Skip to content
Prev 18576 / 20628 Next

Minimum detectable effect size in linear mixed model

Dear Han,

As mentioned earlier, power analysis is only relevant _before_ you do the
study. To avoid running an underpowered study. Doing a post-hoc power
analysis on an underpowered study is putting the cart before the horse.

Once you have done the analysis, look at the confidence intervals of the
estimates.
- non-significant and values in the CI small compared to the
practical range: sufficient power
- non-significant and values in the CI similar or larger than the
practical range: underpowered
- significant and values in the CI similar or larger than the practical
range: sufficient power
- significant and values in the CI small compared to the practical range:
overpowered

Note that you should not only vary the coefficient of interest. At least
also take the uncertainty of the random effect variance into account. Don't
underestimate its effect on the power. The uncertainty on these variances
can be substantial. Especially when the design has a small (<200) number of
levels for the random effect.

Best regards,

ir. Thierry Onkelinx
Statisticus / Statistician

Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
FOREST
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx at inbo.be
Havenlaan 88 bus 73, 1000 Brussel
www.inbo.be

///////////////////////////////////////////////////////////////////////////////////////////
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
///////////////////////////////////////////////////////////////////////////////////////////


<https://www.inbo.be>


Op za 4 jul. 2020 om 23:27 schreef Han Zhang <hanzh at umich.edu>: