CODEWITHSUNDEEP
rbayesian
Dominik Vogel
Dominik Vogel

Reputation: 316

Bayesian meta-analytic t tests - mix one- and two-sample (BayesFactor)

The BayesFactor package offers a function to estimate a Bayes Factor for a meta-analytic t-test (meta.ttestBF). The procedure is based on an article by Rouder and Morey (2011; link).

meta.ttestBF can be used to get a Bayes Factor for a series of "one- and two-sample designs." I assume this refers to paired and independent t-tests.

My question is if there is a way to mix t values from one-sample and two-sample designs. Let's assume I have four studies. Three of them use a two-sample design, and one uses a one-sample design. If all of them were of the same type, it would be easy:

t <- c(-.15, 2.39, 2.42, 2.43)
N1 <- c(100, 150, 97, 99)
N2 <- c(100, 150, 97, 99)

# One-sample
meta.ttestBF(t = t, n1 = N1)

# Two-sample
meta.ttestBF(t = t, n1 = N1, n2 = N2)

So, what should I do if t values 1 to 3 are from a two-sample t-test and t value 4 is from a one-sample t-test?

Upvotes: 0

Views: 68

Answers (1)

richarddmorey
richarddmorey

Reputation: 1170

I have answered this question here: https://github.com/richarddmorey/BayesFactor/issues/122. The short answer is "not really, because effect sizes have different interpretations in the two models."

Upvotes: 1

Related Questions