---
title: "Adding additional levels of bias and indirectness"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Adding additional levels of bias and indirectness}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

```{r setup, include = F}
library(triangulate)
library(dplyr)
```

Say we wish to include 'Very high' or 'Critical' risk of bias as an option in our assessments. In this example, Domain 1 of Study 1 has a "Critical" risk of bias assessment

```{r}
dat_bias_with_critical <- triangulate::dat_bias
dat_bias_with_critical[1,6] <- 'Critical'

dat_bias_with_critical[1,1:6]
```

This requires us to add in another prior ('critical') to the library contained in `dat_bias_values`:

``` {r}
custom_bias_priors <- triangulate::dat_bias_values %>% 
  add_row(domain = 'all',
          j = 'critical',
          bias_m_add = 0.2,
          bias_v_add = 0.12,
          bias_m_prop = 0.08,
          bias_v_prop = 0.04)

custom_bias_priors
```

It is then just a case of passing our custom prior dataset to the prep functions:

```{r}
dat_bias_with_critical_prepped <- dat_bias_with_critical %>%
  tri_to_long() %>%
  tri_absolute_direction() %>%
  tri_append_bias(custom_bias_priors)

dat_bias_with_critical_prepped[1:7,c(2,6:13)]
```

We can see now that the first domain (D1) of Study 1 has been defined as Critical, and the adjustment values we added to our `custom_bias_priors` dataset have been correctly assigned to it.

The same approach is possible if you wish to change/supplement the levels used in the indirectness adjustment.