Note: while I asked if is ok or not to add a constant to the log response ratio here, with this question I focus specifically on the formula of the variance of the log response ratio when adding a constant.

@Wolfgang gave a very clear and useful answer to this similar question. My question builds on his answer: In his example, given the observed means $\bar{x}_1$ and $\bar{x}_2$, if one adds the same constant to the two means, say $\bar{x}_1+k$ and $\bar{x}_2+k$ , then is the variance of the log response ratio this?

$$ \mbox{Var}(\mbox{lnRR}) = \frac{s_1^2}{n_1 (\bar{x}_1+k)^2} + \frac{s_2^2}{n_2 (\bar{x}_2+k)^2} $$

I am also aware that adding a constant is not an encouraged fix to zero values: Koricheva, Handbook of Meta-analysis:

A consequence of using log-transformed ratios, however, is that the ratio must be positive […] In this case, the response ratio is an inappropriate measure and Hedges’ d is instead recommended. One should also not use a ratio as an effect size measure when either the numerator or denominator would be equal to zero; the transform is undefined and trying to adjust the values by adding a tiny fraction to the numerator and denominator usually results in abnormally large estimates of effect size.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.