Normal Distribution

[carriagc]

Anyone know if is posssible to know, using excel, if a certain group o
information (in my case, various numbers sorted in rows and columns
follow a statistical Normal distribution?


Thanks in Advanc

 

[Guest]

Have you tried using the statistical analysis tools?

Perform an analysis of variance (anova)

Correlation analysis tool

Covariance analysis tool

Descriptive Statistics analysis tool

Exponential Smoothing analysis tool

Fourier Analysis tool

F-Test: Two-Sample for Variances analysis tool

Histogram analysis tool

Moving Average analysis tool

Perform a t-Test analysis

Random Number Generation analysis tool

Rank and Percentile analysis tool

Regression analysis tool

Sampling analysis tool

z-Test: Two Sample for Means analysis tool

 

[Harlan Grove]

Have you tried using the statistical analysis tools?

...

And where in this list is there any entry for the two standard test for
distribution fitting: chi squared goodness of fit and Kolmogorov-Smirnov
statistics?

To the extent that you can use Excel to calculate chi squared goodness of fit
and Kolmogorov-Smirnov statistics for the hypothetical normal distribution from
the observed data, you can do this in Excel. If you don't know how to calculate
these, it's an open question whether you should be trying to fit distributions.

As a rough rule of thumb, if the coefficient of skewness of your sample is close
to zero and the coefficient of kurtosis is close to 3, you're probably safe
assuming the sample is from a normal distribution with the sample mean and
sample standard deviation.

 

[Michael R Middleton]

Harlan -

As a rough rule of thumb, if the coefficient of skewness of your sample is

close to zero and the coefficient of kurtosis is close to 3, you're probably
safe assuming the sample is from a normal distribution with the sample mean
and sample standard deviation. <

If you use Excel's KURT function, the online help shows an approximate
"minus three" term at the end, so perhaps a value of KURT close to zero
(along with SKEW = 0) indicates approximate normality. Yes?

- Mike

 

[Harlan Grove]

is
close to zero and the coefficient of kurtosis is close to 3, you're probably
safe assuming the sample is from a normal distribution with the sample mean
and sample standard deviation. <

If you use Excel's KURT function, the online help shows an approximate
"minus three" term at the end, so perhaps a value of KURT close to zero
(along with SKEW = 0) indicates approximate normality. Yes?

Oops! You're right. Guess it shows I don't use KURT().