- Various documentation fixes for re-submission of v0.6 to CRAN (first one since v0.4).
- Added
`summary`

option for`effect_draw`

. - Factor covariates will work (better) now.
- Removed some non-essential code for faster compilation on CRAN.

New

`"mutau_full"`

model is a generalisation of the`"mutau"`

model into individual-level data. The idea is similar as for the recent`"rubin_full"`

changes, see version 0.6.0.I also reparameterised the

`mutau`

model. It should be faster and have fewer divergent transition warnings.Some of the code around the mu and tau model has also been rewritten on the back end.

On the back end the package now follows the rstantools recommended way of compiling models. The user experience should be exactly the same, but this may avoid some problems when installing the package from GitHub or otherwise compiling it locally.

- Spike and slab model can be called via
`model="sslab"`

. See`?baggr`

for basics of working with this type of a model. A vignette will be added soon. - Rubin model with full data is now called via
`model="rubin_full"`

rather than`"full"`

. Old syntax will still work, however. Made some documentation and code improvements around this issue. - Leave-one-out cross-validation works for
`model="rubin_full"`

now. It works the same way as for`model="logit"`

. See`?baggr`

for more information on how to use it. - It’s now possible to use
`model="rubin"`

with the same inputs as`model="mutau"`

. Some data columns are removed automatically in that case.

For v0.6 we added more generic code around plotting, printing, grabbing treatment effects etc. While there are no differences on the front-end, this means that for the next versions we will be able to consider some new models and have more homogeneous syntax for all models.

- Fixed a few issues with formatting data for individual-level data models.
- Fixed a major bug with distributions of baselines in the
`rubin_full`

(`full`

) model. - Fixed glitchy display for some
`baggr_compare`

plots.

- Fixed and random effects for
`baggr`

models now have their own separate functions,`fixed_effects`

and`random_effects`

, in addition to`group_effects`

- LOO CV works for the logistic model (as does general cross-validation).
- Vignette for binary data analysis has been rewritten in parts.
- L’Abbe plots for binary data, see
`labbe()`

. - There is now more automatic conversion between summary-level and individual-level data for binary data (e.g. you can run
`baggr()`

with summary data and`model="logit"`

for automatic conversion) - For logistic model, priors can be specified for rates of events in the control arm, see arguments
`prior_control`

and`prior_control_sd`

in`baggr()`

- There are experimental features for working with models of quantiles. We advise against fitting such models using the package until these features have been fully tested and documented.

- Fixed some issues with printing of coefficients in meta-regressions, where wrong values were given for some models.

- Covariates can now be used in all baggr() models: in “rubin” model they give meta-regression (group-level covariates), while in “full” and “logit” models they can be used for “regular” regression (individual-level covariates)
- Priors for covariates are set through the argument prior_beta
- You can work with regression coefficients for covariates
- you can access and summarise coefficients through fixed_effects(),
- you will also see them when printing baggr objects;
- when using forest_plot() you can request
`show = "covariates"`

- Prototype of pp_check() function now works for Rubin model (thanks to Brice Green) you can apply it to generate new datasets according to posterior distribution of treatment effect and contrast them with the observed quantities as part of model checking
- baggr_compare() function now has standard output which you can print() or plot(), thanks to Brice Green
- Vignettes and documentation were updated to better describe binary data analysis
- We now give more warnings when plugging in stupid inputs.

- Messages for setting priors were accidentally given when e.g. running full pooling models
- All models were re-written to standardise our approach and syntax.
- “Full” model might now behave differently.
- “Mutau” model will be re-written and generalised for next release.
- LOO CV is also disabled for some models. Prompts will be given.

- Binary data models for both summary and individual-level data.
- New vignette for working with binary data; see
`vignette("baggr_binary")`

. - Expanded helper functions (esp.
`prepare_ma`

), esp. for prepping binary data. - Added forest plot functionality for all types of models.
- Various outputs can now be transformed (main use case is
`exp`

, but any transform is allowed). - Reworked vignette sections for pooling and cross-validation.
- Pooling statistics are now calculated for the whole model and better documented.
- More consistent theming, similar to bayesplot (thanks to Brice Green)
- Comparison of leave-one-out cross-validations with
`loo_compare`

(thanks to Brice Green)

- Re-enabled missing Cauchy priors

- Users can now define priors in
`baggr()`

using a syntax similar to`rstanarm`

. Extra priors are available `baggr()`

outputs prior predictive distributions; they can be examined using`baggr_compare`

and`effect_plot`

,`effect_draw`

– 2 new functions- More types of model comparisons are possible
- LOO CV function has been reworked
- Full pooling and no pooling models have been reworked to avoid divergent transitions.

First package version for CRAN.