# Ordered Beta Regression model: is a custom posterior predictive check plot necessary?

**URL:** https://discourse.mc-stan.org/t/ordered-beta-regression-model-is-a-custom-posterior-predictive-check-plot-necessary/28500
**Category:** Modeling
**Tags:** posterior-predictive, brms
**Created:** [August 10, 2022, 6:50pm UTC](https://discourse.mc-stan.org/t/ordered-beta-regression-model-is-a-custom-posterior-predictive-check-plot-necessary/28500 "2022-08-10T18:50:30Z")
**Posts on this page:** 1
**Showing post:** 4

<div class="post-metadata">

### Author: ![jd\_c](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mc-stan.org/jd_c/32/16759_2.png) [@jd\_c](https://discourse.mc-stan.org/u/jd_c)
#### Post date: [August 10, 2022, 9:43pm UTC](https://discourse.mc-stan.org/t/ordered-beta-regression-model-is-a-custom-posterior-predictive-check-plot-necessary/28500/4 "2022-08-10T21:43:56Z")

</div>

I am not sure how to make the custom plot you ask for, but the histogram seems to do a good job already. The density plot isn’t good for this data (as you noted).

I would be interested to see how the [zero-one-mid-inflated beta model](https://discourse.mc-stan.org/t/model-over-estimates-inflation-why/27355/11) works on this data from the vignette, if that spike near the midpoint is right at 0.5 and represents, say, neutral responses from a slider scale.

> [@arthur-albuquerque](#):
>
> probably because the model corresponds to a combination of continuous + discrete outcomes

Note this comment from Michael Betancourt [here](https://betanalpha.github.io/assets/case_studies/mixture_models.html) regarding mixture models:  
"In other words because the the inflated and non-inflated points are essentially modeled by separate data generating processes the entire model decomposes into a binomial model for the total inflated observations and a baseline model for the non-zero observations.

Because these two models are independent we can fit them independently or jointly, whichever is more convenient. In particular if the inflated counts are just a nuisance then we can ignore them entirely and just fit the non-inflated observations directly without any consideration of the inflation!"

That’s worth thinking about when you are fitting some slider scale data where people tend to plop the slider at units of 10 or whatever, with few responses in between.

---

_[View the full topic](https://discourse.mc-stan.org/t/ordered-beta-regression-model-is-a-custom-posterior-predictive-check-plot-necessary/28500)._