Using R for Analytic Graphs: Learn How Data Visualization Can Improve Interpretation in Social Work Research

Background and Purpose: Knowledge translation and research dissemination are critical for social work. The research needs to be accessible to the academy, community partners, and social service agencies. New research methodologies such as data visualization allow us to communicate results and prevent misinterpretation of complex data. These advances have been popularized by Tufte and Few, for example. The purpose of this workshop is to build social work research capacity by sharing reproducible data visualization methods that increase understanding of results.

This workshop will include an introduction to the free and open source statistical package R. This language is considered essential in biostatistics, political science, and demography, disciplines that are also interested in the social and behavioral importance of increased longevity. In R, researchers may work with multiple data sets simultaneously and use scripts to not only run statistical programs, but write presentations and papers.

In the workshop presentation, we will replicate results from published articles to show how data presented in tables may be more effectively presented as a graph or plot. We will provide examples of several graph types. First, for descriptive statistics we will show how to show proportions using mosaic plots and distributions using violin plots. Second, we will also show how to use a parallel dot plot to visualize regression coefficients and plot standard errors. Finally, we will demonstrate the use plots called “small multiples” for group comparisons.

Pedagogical Techniques: This workshop will be presented by social work researchers who utilize R as their primary tool and cover the following topics: 1) installing R and R-Studio, 2) the overall structure of the R/S language, 3) reading and writing data to generate descriptive statistics, 4) generating a simple linear regression model, and 5) graphing (e.g., mosaic plots, violin plots, dot plots, small multiples). We will have time to take questions from the audience.

The workshop pedagogy will include a demonstration and practical exercises with sample data sets. No prior knowledge of R is required. Individuals are encouraged to bring their own laptops (e.g., Linux, Mac OSX, and Windows) and will have the opportunity to install both R and workshop data so they may follow along.

Implications: Social work researchers who wish to be competitive in a contemporary interdisciplinary research environment should learn R because it saves time, money, and has the most powerful analytic graphing capabilities. In contrast to tables, graphs offer a more approachable and engaging method of conveying data in social work research. Graphs allow your audience to visually compare effect sizes and see statistical significance when standard errors are plotted. Currently, researchers present data in tables, but we will show why it can be more effective to present data in graphical form. Effective data visualization scripting in R can also facilitate collaboration through reproducibility.