April 22, 2026 | By Krystal Thomas and Arif Rachmatullah
In our earlier blog post on person-centered analyses, we described how this approach identifies patterns of traits within individuals to reveal distinct profiles of students who approach learning in different ways.
Focusing on the “average student” can often hide what matters most. While all students aim to succeed in their courses, they encounter different barriers shaped by prior preparation, institutional structures, and life circumstances. Thus, students differ in how they manage time, seek help, and respond to challenges—and those differences shape learning outcomes. To design learning environments that genuinely meet students where they are, we need a clearer picture of the varied ways they navigate their coursework. In this post, we offer an example from the Postsecondary Teaching with Technology Collaborative (Postsec Collab) to illustrate the affordances of person-centered analyses and how it can inform decisions for instructors, researchers, and edtech developers.
The Study at a Glance
The Postsec Collab is a five-year research center focused on strengthening self-directed learning in online introductory STEM courses.
- 1,516 undergraduate students in online STEM courses (e.g., biology, chemistry, math, computer science)
- Four broad-access institutions–defined as colleges and universities that accept more than 75% of their applicants including community colleges across the U.S.
- 53% of students were enrolled in courses that introduced self-directed learning strategies through reflective prompts, videos, or structured peer interactions
Using survey data collected at the start of the term, we identified patterns in how students approached learning. See our report for more details of the study.
Three Distinct Student SDL Profiles
We identified three distinct student self-directed learning profiles. These profiles are not labels or deterministic categories; rather, they describe common patterns in the combination of skills, belief, and strategies—such as reviewing notes, creating diagrams, chunking studying)—that students report at the start of their courses. Together, the profiles highlight that students arrive with different configurations of strengths and needs, not simply more or less motivation or ability.
Profile 1: Adaptive self directed learners
These students report high confidence, strong learning strategy use, and active monitoring of their learning. They tend to navigate online learning environments effectively and benefit from autonomy. At the beginning of the term only 20% of students fit into this profile.
Key strengths:
- High self-efficacy
- Effective time management
- Strong metacognition
Profile 2: Reflective but underconfident learners
Students in this group are highly reflective, such that they plan, monitor, and evaluate their learning. However, they often struggle with confidence and belonging. They know what they should doubt may not consistently follow through. Thirty percent of the study sample fit in to this profile at the beginning of the term.
Key strengths and needs:
- Strong metacognition
- Lower self-efficacy
- Lower belonging
Profile 3: Overwhelmed learners
These students report low confidence, more fixed mindset beliefs, and limited use of active learning strategies. They may feel uncertain about how to succeed and navigate in online STEM courses. Half of the study sample (50%) fit in to this profile at the beginning of the term.
Key needs:
- Motivational support
- Strategic scaffolding
- Opportunities for successful early wins
Why Profiles Matter
Profiles help identify where targeted, equity-conscious support can make a difference. These profiles not only describe how students learn—they also reveal how different groups of students are distributed across patterns of strengths and needs. These differences matter for designing supports that improve outcomes overall while also addressing disparities in how students experience and succeed in their courses.
For example, we found that:
- Compared to their white peers, Black/African American students were more likely to be in the adaptive profile and less likely to be in the overwhelmed profile
- Compared to students not receiving Pell Grants, Pell Grant recipients were more likely to fall into a reflective-but-underconfident profile
- Young students were more likely to be in underconfident or overwhelmed profiles relative to their older peers
These patterns suggest that students are not uniformly “at risk” or “prepared”; rather, they face different combinations of assets and barriers.
Importantly, profiles changed over time. Surveys given near the end of each course showed:
- Students were more likely to belong to the adaptive profile by the end of the course, with larger gains for students in courses that introduced the self-directed learning strategies
- Membership in the overwhelmed profile decreased
- Membership in the reflective-but-underconfident profile remained stable, suggesting a need for stronger confidence-building supports
Together, these findings show that short, scalable instructional strategies can strengthen self-directed learning and course success in online STEM courses for all students, while highlighting that different groups may benefit from differentiated support.
How Practitioners Can Use Profiles
Instructors can design layered supports that meet students where they are:
- Provide leadership and peer mentoring opportunities for adaptive learners.
- Normalize help-seeking and pair reflection with confidence-building for underconfident learners.
- Offer structured time-management support and early outreach for overwhelmed learners.
Researchers can use profiles to move beyond averages and ask:
- Who shifts into more adaptive learning patterns—and under what conditions?
- Who remains at risk over time?
EdTech and Learning Management System designers can build tools that:
- Support follow-through for self-directed learning skill development, not just awareness of the strategies.
- Offer adaptive nudges for planning, reflection, and help-seeking that are customized by student profile.
- Detect emerging learning needs before disengagement occurs, especially in online courses.
Looking Ahead
There is no single path through online STEM learning. Students combine skills and strategies in different ways—and those patterns can change. By looking beyond averages and paying attention to student learning profiles, we can design more equitable, responsive, and supportive online learning environments.
Stay tuned for upcoming work extending this approach to elementary STEM and computer science learning.