012 at 01:32 PM.
That's an easy one....calculating the area accounted for within bell-shaped curve using the formula for standard deviations above and below the mean is relatively easy...you've got T-scores and related parametric stats. It's those non-parametric statistics, or worse...any attempt to extract differentials in calculus or to do factorials that leads directly to my sitting in a squatting position under said bridge with a syringe, a dirty spoon, a PRS Zippo lighter, and a copy of Bruning and Kintz's Computational Handbook of Statistics.
brother..could you pass the Sterno, please?