analysis.Rmd

---
title: "CTN Openness"
author: "Deb Lindsay"
date: "`r Sys.Date()`"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(
	echo = TRUE,
	message = FALSE,
	warning = FALSE
)
#required packages
library(tidyverse)
library(Hmisc) #correlations
library(car) #VIF assumptions & allows type II Anova 
library(relaimpo) #calculates the R2 & Beta2 for the regression
library(effectsize) #calculate the Beta2 effect size

#import the clean data set
data <-
     read_csv("clean_data.csv")

```

## Assumptions for running regressions

Linear Regression Assumptions 

- Normality of predictors (Shapiro Wilkes test)
     - Use shapiro.test() 
     - Should be non-significant (there is no skewedness)


```{r Shapiro Wilk}
#Openness
shapiro.test(data$openness)
#non-normal

data %>% 
     ggplot()+
     geom_density(aes(openness))


#CTN Overall
shapiro.test(data$CTN_overall)

#CTN Deep

#CTN Experience

#CTN Emotional

#CTN Presence

```

# Hypothesis 1 - Positive relationship between trait openness and overall CTN.

```{r H1 }
#Linear regression predicting openness from CTN
# Openness ~ CTN
#outcome variable ~ predictor variable

# lm() to run the linear regression
h1_model <- 
     lm(openness_z ~ CTN_overall_z, data = data)


# Anova() to calculate the F value and significance
Anova(h1_model)

# summary() to calculate the variance accounted for by the model (in this case the single variable CTN overall)
summary(h1_model)

#effect size (Beta sq) 
effectsize(h1_model)


#plot relationship
data %>% 
     ggplot(aes(x = CTN_overall_z, y = openness_z)) +
     geom_jitter() +
     geom_smooth(method = "lm")



```

```{r H2}
# Muiltiple Linear regression predicting openness from all four dimensions of DEEP CTN
# Openness ~ Deep + Experience + Emotional + Presence
#outcome variable ~ predictor variables (use + to add additional predictors)

# lm() to run the linear regression
h2_model <- 
     lm(openness_z ~ CTN_Deep_z + CTN_Exp_z + CTN_Emo_z + CTN_Pres_z, data = data)

# Anova() to calculate the variance accounted for by CTN
#use Type II to calcaulte the variance equally among variables (otherwise the model will calculate the variance by the order they are in the model)
Anova(h2_model, Type = "II")

# calc.relimpo() to calculate the unique variance & effect size (Beta sq) 
# type = lmg gives teh unique R2 (variance) of each variable in the model

calc.relimp(h2_model, type = "lmg")

# plot effect sizes

# create table

```