Hi statistical net members :
I am not in the field of statistics but I joined this group to get some of
my basic doubts clarified.I work as a manufacturing engineer and I use
statistical methods for some of my improvement projects.One of these
projects is the tolerance analysis using statistics.I do tolerance analyses
of my design to provide/predict tolerances required / to be provided by
vendors etc.
Question 1)
If I have a variable X which is a function of three other variables A,B,C
and is related by this relation :
X=A+B+C
Can I apply sum of square rule on the variances and say ,
Variance (X) = Variance (A) + Variance (B) + Variance (C)
That is
(STD. DEV X)^2 = (STD. DEV A)^2 + (STD. DEV B)^2 + (STD. DEV C)^2
Am I correct ?
Question 2)
If I have a variable X which is a function of three other variables A,B,C
and is related by this relation :
X=A+B+(C/2)
Can I apply sum of square rule on the variances and say ,
(STD. DEV X)^2 = (STD. DEV A)^2 + (STD. DEV B)^2 + (STD. DEV (C/2) )^2
OR
(STD. DEV X)^2 = (STD. DEV A)^2 + (STD. DEV B)^2 + ( (STD. DEV (C) )^2
divided by 4 )
Which is correct ?
Question 3)
If I have a variable X which is a function of three other variables A,B,C
and is related by this relation :
X=A-B+C
Can I apply sum of square rule on the variances and say ,
(STD. DEV X)^2 = (STD. DEV A)^2 + (STD. DEV B)^2 + (STD. DEV C)^2
OR
(STD. DEV X)^2 = (STD. DEV A)^2 - (STD. DEV B)^2 + (STD. DEV C)^2
Which is correct ?
Before you answer,please note that I am not a statistical expert and
therefore you may need to explain in easier terms !!! .I am not really
looking for equations and stuff like that,but I am interested in knowing the
correct usage of the SUMOF SQUARES rule.
Thanks in advance,
Vishy.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|