- This topic has 3 replies, 2 voices, and was last updated 4 years ago by
.
- Posts
Bpp kit (4a.3)
A regression equation Y = a + bX is used to forecast the value of Y for a given value of X.
Which TWO of the following increase the reliability of the forecast?A) A correlation coefficient numerically close to 1
B) Working to a higher number of decimal places of accuracy
C) Forecasting for values of X outside the range of those used in the sample
D) A large sample is used to calculate the regression equationAnswer is A and D.
My query:
(1) B is more correct than A. A can be wrong if the value is 1.1 or -1.2 or any greater and closer value to 1.
(2) Why B is not possible?General query
(3) Can you please tell strong correlation range? I studied +-0.8 but in BPP kit it is saying that 0.73 is even strong in Question (4a.5).1. It is arithmetically impossible to end up with a coefficient that is great than 1.
2. Working to more decimal places doesn’t make things any more linear.
3. There is no specific value that makes a relationship strong (I don’t know where you were taught that it was 0.8 !!). It is not even possible to say that 0.73 is necessarily strong, but in this sort of question it is better to eliminate the 2 that are certainly not true so as to then be left with the 2 required by the question.
I have studied that strong correlation is >0.8 and <1 while perfect correlation =1 and <0.8 is partial correlation in Kaplan Study Text.
Thank you tutor.
- The topic ‘Forecasting’ is closed to new replies.