R Data Analysis with OLS Regression
Budget: $30 – $250 USD
Two years since it first emerged, the Covid-19 pandemic continues to disrupt lives and livelihoods
across the planet. Almost a full year ago, vaccines started in a few select countries and, despite a very
uneven distribution across countries, over 4 billion people have received at least one dose (according to
the New York Times). However, several of the countries that have an abundance of available vaccines
are plateauing in their vaccine rollout. The United States is one of those countries: despite being the
first country to start vaccinating, only around 60 percent of the total population (70 percent of the
vaccine-eligible population) is fully vaccinated.
There are several potential reasons why people might hesitate to get vaccinated, or let their children
be vaccinated, against Covid-19. Some people may be worried about the fast rollout of the vaccine,
others are worried about side effects or efficacy, and yet others are affected by vaccine-critical news
coverage or partisan politics. Regardless of the cause of concern, however, researchers have noted that
various social-economic or demographic factors may influence individual decision-making. Poverty,
employment, education, and income can all affect an individual’s decision to get vaccinated or not.
Your task in this analysis will be to investigate, using OLS regression, which demographic factors best
explain an individuals’s vaccination decisions for themselves and their children. To that end, you have
been provided with a dataset, hps.csv, which is extracted from one of the waves of the Household Pulse
Survey, conducted by the US Census Bureau during between September 29 and October 11, 2021. It
includes survey responses from several thousand individuals. None of them have been vaccinated yet,
nor have any children in the household received the vaccine yet.
Dependent variable: You may choose one of two potential dependent variables for your analysis (see
important note below on their construction). Either you can study an individual’s decision to get
themselves vaccinated (GETVAC) or get their kids vaccinated (GETKIDVAC) Be sure to adjust your
theory about potential explanatory variables accordingly. Both variables are structured in the same
way, they are interval level variables, that measure an individual’s willingness to get themselves or
their children vaccinated. The scale ranges from 0-100, where 0 indicates an individual that will definitely
get (themselves or their kids) vaccinated and 100 indicates an individual who will definitely not
get themselves or their kids vaccinated.
Independent variables: You will have a variety of potential explanatory variables to choose from.
A codebook has been provided as a separate .pdf document, which lists the variable names in the
dataset, the wording of the question, and the way the reponses were coded. This codebook has been
slightly modified from the original version to fit the variables included. NOTE: Be careful to check
the codebook when interpreting your results. Sometimes, variables are coded such that 1=No and
0=Yes and sometimes it is the opposite (i.e., 1=Yes and 0=No).
across the planet. Almost a full year ago, vaccines started in a few select countries and, despite a very
uneven distribution across countries, over 4 billion people have received at least one dose (according to
the New York Times). However, several of the countries that have an abundance of available vaccines
are plateauing in their vaccine rollout. The United States is one of those countries: despite being the
first country to start vaccinating, only around 60 percent of the total population (70 percent of the
vaccine-eligible population) is fully vaccinated.
There are several potential reasons why people might hesitate to get vaccinated, or let their children
be vaccinated, against Covid-19. Some people may be worried about the fast rollout of the vaccine,
others are worried about side effects or efficacy, and yet others are affected by vaccine-critical news
coverage or partisan politics. Regardless of the cause of concern, however, researchers have noted that
various social-economic or demographic factors may influence individual decision-making. Poverty,
employment, education, and income can all affect an individual’s decision to get vaccinated or not.
Your task in this analysis will be to investigate, using OLS regression, which demographic factors best
explain an individuals’s vaccination decisions for themselves and their children. To that end, you have
been provided with a dataset, hps.csv, which is extracted from one of the waves of the Household Pulse
Survey, conducted by the US Census Bureau during between September 29 and October 11, 2021. It
includes survey responses from several thousand individuals. None of them have been vaccinated yet,
nor have any children in the household received the vaccine yet.
Dependent variable: You may choose one of two potential dependent variables for your analysis (see
important note below on their construction). Either you can study an individual’s decision to get
themselves vaccinated (GETVAC) or get their kids vaccinated (GETKIDVAC) Be sure to adjust your
theory about potential explanatory variables accordingly. Both variables are structured in the same
way, they are interval level variables, that measure an individual’s willingness to get themselves or
their children vaccinated. The scale ranges from 0-100, where 0 indicates an individual that will definitely
get (themselves or their kids) vaccinated and 100 indicates an individual who will definitely not
get themselves or their kids vaccinated.
Independent variables: You will have a variety of potential explanatory variables to choose from.
A codebook has been provided as a separate .pdf document, which lists the variable names in the
dataset, the wording of the question, and the way the reponses were coded. This codebook has been
slightly modified from the original version to fit the variables included. NOTE: Be careful to check
the codebook when interpreting your results. Sometimes, variables are coded such that 1=No and
0=Yes and sometimes it is the opposite (i.e., 1=Yes and 0=No).