battenr

# Title: Overlap Weighting

# Description: A demonstration of overlap weighting using the propensity score

# Setup ----

#... Libraries ----

library(tidyverse) # ol' faithful
library(WeightIt) # for weighting

battenr

# Title: Parametric vs Nonparametric Statistics 

# Description: An example of what is actually meant by parametric with parametric statistics. 

# Setup ----

#... Libraries ----

library(tidyverse) # ol faithful
library(glue) # for easy string formatting

battenr

# Title: A collider and a confounder - To adjust or not adjust? 

# Description: Should we adjust for a variable if it's both a collider and a confounder? 
# The idea for this came from an article I recently read - Liu et al.(2012)
# https://academic.oup.com/aje/article/176/10/938/92975#246684420

# The idea for this code was to keep the situation simple to start. It can be built upon 
# rather easily. A few things to try altering would be the strength of the causal relationships, 
# continuous vs binary vs TTE variables, more/less variables, etc. 

battenr

# Title: Variable Selection - Model Misspecification 
# Subtitle: Two Variables Included That Shouldn't Be

# Description: The idea for this came from trying to show what happens 
# if the wrong variables are selected. Basically, when we use variable selection methods
# based on the p-value the results can be biased due to model misspecification. Furthermore, 
# how would that look if it wasn't just one variable that was wrong...but two!

# Setup ---- 

battenr

# Title: Plotting Propensity Scores and Weights 

# Description: Plotting propensity scores and weights is helpful for checking the 
# entire distribution and assumptions. This code gives an example of doing that. 

# Setup ----

library(tidyverse) # ol' faithful
library(WeightIt) # for using IPTW 
library(patchwork) # for combining plots

battenr

# Title: Understanding Distributions and Parameters

# Description: The below plots show which parameters are needed for four different
# distributions: normal, exponential, binomial and poisson distributions. 

# Setup ----

#... Libraries ----

library(tidyverse) # ol faithful

battenr

# Title: Conditioning on Descendant of a Collider Causes Bias

# Description: This script is intended to show that conditioning/adjusting
# for a descendant of a collider can induce bias. 

# Setup ----

library(tidyverse) # ol' faithful
library(ggdag) # for creating and showing DAG 
library(broom) # for tidying result of the model

battenr

# Title: Adjusting for Sunshine

# Description: Assessing whether adjusting for sunshine improves precision based on a question from David Okunlola on a LinkedIn Post

# Replicate 1000 times ----

# We could use one run as an example, however that is a single run. We need to do it  many more times than that. 
# Let's run it 1000 times. (Note: there are ways to calculate the number of runs required. For simplicity, we'll pick 1000 for now)

#... No Adjustment ----

battenr

# Title: Always Plot Your Data

# Description: A piece of advice I got from my supervisor was to always 
# plot your data first. This is invaluable. The code below is meant to show 
# one reason why 

# Setup ----

#... Libraries ----

battenr

# Title: Showing what the probability of the null is given our data 

# Description: Often the p-value is misinterpreted as the probability of the null. 
# The goal of this post is to show what the probability of the null actually is. 
# We can do this using Bayesian statistics. 

# Note: the p-value is the probability of a test-statistic at least as extreme assuming
# the null distribution

# Setup ----