University students’ fitness falls sharply across three years: linked annual tests reveal where, when and who is most affected

Table of Contents

1. Key Highlights
2. Introduction
3. How the study tracked fitness: data, grouping and analytic approach
4. What changed, by how much and when: an overview of results
5. Why the decline is concentrated in speed, endurance and strength: physiological and behavioral links
6. Discipline-specific patterns: why did Science students show the largest drop?
7. Who is most vulnerable: stage and sex differences that matter for targeting
8. Practical implications: turning annual tests into continuous prevention
9. What the data cannot show: limitations and required next steps
10. Translating findings into actionable campus practice: an implementation checklist
11. Why this matters: health and education stakes
12. FAQ

Key Highlights

• Linked annual fitness records for 4,406 students show total physical fitness scores fell by 6–8 points from 2023 to 2025 across Engineering, Science, and humanities/social sciences/management (HSSM) students, with the largest adjusted drop in Science.
• The deterioration concentrated in 2024–2025 and was driven mainly by declines in speed, endurance, lower-limb explosive power and muscular strength/endurance; vital capacity and flexibility improved modestly.
• Declines were larger among students who were sophomores at baseline and larger in magnitude for female students; pass rates and good-or-excellent grades dropped substantially, signaling both broad and targeted management needs.

Introduction

Three consecutive years of standardized campus fitness testing yield a clear message: university life, as the academic calendar advances, is associated with meaningful declines in measurable physical fitness. Using individual-level linked records from a Chinese university, researchers tracked 4,406 freshmen and sophomores through annual tests in 2023, 2024 and 2025. The dataset—12,795 student-year observations—captures both overall scores and the raw measurements behind them: sprint times, endurance run times, standing long jump, pull-ups or sit-ups, vital capacity and flexibility. These linked records allow the study to move beyond cross-sectional snapshots and population-level trend descriptions to reveal within-student trajectories, discipline-specific patterns and which fitness domains erode first.

The findings have practical consequences for campus health managers and academic leaders. A loss of 6–8 points in the university’s total fitness score is large enough to move many students from “pass” to “fail,” and from “good/excellent” into lower categories. The decline is not identical across majors or student groups. It concentrates in capacities that require repeated aerobic and strength stimuli, and it appears to accelerate as students progress through their studies. This article presents the study’s methods and results, explores plausible mechanisms, evaluates robustness, and outlines targeted policy and program options that universities can deploy to arrest and reverse the trend.

How the study tracked fitness: data, grouping and analytic approach

The analysis draws on routine annual fitness testing organized under the National Student Physical Fitness Standard. Inclusion required freshman or sophomore status in 2023, an identifiable student ID, recorded total score and discipline-group information. The final cohort included 4,406 students: 3,284 Engineering, 342 Science, and 780 HSSM students. Most (4,071; 92.4%) completed all three annual tests; the remainder had two or one records, creating the longitudinal dataset used for mixed-effects modeling.

Why linked records matter Repeated cross-sectional surveys show population trends, but they cannot distinguish whether an observed decline arises from cohort differences or from within-person change. Linking tests at the student level isolates true within-student trajectories. It allows the model to account for each student’s baseline fitness level and individual rate of change, while also modeling clustering at the major level to reflect shared curriculum or training environments.

Discipline groups and background characteristics Students were grouped by baseline major into Engineering, Science (mathematics, applied mathematics and geophysics in this sample), and HSSM (humanities, social sciences and management). Group composition differed sharply by sex: Engineering and Science were male-dominated (77.4% and 71.3% male), while HSSM had more females (59.7%). Baseline mean total scores were similar across groups (≈68–71 points), and baseline pass rates were high (≈92–95%). The discipline grouping is a background indicator of the educational environment—course structure, laboratory and practical demands, time use—and remains fixed for each student through follow-up.

Testing battery and scoring The national battery integrates body morphology and motor-function tests. Key items and their weights in the total score:

• BMI (15%), vital capacity (15%)
• 50-m sprint (20%), endurance run (1,000-m males / 800-m females; 20%)
• Standing long jump (10%), sit-and-reach flexibility (10%)
• Strength: pull-ups for males or one-minute sit-ups for females (10%) Bonus points can add up to 20 points for outstanding performance in some items. Raw measurements were retained and analyzed as exploratory outcomes to pinpoint which domains changed.

Modeling strategy Total physical fitness score was modeled with linear mixed-effects models including fixed effects for year, discipline group, baseline grade and sex, plus interactions of year with these variables. Random effects: a major-level intercept, student-specific intercepts and student-specific time slopes; within-student residuals followed an AR(1) structure. This configuration allowed the model to estimate average trajectories while preserving individual heterogeneity in baseline fitness and rates of change.

Complementary analyses considered binary grade outcomes (pass; good-or-excellent) with conditional logistic regression stratified by student ID, and item-level raw measurements were analyzed to identify the physical domains behind score changes. A broad set of sensitivity analyses—complete-case restriction, inverse probability weighting for missing data, removal of the BMI component, leave-one-major-out, outlier handling and stress tests—assessed robustness.

What changed, by how much and when: an overview of results

Magnitude and timing of change Across all discipline groups the total physical fitness score fell substantially from 2023 to 2025. Model-adjusted mean scores:

• Engineering: 69.07 → 63.12 (change −5.96 points; 95% CI −6.26 to −5.66)
• Science: 70.53 → 62.66 (change −7.87 points; 95% CI −8.77 to −6.97)
• HSSM: 69.05 → 62.40 (change −6.64 points; 95% CI −7.27 to −6.02)

The decline was not linear across years. From 2023 to 2024 scores fell modestly (0.94–2.57 points), while from 2024 to 2025 the drop was larger (5.02–5.56 points). That timing suggests an accelerating loss of fitness after the first follow-up year.

Grade-based outcomes Pass rates fell sharply in all groups between 2023 and 2025:

• Engineering: 91.78% → 68.62% (−23.15 percentage points)
• Science: 93.86% → 66.77% (−27.09 pp)
• HSSM: 94.87% → 72.32% (−22.55 pp)

The proportion of students rated “good or excellent” also declined, with the largest relative losses in Science and HSSM:

• Engineering: 10.26% → 3.88% (−6.38 pp)
• Science: 13.74% → 1.85% (−11.90 pp)
• HSSM: 15.51% → 3.26% (−12.26 pp)

Sex and baseline-grade differences

• Baseline sophomores experienced larger 3-year declines than baseline freshmen. Sophomores fell roughly 9–11 points; freshmen fell about 2.7–4.7 points, depending on discipline.
• Female students showed larger adjusted declines than male students in absolute terms (e.g., females −7.35 to −9.25 points across groups versus males −5.38 to −7.29 points), but the three-way interaction (year × discipline group × sex) did not reach statistical significance, indicating the discipline-specific patterns were not clearly sex-specific.

Which domains moved and which did not Raw item analyses revealed targeted deterioration:

• Declines: 50-m sprint times increased (worse), endurance run times increased, standing long jump distance decreased, and strength measures (one-minute sit-ups in females; pull-ups in males) decreased.
• Stable or improved: Vital capacity and sit-and-reach flexibility improved modestly.
• BMI increased across groups; interpreted cautiously because BMI conflates fat and lean mass.

Robustness of findings Results proved robust to multiple sensitivity analyses. The year × discipline-group interaction remained significant when the sample was restricted to complete three-year cases, after removing BMI from the total score, after inverse probability weighting for missing data, and in leave-one-major-out tests. Even conservative stress tests assuming worse 2025 outcomes for missing students did not change the principal conclusion: fitness declined in all groups, with Science showing the largest drop.

Why the decline is concentrated in speed, endurance and strength: physiological and behavioral links

The pattern of item-level change aligns with principles of exercise physiology and typical university behavior changes.

Physiological precursors Speed, explosive jump performance, endurance running and muscular strength/endurance are all capacities that require repeated, sufficiently intense training stimuli to maintain. Sprinting and jumping depend on neuromuscular power and lower-limb strength mechanics; strength training transfers to improved sprint and jump performance. Endurance depends on regular aerobic stimulus to preserve cardiovascular conditioning. When stimulus frequency or intensity falls, these domains deteriorate faster than traits like flexibility or certain measures of lung function, which may respond differently to general activity patterns.

Behavioral shifts during university University progression often entails:

• Decreased structured physical education or group sport requirements.
• Increased academic load, laboratory work, internships and exam preparation that elevate sedentary time.
• Greater need for self-directed exercise, which many students fail to sustain or to deliver at sufficient intensity.
• Possible reduced motivation for maximal test performance after repeated annual assessments.

These behavioral shifts reduce the specific training that preserves sprint speed, explosive power, endurance and muscular strength. Vital capacity and sit-and-reach can improve with low-to-moderate general activity or targeted practices like breathing exercises or stretching routines and therefore show different trajectories.

Real-world parallels

• National and international surveillance repeatedly finds insufficient physical activity among adolescents and young adults, and university students worldwide report falling activity levels during academic progression.
• Intervention studies cited by the study—blended learning models for basketball education and active classroom movement breaks—show that structured curricular or extracurricular programs can counter sedentary displacement and improve fitness outcomes.

Discipline-specific patterns: why did Science students show the largest drop?

The study shows a statistically significant difference in trajectories by discipline group: Science students suffered the largest adjusted decline in total score (−7.87 points), followed by HSSM (−6.64) and Engineering (−5.96). The difference does not stem from baseline disparities in scores; rather, it reflects distinct longitudinal trajectories.

Possible explanatory mechanisms

• Curriculum rhythm: Engineering programs in the sample had a higher proportion of practical-teaching credits across academic years (about 31–35% depending on academic year) than Science and HSSM. Practical teaching like laboratory sessions, engineering training or internships may involve more movement related to the learning environment, or at least less continuous sedentary attendance, than theory-heavy majors.
• Time demands and study patterns: Science majors represented here (mathematics, geophysics) often include intensive theory and individual study sessions that can increase sedentary time, especially in later academic stages.
• Extracurricular opportunity and culture: Differences in club participation, team sports and informal peer exercise culture can vary by faculty and affect students’ likelihood of maintaining training habits.

Caveats Practical-teaching credits do not directly measure physical activity intensity or duration. Laboratory or internship tasks can be physically light or heavy; the credit counts are contextual indicators, not direct behavioral measures. The Science group composition in this university (mathematics, geophysics) shapes the result; results may differ in institutions where biology, chemistry or field-based sciences comprise the majority of science students. Multi-institution replication would clarify external validity.

Who is most vulnerable: stage and sex differences that matter for targeting

Baseline-grade vulnerability Students who were sophomores at the 2023 baseline experienced larger declines than baseline freshmen—roughly three to four times larger. This likely reflects stage-related vulnerability: as students move from early to later undergraduate years, structured PE obligations often fall away, internships and professional training intensify, and self-directed study demands rise. Universities should not confine fitness promotion to orientation or first-year programs alone.

Sex differences: magnitude not pattern Female students showed larger mean declines than male students, but sex did not significantly modify the discipline-specific patterns. Two caveats explain why simple male–female comparisons must be cautious:

• Test items differ by sex (e.g., endurance distances and strength tests), and total scores use sex-specific standards.
• Social, motivational and behavioral drivers differ by sex (e.g., types of physical activity preferred, time use, perceptions of exercise), which can influence both participation and test outcomes.

Programmatic implication: stratify by risk, not stereotype Discipline group, baseline grade and sex help identify risk strata for targeted prevention, but none should be used alone to label individual students. The most effective approach combines group-level risk signals with individual monitoring.

Practical implications: turning annual tests into continuous prevention

Annual testing offers a surveillance snapshot. The study demonstrates the additional value of linking tests at the student level. That link enables early warning and stage-specific responses.

Three practical shifts universities can make

1. Turn annual tests into a continuous monitoring system
   • Maintain and link raw item measurements, grade classifications and total scores at the student level.
   • Define trajectories per student (stable, mildly declining, severely declining) and trigger differential responses: automated guidance for mild decline, targeted coaching or referrals for persistent decline.
   • Use the raw-item profile to match interventions (e.g., prescribe strength training for students losing pull-ups; prescribe interval or endurance runs for those losing cardio capacity).
2. Preserve and adapt structured activity opportunities across the undergraduate period
   • Keep core PE or movement opportunities available beyond the first year. Short, frequent activity sessions embedded into schedules—e.g., 10–15 minute high-frequency workouts—can sustain intensity without large time commitments.
   • Provide flexible practical offerings that fit professional-training demands (e.g., early-morning or evening sessions, modular micro-courses that students can fit between labs and internships).
3. Apply evidence-backed interventions and evaluate them
   • Blended learning models for PE (classroom + online, plus practical skill sessions) and e-health tools (digital coaching, activity tracking, social support platforms) have evidence of improving student physical activity and fitness.
   • Classroom movement breaks and physically active learning reduce sedentary time and improve focus.
   • Evaluate interventions with linked fitness data and randomized or quasi-experimental designs where feasible.

Examples from practice

• A cluster randomized trial in basketball education found blended learning could improve fitness outcomes; such curricular innovation can be extended to other sports or general fitness modules.
• Systematic reviews and meta-analyses support the effectiveness of digital interventions in promoting activity among college students; combining digital coaching with campus facilities can increase reach and adherence.
• Short “movement break” protocols during long lectures have reduced sedentary behavior and fatigue in university settings and may preserve fitness when scaled.

Operational considerations

• Data governance and privacy: linked fitness records must be anonymized and handled under appropriate ethical oversight; de-identified datasets can inform program planning while protecting students.
• Motivation and behavioral support: test results should be framed constructively to avoid stigmatization and to encourage participation in programs.
• Integration with health services: students with worrying declines or large BMI increases should be offered holistic assessments (activity, diet, sleep, mental health) rather than BMI-driven judgments alone.

What the data cannot show: limitations and required next steps

Limitations to bear in mind

• Single-university sample: the discipline composition and curriculum structure are specific to the institution studied. The Science group here consisted mainly of mathematics and geophysics; other universities may show different discipline effects.
• Missing behavioral measures: the dataset lacks direct measures of sport history, weekly activity frequency, sedentary time, screen use, sleep, diet or academic stress. These behaviors likely mediate the observed declines but require direct measurement to confirm causation.
• Test motivation and technique: item-level results can be influenced by how seriously students take annual testing, by their movement technique and by testing conditions, despite standardized procedures.
• Observational design: associations between discipline group or grade and fitness trajectories cannot be interpreted as causal effects of curriculum or course structure.

What research should follow

• Multi-university replication with diverse discipline mixes to test external validity.
• Studies that combine linked fitness records with repeated behavioral and environmental measures (wearable activity monitoring, surveys of time use, course schedules) to identify causal pathways.
• Trials of stage-specific and domain-specific interventions (e.g., micro-strength programs, interval running protocols, blended PE curricula) evaluated with linked fitness outcomes.
• Qualitative work exploring student motivations, barriers to self-directed exercise and the role of campus culture.

Translating findings into actionable campus practice: an implementation checklist

For university administrators and health promotion teams seeking to act, a pragmatic checklist:

• Retain and link raw item measurements, total scores and grades across years for every student, under robust data governance.
• Define clear thresholds for automated alerts: for example, a drop of ≥4 points in total score in one year or a downward trend across two years triggers outreach.
• Offer tiered interventions:
  • Universal: brief activity modules, movement breaks in lectures, campus campaigns promoting daily activity.
  • Targeted: short-term coaching, group strength sessions, interval-running clinics for identified students.
  • Intensive: individualized training plans, nutrition and sleep counseling, clinical referral if appropriate.
• Protect time for sustained activity: schedule protected activity windows or integrate active learning to reduce prolonged sedentary time.
• Monitor and evaluate: use subsequent linked tests to measure intervention impact; collect process metrics (participation, adherence) and student feedback.

Why this matters: health and education stakes

Physical fitness in young adulthood is not a cosmetic metric. Cardiorespiratory fitness and muscular fitness predict later-life cardiovascular and metabolic risk. Declines during university can lock in longer-term inactivity patterns, with implications for population health and workforce readiness. Moreover, physical fitness supports cognitive function, stress resilience and academic performance—outcomes universities have a stake in protecting. The findings make clear that annual testing alone is insufficient. Linked fitness data can be an early-warning tool that points to where stage-specific, discipline-informed and domain-targeted interventions are needed.

FAQ

Q: How large is the fitness decline and is it practically important? A: Adjusted declines from 2023 to 2025 were about −6.0 points in Engineering, −7.9 points in Science and −6.6 points in HSSM. These drops are large enough to shift many students between grade categories (e.g., from pass to fail, or from good/excellent down to lower categories). The decline therefore has practical consequences for student classification and indicates meaningful loss of capacities that matter for health.

Q: Which abilities deteriorated most? A: The decline concentrated in speed (50-m sprint), endurance (800-m/1,000-m runs), lower-limb explosive power (standing long jump), and muscular strength/endurance (pull-ups or one-minute sit-ups). Vital capacity and sit-and-reach flexibility improved modestly, indicating a non-uniform pattern of change.

Q: Did decline differ by major or sex? A: Yes. Science students in this sample experienced the largest adjusted decline. Pass-rate decline was widespread across groups, while loss of good-or-excellent status was most pronounced in Science and HSSM. Female students showed larger absolute declines than male students, but discipline-specific trajectory patterns did not differ significantly by sex.

Q: When did most deterioration occur? A: Most of the observed decline occurred between the second and third year of follow-up (2024 to 2025), suggesting an accelerating loss as students progress through their studies.

Q: Could changes in BMI explain the findings? A: BMI increased across groups, but removing the BMI item score from the total score did not change the primary discipline-specific findings. BMI conflates fat and lean mass and is not a direct measure of fitness; item-level declines in sprint, endurance and strength provide clearer evidence that reduced training stimulus explains much of the score drop.

Q: Are the results generalizable to other universities? A: Findings are robust within the studied institution and stable across multiple sensitivity checks, but generalizability requires replication in other settings with different major mixes, curriculum plans and campus cultures. The Science group’s composition here (mathematics and geophysics) shapes the discipline-specific result.

Q: What should universities do now? A: Shift from seeing annual tests as discrete evaluations to using linked annual records as continuous monitoring. Maintain structured activity opportunities beyond first year, introduce short high-frequency activity sessions, deploy digital interventions and blended PE where appropriate, and target interventions to students with early signs of decline—particularly those in later undergraduate stages.

Q: Can students reverse these declines? A: Yes. Evidence from intervention studies shows that targeted strength training, interval and endurance programs, and consistent activity habits can restore or improve sprint, endurance and strength capacities. Universities can help by providing structured, time-efficient programs and by integrating activity into academic timetables.

Q: What data should universities collect alongside fitness tests? A: To design effective interventions, collect repeated measures of physical activity (e.g., accelerometers or activity logs), sedentary time, sleep, diet, mental health indicators and participation in physical education or extracurricular sport. These data help identify causal pathways and inform tailored programs.

Q: What are the next research priorities? A: Replication across multiple institutions, integration of behavioral and environmental measures with linked fitness data, randomized evaluations of stage-specific and domain-specific interventions, and qualitative studies on student barriers and motivations for sustained exercise.