Data management

An essential set of skills for data analysis

In this chapter we will introduce you to the basics of data management in R, using the powerful tidyverse framework. Data management is the process of importing, cleaning, transforming, and exporting data. This is an essential part of any data analysis project, where you will often have to deal with data that is messy, incomplete, or in the wrong format. Knowing your way around data management will save you a lot of time and frustration, and opens up new possibilities for analysis and visualization.

One of the greatest benefits of learning how to manage data in R is that it allows you to work with data in a structured and reproducible way. All the steps from importing the raw data to analyzing the final results are written in your script. If you made a mistake anywhere in the process, you can easily go back to fix it, and then rerun the script from the start. It is also common that you will need to rerun your analysis at a later time, for example when you get new data, or when someone (e.g., a client, a reviewer) asks you to make some changes. By the end of this chapter, you will be able to manage your data in a way that is transparent, reproducible, and efficient.

Practice data

Throughout this chapter, and further throughout this book, we’ll use a practice data set that contains simulated data. The data looks like real data used in communication science, but it’s entirely made up.

The benefit of using simulated data is that it’s tailored to give you the best learning experience. We can the re-use the same data across the learning material, so that you can focus on learning the concepts, rather than spending energy on getting acquinted with new data all the time. Just do keep in mind that this data is not real, and should not be used for any real-world analysis.

Description of the data

The practice data describes a fictional study about how entertainment media might influence people’s attitudes. In an experiment, 600 participants were shown hollywood movies with a positive, negative or neutral portrayal of journalists, and we measure whether this affects the participants’ trust in journalists before (trust_t1) and after (trust_t2) the experiment.¹

There were three types of movies (experiment_group): a control group that watched a neutral movie, a group that watched a movie with a positive portrayal of journalism, and a group that watched a movie with a negative portrayal of journalism. Participants were randomly assigned to one of these groups, with balanced group sizes (i.e., 200 participants per group). We also collected some demographic information (age), asked whether they have a newspaper subscription (np_subscription), and how much hours of news they consume per week (news consumption).

Since trust is a complex construct, we measured it with five items that participants rated on a scale from 1 to 10. These items are based on the questions proposed by Strömbäck et al. (2020), but with some modifications. For example, we inversed the scale for one of the items, so that higher values indicate lower trust. The items are:

1. Journalists are fair when covering the news
2. Journalists are unbiased when covering the news
3. (inversed) Journalists do not tell the whole story when covering the news
4. Journalists are accurate when covering the news
5. Journalists separate facts from opinion when covering the news

Variables

Variable name	Type	Description
id	numeric	Unique identifier for each participant
age	numeric	Age of the participant in years
political_orientation	character	Political orientation of the participant (left, center, right)
np_subscription	factor	Whether the participant has a newspaper subscription (yes, no)
news consumption	numeric	Average number of hours per week the participant consumes news
experiment_group	factor	Type of movie shown: neutral (control group), positive (positive portrayal of journalism), negative (negative portrayal of journalism)
trust_t1	numeric	Trust in journalists before the experiment (range: 1-10)
trust_t2	numeric	Trust in journalists after the experiment (range: 1-10)
trust_t1_item[1-5]	numeric	Scale items for trust_t1
trust_t2_item[1-5]	numeric	Scale items for trust_t2

Reading the data into R

The following code reads the practice data into R. We will explain this code in the next tutorials, and include it at the top of every tutorial that uses it. Note that in order to run this code, you need to have installed the tidyverse package.

library(tidyverse)
d <- read_csv('https://tinyurl.com/R-practice-data')

This is what the full data looks like:

References

Strömbäck, Jesper, Yariv Tsfati, Hajo Boomgaarden, Alyt Damstra, Elina Lindgren, Rens Vliegenthart, and Torun Lindholm. 2020. “News Media Trust and Its Impact on Media Use: Toward a Framework for Future Research.” Annals of the International Communication Association 44 (2): 139–56.

Footnotes

1. The t1 in trust_t1 stands for “time 1”. This is a common notation in studies where you measure the same construct at different points in time. Here t1 and t2 refer to the time before and after the experiment, respectively.