# Modelling accuracy scores with ceiling effects

**URL:** <https://discourse.mc-stan.org/t/modelling-accuracy-scores-with-ceiling-effects/24966>\
**Category:** Modeling\
**Tags:** brms\
**Created:** [October 23, 2021, 6:10am UTC](https://discourse.mc-stan.org/t/modelling-accuracy-scores-with-ceiling-effects/24966 "2021-10-23T06:10:30Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![LachlanC](https://avatars.discourse-cdn.com/v4/letter/l/ea5d25/32.png) [@LachlanC](https://discourse.mc-stan.org/u/LachlanC)\
**Post date:** [October 23, 2021, 6:20am UTC](https://discourse.mc-stan.org/t/modelling-accuracy-scores-with-ceiling-effects/24966/2 "2021-10-23T06:20:46Z")

</div>

As it happens, it seems like there was a very recent post discussing a similar issue which I have missed: [Non linear distribution with some zeros in data](https://discourse.mc-stan.org/t/non-linear-distribution-with-some-zeros-in-data/24915)

In this case, I believe that I _do_ have access to the number of correct responses as raw data. So my question may be solved. A natural option may be to use a binomial regression with _p_ correct responses of _n_ trials.

But if anyone has any other recommendations, I am all ears!

---

_[View the full topic](https://discourse.mc-stan.org/t/modelling-accuracy-scores-with-ceiling-effects/24966)._