where x and y are the values of the two variables considered and # is 
the number of observations. 
Its significance has been tested accordingly: 
(2) 
€ esr iet 
V1—r 
Where / is distributed in the S/zudent’s form with n — 2 degrees of freedom. 
Having computed the correlation coefficients r, we have then carried 
7 = 
out the determination of the multiple correlation coefficient À between 
vn, and 7», V3; v2 and £n, vs; vs and n, ve. 
The values of A have been directly computed from the values of the 
linear correlation coefficient just found out, through the formulas: 
2 2 7 
I +11 — 2 f fa 15 
Ri, 2 = 1 2 
7123 
(3) Re qug m 
T1? d- £2; — Z f fi3 15 
R, 13.5 ] 1 2 
eei 
th =p 1% — 211431; 
Rı 12 —— T 2 
tr 
  
The significance of the multiple correlation coefficients been have 
tested through: 
1 R? 
log, - 
2 1—R” 
n— P 
4) x 
p—1 
where p is the number of variable »; and # is the number of observa- 
tions [4]. 
The numerical results obtained during the new research set have been 
summarized in Table III. 
TABLE III 
  
  
  
  
  
| | 
| #Student distri- Multiple z-Distributon 
Variabl. | Correlation bution with 23 correlation v=p—1=2 
| coefficient degrees of freed. coefficient va—n — p—22 
e te 
| 
V1 rı2 0,306 F2 — 1,543 Rı,23 = 0,315 E123 — 1,248 
va | rıs=0,013 Hs = 0,064 Ras — 0,413 22,13 = 1,558 
vo | tas 0,281 ha 1,404 Rois = 0,291 $213 — 1,160