Arc of Essentials ©

 -  Method for Causal Analysis



Statistical results applied for practical use: Deductive Validity



[dd. 2012:01:16 - 22h:09m (40s:405ms) ]

Predictive power of sample correlation for deduction to N=1 samples



- Structure -


Sample size:
  

N

0= 1000000000;
  

N

1= 1000;
  

N

2= 100;
  

N

3= 20;

Max. false negative conclusion of difference (prb.): alpha=  0.050 ;
Max. sign estimation error (prb.):  0.050 ;
Min. confidence level (pct.):  95.000 .

I. Sample values
(observation)

(1) Min. Sample count of hits (equal in both groups) (pct.).
(2) Min. Sample corr. .
(3) Min. Sample corr. cumulat. prob. .

II. Population values
(generalization)

(4) Max. Popul. corr. (gradient) .
(5) Max. Variance accounted for, total (ppt.) .
(6) Max. Mutual info., average (bits) .

III. Population values, averages for individuals
(re-test prediction)

(7) Max. Accuracy of sign (+/-) estimate, overall (pct.) .
(8) Max. Accuracy of quintle (5th) estimate, overall (pct.) .
(9) Max. Accuracy of decile (10th) estimate, overall (pct.) .
(10) Max. Accuracy of vigintile (20th) estimate, overall (pct.) .

IV. Values for individuals
('N=1' prediction)

(11) Max. Confid. of sign (1/2) predicted, indiv. (pct.).
(12) Max. Confid. of quintile (5th) predicted, indiv. (pct.).
(13) Max. Confid. of decile (10th) predicted, indiv. (pct.).
(14) Max. Confid. of vigintile (20th) predicted, indiv. (pct.).