Statscraft 0.15.0, release v0.15.0
BAYES INDEPENDENT
Bayesian comparison of the means of two independent groups.
Syntax
BAYES INDEPENDENT /INFERENCE VARIABLES=varlist GROUP=var
SELECT=LEVEL(a b) | MIDPOINT | CUTPOINT(v)
[ANALYSIS=POSTERIOR | BAYESFACTOR | BOTH] [DATAVAR=v1 v2]
[/ESTBF COMPUTATION=ROUDER[(1)] | GONEN([0] variance) | HYPER[(-0.75)]]
[/PRIOR [EQUALDATAVAR=FALSE | TRUE] [VARDIST=DIFFUSE | JEFFREYS | INVCHISQ(df scale)]
[MEANDIST=DIFFUSE | NORMAL(mean1 mean2 variance1 variance2)]]
[/CRITERIA [CILEVEL=95] [TOL=0.000001] [MAXITER=2000]] [/MISSING SCOPE=ANALYSIS | LISTWISE].
LEVEL takes the two values of GROUP that form group 1 and group 2; MIDPOINT (the middle of
the range of GROUP) and CUTPOINT put values below the point in group 1 and the rest in
group 2. The difference is group 2 minus group 1. Bayes factors use the pooled t: ROUDER
gives BF01 for a Cauchy prior of that scale on the effect size, GONEN BF01 for a normal
prior with that mean and variance, HYPER BF10 for a hyper-prior with parameter a (above -1,
at most -0.5). The posterior of the difference: with DATAVAR (known variances) normal, from
flat or NORMAL mean priors; else with EQUALDATAVAR=TRUE a t from VARDIST=JEFFREYS or
INVCHISQ, and in all other cases the unequal-variances posterior of flat priors (each
group needs more than five cases). Weights as on BAYES ONESAMPLE.
Example
BAYES INDEPENDENT /INFERENCE VARIABLES=yield GROUP=fertilizer SELECT=LEVEL(1 2)
ANALYSIS=BOTH /ESTBF COMPUTATION=GONEN(0.5)
/PRIOR EQUALDATAVAR=TRUE VARDIST=JEFFREYS.
statscraft help BAYES INDEPENDENT prints the same text in the terminal. All commands