Quote:
Originally Posted by Jamal NUMAN
Could you please elaborate more regarding the method that you have used to check whether the “marks” are normally distributed and namely the equation?
=8.74-ABS(AVERAGE(A$2:A$20)-A2)
What does 8.74 represent?
|
As indicated in my previous post:
Quote:
|
The '8.74' is just to 'normalise' the balance of the formula to a (rough) zero base
|
It has no other significance.
The figures I posted were not those of an actual normal distribution curve - they were just my manual approximation of what one might look like for data spanning the number range you posted. The actual curve might be flatter or steeper than my figures suggest. In any event, for so few results, it would be unreasonable to expect a close approximation of the true normal distribution curve. You might need a few hundred results before you would have a really close approximation of the true normal distribution curve - it all depends on how variable the data are. For example, to attain 90% confidence that your sample data are within 5% of a normal distribution, you'd need a minimum sample size of 271. 95% confidence would require a minimum sample size of 385.
If you need to analyse the data statistically, you'll need some understanding of the theory. You should perhaps get a copy of Freund's
Modern Elementary Statisics. See also:
http://en.wikipedia.org/wiki/Normal_distribution
http://en.wikipedia.org/wiki/Sample_size