randomization.Rmd

---
title: How to block and randomize
author:
- name: Nathaniel Higgins
  affiliation: SBST
- name: Jake Bowers
  affiliation: SBST and University of Illinois
date: '`r format(Sys.Date(), "%B %d, %Y")`'
bibliography: references.bib
output:
    html_document:
      theme: cosmo
      toc: yes
    pdf_document:
      keep_tex: true
      number_sections: true
      fig_width: 5
      fig_height: 5
      fig_caption: true
      template: bowersarticletemplate.latex
...



```{r include=FALSE, cache=FALSE}
# Some customization.  You can alter or delete as desired (if you know what you are doing).
# knitr settings to control how R chunks work.
## To make the html file do
## render("thefile.Rmd",output_format=html_document(fig_retina=FALSE))
## To make the pdf file do
## render("thefile.Rmd",output_format=pdf_document())
require(knitr)
opts_chunk$set(tidy=TRUE,echo=TRUE,results='markup',strip.white=TRUE,cache=FALSE,highlight=TRUE,width.cutoff=60,size='footnotesize',out.width='.5\\textwidth',message=FALSE,comment=NA,fig.env="figure", fig.align="center",fig.lp="fig:",fig.pos="H")
```

Before you randomize, you need to know how you want to organize the randomization, and how you plan to evaluate a better or worse randomization.

# Blocking

```{r}
library(blockTools)
```

Evaluating blockings

```{r}
library(RItools)
```

# Randomizing

First, we show how to do this by hand.

## Complete Randomization

This is the most common type of randomization because it guarantees the same numbers of observations in each treatment arm.

```{r}

N <- 20
numtrts <- 2
stopifnot( N %% numtrts == 0)
set.seed(180046)
Z1 <- sample(rep(seq(1,numtrts),N/numtrts))
Z1[1:10]
table(Z1)

```

Currently proposing to use the `randomizr` package

```{r}
library(randomizr)
set.seed(180046)
Z2 <- complete_ra(N=N,num_arms=numtrts)
Z2[1:10]
table(Z2)

```




# References