# Model comparison via PSIS-LOO for 3-level hierarchical growth model

**URL:** <https://discourse.mc-stan.org/t/model-comparison-via-psis-loo-for-3-level-hierarchical-growth-model/2905>\
**Category:** Modeling\
**Tags:** loo\
**Created:** [January 5, 2018, 8:31am UTC](https://discourse.mc-stan.org/t/model-comparison-via-psis-loo-for-3-level-hierarchical-growth-model/2905 "2018-01-05T08:31:24Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![avehtari](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mc-stan.org/avehtari/32/5935_2.png) [@avehtari](https://discourse.mc-stan.org/u/avehtari)\
**Post date:** [January 5, 2018, 3:37pm UTC](https://discourse.mc-stan.org/t/model-comparison-via-psis-loo-for-3-level-hierarchical-growth-model/2905/2 "2018-01-05T15:37:02Z")

</div>

Cool model.

> [@laws1411](#):
>
> Is that mean, I am actually not comparing apple to apple (due to the logarithm in the log-linear model) via PSIS-LOO?

This is correct. Both models need to use same y (ie no transformations) or you need to take into account the Jacobian of the transformation. You can add that (log) Jacobian term to your log\_lik computation in the generated quantities.

---

_[View the full topic](https://discourse.mc-stan.org/t/model-comparison-via-psis-loo-for-3-level-hierarchical-growth-model/2905)._