cochran.Q {mada}

Cochran's Q statistic

Description

Given estimates from primary studies and the weights of the single studies calculate Cochran's Q as a measure of heterogeneity.

Usage

cochran.Q(x, weights)

Arguments

x

numeric, typically a vector of effect sizes like (log-)OR

weights

numeric, see Details

Details

In fixed effects settings the weights are often inverse proportional to the variances of the primary studies. Cochran's Q is known to have low power to detect heterogeneity.

Value

A named vector of length 3. First element is Q followed by the p-value and the degrees of freedom.

Author(s)

Philipp Doebler <[e-mail]>