How to check normality in SPSS?

[Lawrence]

Hello Guys!

How to check the normality other than Box Plot, Histogram and
Stem&Leaf?

Thanks,

Lawrence

[Bruce Weaver]

Statistical tests of normality are usually not very useful. If you are
concerned about meeting the (approximate) normality assumption for some
statistical test, the test of normality will be under-powered for small
samples, and (vastly) over-powered for large samples. Graphical methods
of assessing normality are usually much better. Try the Q-Q plot.

Graphs -> Q-Q Plot

--
Bruce Weaver
bwe...@lakeheadu.ca
www.angelfire.com/wv/bwhomedir

[Lawrence]

Hello Bruce,

Thanks for your reply. In what way you are saying "Statistical tests of
normality are usually not very useful"? According to the distribution
theory, it is the most powerful. Many of the Statisticcal techniques
based on the normality assumption. I aggreed the graphical
representation gives overall picture. But we are unable to find exactly
what is the normlity values for particular variable/data. In the theory
we are saying Normality => Mean=Median=Mode. In all the cases it should
not equal. If it is vary how we interpret?

Thanks,

Lawrence

[Bruce Weaver]

Lawrence wrote:
> Hello Bruce,
>
> Thanks for your reply. In what way you are saying "Statistical tests of
> normality are usually not very useful"?

"If you are concerned about meeting the (approximate) normality

assumption for some statistical test, the test of normality will be

under-powered for small samples..."

I.e., you will often fail to detect important departures from normality
when sample sizes are small.

"...and (vastly) over-powered for large samples."

I.e., the test of normality will often be significant for unimportant
departures from normality when sample sizes are large.

Also, remember that nothing in nature is truly normally distributed. So
all of your tests that assume normality are really approximate tests,
and the question you should be asking yourself is whether the
approximation is useful. See Box's 1976 paper "Science and Statistics",
available through JSTOR.

> According to the distribution
> theory, it is the most powerful. Many of the Statisticcal techniques
> based on the normality assumption. I aggreed the graphical
> representation gives overall picture. But we are unable to find exactly
> what is the normlity values for particular variable/data. In the theory
> we are saying Normality => Mean=Median=Mode. In all the cases it should
> not equal. If it is vary how we interpret?
>
> Thanks,
>
> Lawrence
>

[Lawrence]

Thanks a lot Bruce.

Lawrence