⁸Egger’s regression model can be represented as _i = α + ρSe_i + ϵ_i with the variance of ϵ_i proportional to 1∕Se_i² where _i and Se_i are the observed effect size and the associated standard error obtained from study i respectively, α is the intercept and ϵ_i is the error term. The bias is measured by ρ. If ρ is significantly different from zero, this is a sign of publication bias (Peters et al. 2010)