U-Behavior: A method to promote spaced and interleaved retrieval practice in challenging undergraduate microbiology and biomedical sciences courses

Published Feb 12, 2026 on Dryad. https://doi.org/10.5061/dryad.cnp5hqck8

Data files

Feb 12, 2026 version files 58.75 KB

Abstract

The underlying data for this study consist of anonymized student performance metrics, retrieval practice activity (RPA) usage logs (including timestamps of quiz attempts, scores, spacing intervals, and mixing patterns), UB-rubric scores, demographic descriptors (aggregated only), and exam/retention test scores from two undergraduate courses: Introduction to Biomedical Sciences (VMBS 100) and General Microbiology (MIP 300). These data were extracted from the university's learning management system (LMS) and course records. No raw individual-level data are shared due to privacy protections under institutional review board (IRB) guidelines and FERPA regulations; all analyses use de-identified, aggregated, or summarized forms.

Dataset DOI: 10.5061/dryad.cnp5hqck8

Description of the data and file structure

This dataset contains anonymized, aggregated, and processed student learning behavior and performance data collected to evaluate the effectiveness of the U-Behavior teaching and learning method—a pedagogy tool designed to promote evidence-based study strategies (spaced retrieval practice and interleaving) in undergraduate STEM courses. Data were gathered from two quasi-/randomized controlled experiments conducted at a large land-grant university in the Western United States.

Experiment 2 (randomized assignment) involved first-year students in Introduction to Biomedical Sciences (VMBS 100), a gateway course for biomedical sciences majors.

Experiment 3 (quasi-experimental, section-level assignment) involved junior-level students in General Microbiology (MIP 300), a foundational, content-intensive course required for microbiology majors and related STEM programs.

In both experiments, students used identical online retrieval practice activities (RPAs; low-stakes quizzes drawn from large question pools) accessible via the learning management system (LMS) throughout the semester. The control condition received standard low-stakes quizzing instructions, while the U-Behavior treatment condition added a tutorial on effective strategies, personalized RPA graphs visualizing spacing and mixing behaviors, periodic reflections on practice patterns, and rubric-based grading of final behavior graphs.

LMS timestamps captured quiz attempt details (dates, times, scores) to compute spacing scores (distribution of attempts over days), mixing scores (alternation between topics), total attempts, and UB-rubric levels (categorized practice quality). Course exam scores (midterm and final) and a delayed retention test (MIP 300 only, administered 4 weeks post-course) provided performance measures. All data were de-identified to comply with IRB and privacy regulations.

The experiments tested whether U-Behavior increases use of high-utility study behaviors compared to control, and whether these behavioral changes predict improved exam performance and long-term retention in challenging microbiology-related courses. Processed summaries, descriptive statistics, and regression outputs support the manuscript's findings on behavior change and learning outcomes

Files and variables

File: `1._MIP300_001_-_Dryad.xlsx`

Description: This file supports analyses in the paper's Experiment 2: comparing behavior change (Spacing, Mixing, UB-Level) between conditions, relating them to exam performance (Exam1–FinalExam) and overall outcomes (FinalScore, FinalGrade), and controlling for prior achievement (HS_GPA, CSU_GPA) and practice quantity (TotalAttempts). It also includes a retention test component indirectly through outcomes, though raw retention scores are not in this sheet.

Variables

Section

Course section identifier (string, e.g., "2020FA-MIP-300-002", "2020FA-MIP-300-003", "2020FA-MIP-300-232", "2020FA-MIP-300-233"). Indicates which section the student was enrolled in (used for quasi-experimental assignment in Experiment 3).

Consent

Indicates whether the student consented to data use for research (all rows = "Yes").

Condition

Experimental group assignment: "U-Behavior" (treatment: full intervention with tutorial, RPA graphs, reflections, and graded behavior) or "Control" (low-stakes quizzing only).

AttentionValue

attention/engagement/participation score or data quality flag based on LMS active (2–4; higher values indicate better attention or completeness).

Exam1

Score on Exam 1 (percentage or points; float, range includes 0–101; some 0s indicate missed exams).

Exam2

Score on Exam 2 (percentage or points; float).

Exam 3

Score on Exam 3 (percentage or points; float).

FinalExam

Score on the comprehensive final exam (percentage or points; float).

ExamsFinalScore

Weighted average or composite score from the four exams (Exam1 + Exam2 + Exam3 + FinalExam) (percentage; float).

FinalExamGrade

Letter grade assigned specifically for the final exam component (categorical: A, A-, B+, B, B-, C+, C, C-, D, F).

FinalScore

Overall final course score (percentage; float), the comprehensive weighted grade for the entire course.

FinalGrade

Overall letter grade for the course (categorical: A+, A, A-, B+, B, B-, C+, C, C-, D, F).

RetrievalPracticeActivities

Points or percentage earned for the retrieval practice activities (RPAs) component

Spacing

Spacing behavior score (%) — quantifies how well the student distributed RPA attempts across multiple days (0–100; higher = better distributed/spaced practice). Derived from attempt timestamps.

Mixing

Interleaving/mixing behavior score (%) — quantifies how well the student alternated between different RPAs/topics within study sessions (0–100; higher = better mixing).

UB-Level

U-Behavior rubric level (integer 1–4): overall categorization of practice quality based on Spacing and Mixing thresholds (1 = lowest/least effective; 4 = optimal/highly effective spaced and mixed practice).

Original-UB-Grade

Points awarded for the U-Behavior practice behavior component based on UB-Level and rubric (integer, e.g., 16 for level 1, 40 for level 2, 80 for level 4 in U-Behavior group; lower/blank in Control).

U-BehaviorFinalPoints

Final points assigned for the U-Behavior activity after rubric application (integer, e.g., 41, 65, 80; reflects adjusted or final graded value for behavior in U-Behavior group).

PracticeBehaviorCheckALL

Categorical or flag variable indicating overall practice behavior review status or compliance check

U-BehaviorFinalScore

Final percentage or score contribution from the U-Behavior component to the overall course grade

TotalAttempts

Total number of RPA quiz attempts across all available RPAs (integer; measures overall practice quantity).

HS_GPA

High school grade point average (float, scale 0–5.0 or 4.0+; prior academic performance covariate; some missing/blank).

CSU_GPA

Cumulative GPA at Colorado State University (CSU) at the time of the course (float; university-level prior performance covariate; used in regressions for MIP 300).

Additional notes on the file and variables:

The dataset is anonymized (no student IDs/names).

Rows represent individual students (n ≈ 165+ based on partial view; full sheet has more).

Many Control group rows have blanks/NaN for U-Behavior-specific variables (e.g., Spacing, Mixing, UB-Level, U-BehaviorFinalPoints, U-BehaviorFinalScore) since behavior was not graded or visualized the same way.

Units: Percentages (%) for exam/course scores (Exam1–FinalScore, Spacing, Mixing, RetrievalPracticeActivities); points for graded components (Original-UB-Grade, U-BehaviorFinalPoints); count for TotalAttempts; GPA on standard scale.

File: VMBS100_-_Dryad.xlsx

Description: This file supports analyses in the paper's Experiment 2: comparing behavior change (Spacing, Mixing, UB-Level) between conditions and relating them to performance (Midterm, FinalExam, FinalScore), while controlling for prior achievement (HS_GPA) and practice quantity (TotalAttempts).

Consent

Indicates whether the student consented to data use for research.

Condition

Experimental group assignment: "U-Behavior" (treatment: received the full U-Behavior intervention with tutorial, graphs, reflections, and graded behavior) or "Control" (low-stakes quizzing only).

Attention Value

Measure of student attention/engagement in LMS;

Midterm

Score on the midterm exam (percentage or points; float, range ~32.5–50).

FinalExam

Score on the final exam (percentage or points; float, range ~0–98).

FinalScore

Overall final course score (percentage; float, range ~36.4–98.18), a weighted combination of all assessments.

FinalGrade

Letter grade assigned for the course (categorical: A, B, C, F).

Retrieval Practice Activities

Points earned or percentage achieved for the retrieval practice activities (RPAs) component of the grade (float, 0–120 in U-Behavior group; NaN or blank in Control where not graded the same way).

Spacing

Spacing behavior score (%) — measures how well the student distributed RPA attempts across different days (0–100; higher = better spacing / more distributed practice). Calculated from attempt timestamps.

Mixing

Interleaving / mixing behavior score (%) — measures how well the student alternated between different RPAs/topics in study sessions (0–100; higher = better mixing).

BehaviorCheck

Categorical or flag variable for behavior quality check (integer, mostly 1 or 2); Indicates whether behavior met minimum thresholds or a review status.

UB-Level

U-Behavior rubric level (integer 1–4): overall categorization of practice quality based on Spacing and Mixing thresholds (1 = lowest / least effective; 4 = optimal/highly effective spaced & mixed practice).

Original-UB-Grade

Points awarded for the U-Behavior component based on UB-Level and rubric (integer, e.g., 24 for level 1, 60 for level 2, 120 for level 4 in U-Behavior group.

HS_GPA

High school grade point average (float, scale likely 0–5.0 or 4.0; used as a prior academic performance covariate; some missing/NaN).

TotalAttempts

Total number of RPA quiz attempts across all available RPAs (integer; overall practice quantity).

Additional notes on the file and variables:

The dataset is anonymized (no student IDs/names).

Rows represent individual students (n ≈ 125 based on visible data).

NaN/blank values appear in Control group for variables graded only in U-Behavior (e.g., Retrieval Practice Activities, detailed behavior scores).

Units: Percentages (%) for scores (Midterm, FinalExam, FinalScore, Spacing, Mixing); points for some grade components (Retrieval Practice Activities, Original-UB-Grade); count for attempts; GPA on standard scale.

This file supports analyses in the paper's Experiment 2: comparing behavior change (Spacing, Mixing, UB-Level) between conditions and relating them to performance (Midterm, FinalExam, FinalScore), while controlling for prior achievement (HS_GPA) and practice quantity (TotalAttempts).

Code/software

Microsoft Excel or another spreadsheet.

Human subjects data

I confirm that I received explicit consent from your participants to publish the de-identified data in the public domain. All individual identifiers were removed from these data files.