Determinants of COVID-19 vaccination intention

Multiple linear regression
COVID-19
Search the determinants of COVID-19 vaccination intention using multiple regression analysis
Author

Elías Aburto

Published

March 12, 2023

Research information

Original Research: Reactance and perceived disease severity as determinants of COVID-19 vaccination intention: An application of the theory of planned behavior

Authors: Dariusz Drążkowski, Radosław Trepanowski, Adam Mickiewicz

Year: 2021

Sample: 551 Polish participants

DOI: https://doi.org/10.1080/13548506.2021.2014060

Multiple Regression Analysis: Step by step

Problem What are the determinants of COVID-19 vaccination intention?

Objective: To identify the determinants of COVID-19 vaccination through an application of the Theory of Planned Behavior.

Hypothesis: H1: Control beliefs have a direct impact on COVID-19 vaccination intention H2: Utility beliefs have a direct impact on COVID-19 vaccination intention H3: Social norm beliefs have a direct impact on COVID-19 vaccination intention

Steps to prove the hypothesis

A. Model Evaluation

  1. Regression coeficients and r-square
  2. Interpretation

B. Regression Model Assumtions

  1. Lineality between IVs and Dv
  2. Independence of observations: The observation from our model are independent.
  3. Homoscedasticity: The errors from our model have equal variance.
  4. Normality of Errors: The errors from our model are normally distributed.
  5. Multicollinality: evaluate if the IVs are redundant.

Load the packages

library("foreign")
library(ggplot2)
library(dplyr)
library(broom)
library(ggpubr)
library(tidyverse)
library(readr)
library(car)

A. Model Evaluation

A1. Regression coeficients and r-square

Load the database

Mydata=read.spss("data/COVID_data.sav",to.data.frame=T,use.value.labels=FALSE)
summary(Mydata$SE_Total)
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
   3.00   12.00   15.00   15.42   18.00   21.00

Perform the linear regression analysis

Intention_lm <- lm(IN_Total ~ SN_Total + BC_Total + AT_Total, data = Mydata)
summary(Intention_lm)

Call:
lm(formula = IN_Total ~ SN_Total + BC_Total + AT_Total, data = Mydata)

Residuals:
     Min       1Q   Median       3Q      Max 
-11.4622  -1.2412   0.1199   1.3464  11.0907 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) -1.71571    0.36855  -4.655 4.07e-06 ***
SN_Total     0.36218    0.03976   9.109  < 2e-16 ***
BC_Total     0.29052    0.03483   8.341 6.00e-16 ***
AT_Total     0.45369    0.03504  12.946  < 2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 2.457 on 547 degrees of freedom
Multiple R-squared:  0.7865,    Adjusted R-squared:  0.7853 
F-statistic: 671.5 on 3 and 547 DF,  p-value: < 2.2e-16

A2. Interpretation

Bs ajusted:

For each point obtained in the scale of subjective norms, the intenton to COVID-19 vaccination intention will increase 0.36 points. Also, subjective norms had a direct effect on COVID-19 vaccination intention it was statistically significant.

For each point obtained in the scale of Behavioral Control, the intenton to COVID-19 vaccination intention will increase 0.29 points. Also, Behavioral Control had a direct effect on COVID-19 vaccination intention and it was statistically significant.

For each point obtained in the scale of Attitude Toward Covid-19 vaccination, the intenton to COVID-19 vaccination intention will increase 0.45 points. Also, Attitude Toward Covid-19 vaccination had a direct effect on COVID-19 vaccination intention and statistical significance.

The model with these three determinants explain 78% of variance of COVID-19 vaccination intention.

B. Regression Model Assumtions

B1. Lineality between IVs and Dv

plot(IN_Total ~ SN_Total, data = Mydata)

plot(IN_Total ~ BC_Total, data = Mydata)

plot(IN_Total ~ AT_Total, data = Mydata)

B2. Independence of observations

The observation from our model are independent.

This was fulfilled when each observation was made by one participant.

B3. Homocedasticity

The errors from our model have equal variance.

par(mfrow=c(2,2))
plot(Intention_lm)

par(mfrow=c(1,1))

B4. Normality of Errors

The errors from our model are normally distributed.

par(mfrow=c(2,2))
plot(Intention_lm)

par(mfrow=c(1,1))

B5. Multicollinality: evaluate if the IVs are redundant.

vif(Intention_lm)
SN_Total BC_Total AT_Total 
2.984366 1.899331 2.991759