.-
help for ^normtest^
by Herve M. CACI, September 2000 (v 1.2)
.-

Tests of univariate/multivariate kurtosis and skewness
------------------------------------------------------

     ^normtest^ varlist [^if^ exp][^in^ range]

Description
-----------

Computes measures and tests of univariate skew (g1, sqrt(b1)) and
kurtosis (g2 and b2-3), and two omnibus tests of normality: namely
D'agostino & Pearson K sq, and Jarque-Bera LM.

Fisher's g statistics are given, while "sqrt(b1)" and "b2" are
already computed by Stata's ^summarize^ command when specifying
the "detail" option.

D'agostino & Pearson Ksq is similar to Shapiro-Wilk W.

Computes measures and tests of multivariate skew (Small's Q1 and
Srivastava's b1p) and kurtosis (Small's Q2, Srivastava's b2p, and
Mardia's b2p), plus an omnibus test which combines VQ1 and VQ2.

Finally shows the 5 observations with larger Mahalanobis
distances, and graphs the ordered squared distances.

Note
----

This ado draws on a SPSS macro by Prof. Lawrence DeCarlo
(decarlo@exchange.tc.columbia.edu). I'm also grateful to Brendan
Halpin (brendan@essex.ac.uk).

Also see
--------

 Manual:  ^[R] sktest^
          ^[R] swilk^
          ^[R] sfrancia^

 Lawrence T. DeCarlo (1997) "On the meaning and use of kurtosis"
 Psychological Methods, 2(3), pp292-307.