Children’s success in school is bolstered by their persistence in the face of difficult learning tasks and their ability to overcome the various academic setbacks they may face. Developing these qualities can help students navigate academic challenges, supporting their long-term educational success. But what do we really mean by persistence and resilience? What do these qualities look like in classrooms and learning environments? How can we measure whether students are persisting, or how resilient they truly are in the face of setbacks?

As part of INVITE, a multi-institutional collaborative AI National Institute funded by the National Science Foundation, we are exploring technology-enabled approaches to support K-12 students in developing the persistence, resilience, and collaboration skills necessary for academic success, especially in the context of computer science courses.

Recently, our team undertook an important first step toward this goal by trying to understand how persistence and resilience are defined, and how they can be measured. To do this, we reviewed over 70 publications that used and provided validity evidence for measures of persistence or resilience and examined how each construct was defined as well as the characteristics and technical quality of each measure. Having clear operational definitions and appropriately reliable and valid surveys and indicators of these attributes is necessary for understanding how learning environments or interventions support, or fail to support, students’ persistence and resilience.

There are various ways to define persistence and resilience. Our review of educational and psychological studies of these two constructs revealed a lack of consensus on their definitions. Sometimes, persistence was viewed as a component of resilience, or vice versa, and some authors treated them interchangeably.

After considering the range of published definitions, we centered on the following definitions for persistence and resilience, using the most important facets of each construct.

• Persistence involves sustained effort toward completion of a goal-directed task despite challenges or difficulties.
• Resilience involves a process of using adaptive coping behaviors and strategies to bounce back or recover from adversities and achieve successful outcomes.

These definitions focus on persistence within tasks and resilience within academic learning contexts, rather than persistence across grade levels (for example, persistence to graduate or resilience in the face of more acute adverse childhood experiences).

We observed substantial diversity in approaches to measuring persistence and resilience. Most of the measures reported in prior research were self-report surveys. These instruments vary in length and depth, with some covering a single dimension (e.g., persistence) and others capturing multiple dimensions (e.g., persistence as one of several subscales within a larger survey of motivational constructs like self-efficacy, mastery orientation, and adaptive behaviors). These measures allow students to report on their typical behaviors as evidence of their tendency to persist or be resilient. This could include reporting how likely they are to finish tasks they begin or how likely they are to use coping behaviors or get derailed by setbacks.

Behavioral measures can also be derived from students’ interactions within learning environments. Several behavioral indicators of persistence have been used in prior studies, including measures related to effort like the number of actions taken, tasks or questions completed, and time spent, especially on difficult problems. Some studies have even investigated measures of what might be considered unproductive persistence or “wheel spinning”, where students spend time and effort without making progress toward task completion. When these states are detected in learning environments, educators can intervene at the desired moment to help students continue their learning journey in more productive ways.

While we observed many self-report measures for resilience, we only observed one example of a behavioral indicator of resilience available in prior research with K-12 students. This indicator was based on students’ performance on three easy problems after seeing four difficult problems, which does not fully capture our definition of resilience. This lack of appropriate behavioral measures is especially concerning given that we consider resilience as a process, suggesting that it might most validly be observed through adaptive coping behaviors demonstrated over time. We advocate for researchers to continue exploring methods for evaluating evidence of resilience from students’ long-term interactions within learning and assessment systems to better understand how resilience processes can be detected and supported during students’ educational journeys.

We hope this review will serve as a resource for other researchers interested in understanding the role of student characteristics in learning. We also aim to build on these findings in our work within the ETS Research Institute, examining ways to measure persistence and resilience to better consider how these factors affect students’ engagement in and performance on meaningful assessment tasks, and how data on these constructs can be used to support dynamic personalization within assessment contexts.

For further details on the definitions and measures reviewed, access our full paper (available open access) from Frontiers in Education.