Hi k9queen,
I don't know what statistical software you're using, so I can't
provide you with appropriate instructions for running the regression.
I must also admit that I'm not familiar with that many packages and I
wouldn't want to try to figure out a package I don't know well. I
used TSP to calculate this regression, and I can provide the code for
you if you're interested.
After regressing quarter on sales the computed R2 is 0.020979, which
means that only 2.09% of the variation in sales is predicted by its
quarter. Essentially the quarter has no predictive power.
This makes sense because sales do not follow any linear growth
pattern, but instead just bounces between two values. Quarter grows
linearly from 1 to 12. So, since there is no growth (or decline) in
the sales over the period 1 to 12 the quarter could not have any
predictive power.
You might want to try running the regression with some different
numbers to convince yourself of this. If you made the numbers 450,
460, 470 ... 560, you would find the R2 to be 1.00 because the growth
in both variables would be totally linear. If you then changed some
of the sales numbers a bit, but left most of them the same (460,
470...) you would see the R2 fall as the quarter became less of a
predictor. In the case where changes are essentially random, or have
no growth pattern (positive or negative), the R2 will be close to
zero.
I hope this helps you out. Good luck with your studies. If you need
any clarification, please let me know.
Hibiscus
options crt;
in haus; |
Clarification of Answer by
hibiscus-ga
on
19 Nov 2003 21:24 PST
Like I say, my familiarity with other packages is not too great.
Using Excel for this sort of thing is much, much more cumbersome than
using TSP, SAS, or some other stats software. On the other hand, TSP
is rather horrible to learn since it is basically FORTRAN, a rather
arcane programming language to say the least!
|
