0
$\begingroup$

In complement to Aikake Information Criterion, I want to use posterior probability to select the best model for group-based trajectory modeling. In this paper:

https://drc.bmj.com/content/bmjdrc/4/1/e000243/DC1/embed/inline-supplementary-material-1.pdf?download=true

Which is a supplement to the methods section. The paragraph to which I'm referring is: "We used the Bayesian information criterion (BIC) as the criterion for model selection. However, this was moderated by /../ (c) an average posterior probability (AvePP) value >0.7 for each group; /../ and (f) the odds of correct classification based on the posterior probabilities of group membership >5 for each group".

the average posterior probability value is set at >0.7 - how is this value selected? I understand that probability is a value which ranges between 0-1 and that a model of 0.5 is a very good model predictor. Therefore, should the selected value not be in the range of >0.3 and <0.7? I have been unable to find other papers using a similar approach (in which to gage what the "standard" selected value is.

$\endgroup$
  • $\begingroup$ The question does not make sense unless one reads the paper in the link, so please produce some summary of said paper that makes the question self-contained. $\endgroup$ – Xi'an Jan 3 at 7:53
  • $\begingroup$ Hi Xi'an. The paper is a description of how group-based trajectory modelling was executed (as a supplement to the results section). The paragraph to which I'm referring is: "We used the Bayesian information criterion (BIC) as the criterion for model selection. However, this was moderated by /../ (c) an average posterior probability (AvePP) value >0.7 for each group; /../ and (f) the odds of correct classification based on the posterior probabilities of group membership >5 for each group". $\endgroup$ – JennyElsa Jan 3 at 10:20
  • $\begingroup$ To clarify. I am not asking why this particular paper has selected a value of 0.7. Rather, I am asking whether 0.7 is the standard "cut-off point" for Bayesian statistics, similar to a probability value of 0.05. $\endgroup$ – JennyElsa Jan 3 at 10:25
  • $\begingroup$ Please add this additional information in the question rather than as a comment. $\endgroup$ – Xi'an Jan 3 at 10:26

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.