In response to a previous question posed about effect sizes with a generalized linear mixed model with a binomial distribution, it's been clarified that estimating effect sizes is not straightforward. I have a dataset with continuous predictors where family = gamma (link = "log") for each of my response variables.
After conducting model selection for each response, I've identified one predictor that consistently appears in my best models. Is there a method to visually represent the magnitude of the relationship between this predictor and each response variable, akin to using sjPlot::plot_model? Although I've already utilized ggeffects to graphically depict the model, I'm seeking a straightforward approach to demonstrate the strength of the relationship between this predictor and each response.
