Crosswalking an Economic Security Dataset

This resource is intended to support researchers who want to use the U.S. Survey of Household Economics and Decisionmaking data.

by Shaun FieldJoseph de la Torre DwyerKate FlinnerUduak Grace Thomas
Feb 9, 2021

As part of A New History of Investment in Financial Education across the United States project, Knology used a public dataset from the United States Federal Reserve known as the Survey of Household Economics and Decisionmaking (SHED). This dataset has measured the economic well-being of U.S. households and potential risks to their finances since 2013. Based on our experiences with analyzing the data, we have put together a resource to help guide other researchers who are seeking to use the SHED in their projects.

The Federal Reserve collects and publishes data featured in the SHED once each year. With some variation, the dataset measures many of the same things over and over so that researchers can understand change over time. One important part of using a public data source in research is ensuring the consistency and comparability of the data. Are the survey questions asking for the same information each year? Are the questions rephrased in ways that might shift responses from year to year?

If longitudinally collected data is not consistent year-over-year, researchers may not be able to accurately study change over time. A vital step to ensure the consistency and comparability within a dataset is to scrutinize the data and codebooks. A question-by-question assessment -- called a crosswalk -- identifies where changes were made to questions, response options, and codes over time.

For our study that incorporated the SHED data, a Knology researcher conducted a crosswalk, where they reviewed and compared each year’s codebook and question response categories. Crosswalking the codebook and question response categories from 2013-2020 provided our team with an understanding of which questions could be incorporated into a longitudinal analysis.

Knology then used the SHED data to measure a handful of economic security outcomes for U.S. residents, and compared them to financial education investments at the state level. The economic security outcomes include subjective financial well-being, un-affordability of healthcare, retirement savings, and financial fragility. We tested whether and to what extent financial education in school and other programs had an impact on financial security.

Let’s Put It to Work

This crosswalk offers a breakdown of SHED survey and survey supplement codebooks from 2013 through 2021, mapping questions and question IDs year-over-year. The first appearance of a question and question ID is highlighted in a blue color, and changes made to that question in subsequent years are highlighted in a purple color. This crosswalk, presented as a downloadable .CSV workbook, can not only be used by researchers who wish to incorporate longitudinal SHED data into their research, but can also serve as a model for those who plan to use other public datasets with longitudinal data.

About the Project

From 2019 to 2020, Knology led A New History of Investment in Financial Education across the United States, a research initiative funded by the National Endowment for Financial Education® (NEFE) and supported by a range of experts. Knology built a robust database of historical spending on financial education across all 50 states, as well as outcomes of that spending. The team then used the database to study how investments in financial education through public schools and non-profits contributed to indicators of financial health for U.S. residents, such as retirement savings and financial fragility. Learn more about the project here. If you have questions about this project, contact us at info@knology.org.

The National Endowment for Financial Education is the independent, centralizing voice providing leadership, research, and collaboration to advance financial well-being. Find out more about NEFE at NEFE.org.

Photo credit: 13on on Unsplash.

Join the Conversation
What did you think of this? How did you use it? Is there something else we should be thinking of?
Support research that has a real world impact.