# Recent AFNI Changes - Dec 2005

Submitted by rickr.
on 2005-12-29 17:18.

Updates to programs 3dANOVA, 3dANOVA2, and 3dANOVA3.

- The computation of variance for the t-statistic was altered. See http://afni.nimh.nih.gov/sscc/gangc/ANOVA_Mod.html for details.

- Options were added to allow use of the old method either for testing, or for allowing computations that assume sphericity.
: request to perform ANOVA using the previous functionality (requires**-old_method**, also)**-OK**: confirm that you understand issues with the old method of computing t-statistics**-OK**: assume sphericity using the old method (requires that contrasts sum to zero)**-assume_sph**- Options were added for second-order contrast testing. These apply to
, model types 4 and 5 only.**3dANOVA3** - compute 2nd order contrast in A, at fixed B level j
**-aBcontr c1 ... ca : j prefix**- compute 2nd order contrast in B, at fixed A level i
**-Abcontr i : c1 ... cb prefix**- Convenience options were added for the more trivial contrasts.
- compute the difference between levels a1 and a2 of factor A, with factor B fixed at level j (this is identical to using
with all zeros, except for c_a1 = 1, c_a2 = -1, and using the same j for the factor B level)**-aBcontr** **-aBdiff a1 a2 : j prefix**- compute the difference between levels b1 and b2 of factor B, with factor A fixed at level i (this is identical to using
with all zeros, except for c_b1 = 1, c_b2 = -1, and using the same i for the factor A level)**-Abcontr** **-Abdiff i : b1 b2 prefix**