# Extra stats

This menu item adds some handy statistics to Excel.

**Autocorrelation**Enters an array formula that returns the
autocorrelation function for a data vector. The first element of
the result is the correlation at lag 0, the next is the correlation
at lag 1, and so on (see also function dACF).

**ANOVA** Enters an array formula that returns a one-way ANOVA (Type
II sums of squares) for balanced or unbalanced data. The output is
a 4 row x 6 column ANOVA table. Unlike the Excel built in routines,
the output is dynamic and unequal group sizes are allowed (see also
function ANOVA).

**Chi square** Enters an array formula that gives the Chi squared
value and associated *P*-value) for a set of observed data.

**Regression** This procedure replaces Excelâ€™s procedure for
linear regression (LINEST). It calculates the results in extended
precision and also gives correct *R*^{2} values if the intercept is
fixed (see also function Regress).

**Goodness-of-fit.** This routine computes the G-statistic (see
Sokal, R. R. & Rohlf, F. J. ,1995,
*Biometry: the principles and practice of statistics in biological research*.
W. H. Freeman and Company, New York) for comparing the
goodness-of-fit of a set of observed data to some specified model
(or frequency distribution).

**G-test.** This routine computes the G-statistic to determine if
the frequency distributions of two or more groups of observed data
differ.

**Mantel test** This routine does a computation of the Mantel test
on large matrices. It is fast and handles large matrices (tested on
matrices of random variables as large as 256 x 256 ). When invoked
you will be asked to nominate ranges containing two distance
matrices. These are copied to text files (this is slow for large
matrices - so be patient) and then loaded by the PopTools DLL
(RANDEVS.DLL). The DLL computes the test and you can then copy the
results of the randomisation procedure back into the spreadsheet.
The result of the test is a sorted vector of correlation
coefficients for each randomisation, with the correlation
coefficient of the original matrix as the first element in the
vector. See also function MANTEL function.

**Principal components analysis.** Enters an array formula for a
dynamic PCA of a data matrix (see also function dPCA).

**Summary Stats**. See this page.