Why Every Master’s Degree in Development Needs a Course on Data Quality
The importance of high-quality data has long been recognized by development researchers and practitioners. Having reliable data is essential for understanding trends in key indicators such as poverty, employment, and sustainable development goals, as well as for testing theoretical models and informing policy decisions.
However, the quality of the data used is sometimes questionable. While the costs of using poor-quality data can be difficult to quantify, numerous studies have shown the negative consequences.
A course on data quality is thus a must-have for every international development curriculum. These learning topics are critical for preparing students for a career in international development, and should include learning how to design surveys, selecting a survey firm, knowing how to adjust for biases in existing data, and understanding the influence of poor-quality data on decision making. In this blog post, I explore five reasons why every university that teaches international development should include a course on data quality.
- Understanding what the phrase ‘garbage in, garbage out’ actually means. Studying data quality issues can help students in critically assess the data they use for their own research as well as evaluate the validity of other work’s design and findings. In my field of applied development economics, I’ve frequently found that students are more interested in estimating the most advanced econometric models than in comprehending sample bias, measurement errors, and other possible data quality issues. Actually, they should be as concerned by data quality than by endogeneity concerns.
- There is no such thing as a perfect data set, however there are several approaches for evaluating data quality and adjusting data. Students should learn how to use advanced statistical techniques to improve the quality of their data sets.
- Recognizing that poor data quality can lead to flawed policy decisions. Students should learn how survey biases and other data quality issues can lead to incorrect conclusions about key indicators and to improper interpretation of empirical study results, thereby leading to poor policy recommendations. Poor policy choices based on bad data are common throughout history.
- Students who understand data quality will be better development practitioners. As data producers, they will design better surveys and field experiments to collect more accurate data. As data users and decision makers, they will grasp the limitations of the data they use and the implications for findings and decision making. As a client of a survey firm, they will understand of what is takes to produce high quality data, from the design to writing Terms of Reference to select a survey firm.
- As the development field grows more evidence and data-driven, it is more critical than ever for students to have a full awareness of data quality issues. This knowledge is simply essential for their career.
University is an excellent place to begin explaining how flawed policy recommendations based on poor data quality may result in missed opportunities for development and a waste of resources at the expense of the populations we are intended to help. The bottom line is that every development studies/economics student should learn about the importance of data quality, how surveys are (should be) developed and implemented, the value of reviewing data set documentation such as sample design and variable definitions (and footnotes!), and so on. Ensuring that all students are equipped with this knowledge and these skills is critical in a sector that is heavily data driven.
And if you are a student in development, don’t skip the course on data quality. Always explore your data, its biases and limitations just as much as you invest into the statistical and econometric approaches, if not more!
Author: Dr Johanna Choumert Nkolo, Director of Research