2B Lab 4 Week 5

This is the pair coding activity related to Chapter 11.

Task 1: Open the R project for the lab

Task 2: Create a new .Rmd file

Task 3: Load in the library and read in the data

The data should already be in your project folder. If you want a fresh copy, you can download the data again here: data_pair_coding.

We are using the packages tidyverse, sjPlot, and performance today.

Just like last week, we also need to read in dog_data_clean_wide.csv.

Task 4: Tidy data & Selecting variables of interest

Let’s define a potential research question:

To what extent do pre-intervention loneliness and pre-intervention flourishing predict post-intervention loneliness, and is there an interaction between these predictors?

To get the data into shape, we should select our variables of interest from dog_data_wide and remove any missing values .

Solution

library(tidyverse)
library(sjPlot)
library(performance)

dog_data_wide <- read_csv("dog_data_clean_wide.csv")

dog_mult_reg <- dog_data_wide %>%
  select(RID, Loneliness_post, Loneliness_pre, Flourishing_pre) %>% 
  drop_na() 

Furthermore, we need to mean-center our two continuous predictors. Since this is a new concept, simply run the code below.

dog_mult_reg <- dog_mult_reg %>% 
  mutate(Flourishing_pre_centered = Flourishing_pre - mean(Flourishing_pre),
         Loneliness_pre_centered = Loneliness_pre - mean(Loneliness_pre))

Task 5: Model creating & Assumption checks

Now, let’s create our regression model. This follows the same approach as Chapter 10, but with additional predictors.

According to our research question, we have the following model variables:

• Dependent Variable (DV)/Outcome: Loneliness post intervention
• Independent Variable (IV1)/Predictor1: Flouring before the intervention
• Independent Variable (IV2)/Predictor2: Loneliness before the intervention
• Does our model require an interaction term? Yes No

As a reminder, the multiple linear regression model has the following structure:

lm(Outcome ~ Predictor1 * Predictor2, data)

The asterisk (*) means that the model includes main effects for both predictors (i.e., Pre-intervention flourishing & Pre-intervention loneliness) as well as their interaction term (which tests whether the effect of one predictor depends on the other).

Your Turn

Step 1: Create the model

Compute the linear regression model using the formula above. Store the model in an object called mod, ensuring that you use the mean-centered predictors rather than the unstandardised values.

Step 2: Run the regression

Just like last week, use the summary() function on mod to display the regression output.

Step 3: Check assumption

Use the check_model() function from the performance package to test whether the model meets its assumptions.

Solution

# Step 1
mod <- lm(Loneliness_post~Loneliness_pre_centered*Flourishing_pre_centered, data = dog_mult_reg)

# Step 2
summary(mod)

# Step 3
check_model(mod)

Now, answer the following questions:

1. Are all of the assumptions met? Yes No

2. How much of the variance is explained by the model? Enter the percentage value with 2 decimal places. %

3. Is the interaction term statistically significant? Yes No

4. What is the p-value for the interaction term? Enter the value with 3 decimal places.

5. What is \(\beta\) coefficient for pre-intervention loneliness? Enter the value rounded to two decimal places.