test file.Rmd

```{r, include = FALSE}
source("include/globals.R")

# needed to make the chapter (not visible)
library(kableExtra)
library(ggplot2)

```

```{r, include = FALSE}
# Add Type 1 error rate function
add_type1_error <- function(N, 
                            side = "right", 
                            col = rgb(1, 0, 0, 0.5)) {
  mult <- ifelse(side == "right", 1, -1)
  crit_d <- mult * abs(qt(0.05 / 2, (N * 2) - 2)) / sqrt(N / 2)
  
  if (side == "right") {
    y <- seq(crit_d, 10, length = 10000)
  } else {
    y <- seq(-10, crit_d, length = 10000)
  }
  
  # determine upperbounds polygon
  suppressWarnings({
    z <- (dt(y * sqrt(N / 2), df = (N * 2) - 2) * sqrt(N / 2))
  })
  
  if (side == "right") {
    polygon(c(crit_d, y, 10), c(0, z, 0), col = col)
  } else {
    polygon(c(y, crit_d, crit_d), c(z, 0, 0), col = col)
  }
}

# calculate distribution of d based on t-distribution
calc_d_dist <- function(x, N, ncp = 0) {
  suppressWarnings({
    # generates a lot of warnings sometimes
    dt(x * sqrt(N / 2), df = (N * 2) - 2, ncp = ncp) * sqrt(N / 2) 
  })
}
```

# test example 

## Enlish version

The interpretation of a *p*-value depends on the statistical philosophy one subscribes to. [Ronald Fisher](https://en.wikipedia.org/wiki/Ronald_Fisher) published 'Statistical Methods for Research Workers' in 1925 which popularized the use of *p*-values. In a Fisherian framework a *p*-value is interpreted as a descriptive continuous measure of compatibility between the observed data and the null hypothesis [@greenland_statistical_2016]. 

## Chinese version

通常，我们对*p*值的理解取决于个人所认同的统计理念。[Ronald Fisher](https://en.wikipedia.org/wiki/Ronald_Fisher)在1925年出版了《Statistical Methods for Research Workers》一书，该书向大众普及了*p*值的应用。在Fisher的理念中，*p*值被定义为所得数据和零假设之间的描述性连续测量(Greenland, 2016)。