Impact of video-based learning in business statistics: a longitudinal study

The study tests the concept that engagement with learning materials outside the classroom promotes more positive student–subject interaction. With the expansion of learning environments (accelerated by Covid-19) and student expectations for technology-driven support, the paper examines video-based learning (VBL) as an approachable medium to empower first-year business undergraduates to engage with statistics. Given employability demands and TEF-driven emphasis on numeracy, the research aims to evaluate VBL’s effectiveness in improving achievement in a compulsory Business Statistics module. Admissions policies in UK Business Schools vary on maths requirements, leading to heterogeneous preparedness; thus, the study explores relationships between media-enhanced learning, student experience, and achievement in business statistics. The research adopts a longitudinal design (11 years, 2006/07–2016/17) to assess both immediate and sustained impacts of VBL on attainment.

Mathematics for non-specialists in HE: Many business students have limited post-compulsory maths experience, prompting universities to include maths/statistics in curricula. While support centres can be effective, engagement is typically low, allowing competence gaps to persist; negative perceptions and maths anxiety further reduce engagement. Technology in HE is widespread but inconsistently applied, impeding sector-wide impact evaluation. Video-based learning (VBL): Prior literature identifies VBL as widely used and potentially effective but notes mixed results due to underinvestment and limited tutor capability to produce suitable materials. Publicly available videos often lack alignment to HE learning outcomes and quality standards (OfS). Effective VBL typically requires bespoke, educator-created videos aligned to curriculum, with features such as clear structure, concise explanations, and appropriate duration. A doctoral review (Lewis, 2019) highlighted gaps concerning effective VBL application to maths/statistics in business school settings. Research questions: (1) To what extent has VBL enhanced achievement in a compulsory Business Statistics module? (2) How was VBL developed over time within the module? (3) How does VBL raise attainment/achievement in statistics?

Design and setting: Longitudinal comparative study at a modern UK university Business School (Southeast England) across 11 academic years (2006/07–2016/17). The compulsory first-year Business Statistics module enrolls ~350–400 students/year at two campus locations (A and B) with identical curriculum, assessments, and teaching materials; same module leader ensured consistency. No pre-/co-requisites; admissions required 96 UCAS points plus GCSE English and Maths, leading to wide variation in maths preparedness. Assessments: Two summative elements, each 50%: (1) Computer-based exam (COM) using Microsoft Excel; (2) Multiple-choice exam (MCQ) using statistical tables and calculators. COM requires selecting appropriate methods, applying Excel formulae/graphs, and interpreting results for business decision-making. Only first-sit results were analysed. Baseline (pre-VBL) data: Historical datasets 2006/07–2012/13 established achievement profiles prior to VBL. Intervention: In 2013/14, a VBL pilot at Location A (experimental) was introduced; Location B served as control. Twenty bespoke instructional videos aligned exactly to workshops and mock exam materials, plus five mock-exam walkthrough videos, were hosted on the VLE. Video design principles: (1) step-by-step instructions; (2) explicit linkage to class content; (3) concise explanations/narration; (4) logical structure to sustain engagement; (5) duration ≤15 minutes. From 2014/15 onward, VBL was integrated into curricula at both locations and promoted. Teaching structure: Weekly 12-week module with 1-hour lecture, 1-hour seminar (calculator-based problem solving), 1-hour computer workshop (Excel-based tasks), and expectation of ~114 hours of independent study supported via VLE. Data collection: Anonymous achievement data (COM and MCQ marks), VLE engagement metrics (visit frequency, video views), and student perceptions via online questionnaire and semi-structured interviews in 2015/16 and 2016/17. Survey sample sizes exceeded thresholds for 95% confidence with 10% margin of error. Statistical analysis: Descriptive statistics; tests for normality; independent-samples t-tests for normal data and Mann–Whitney U-tests for non-normal comparisons between locations within each year; effect size eta squared; correlations (Spearman’s rho) to examine relationships among resource usage (VLE, video, books, eBooks, game-based learning); graphical visualisations of grades and VLE usage.

Pre-VBL performance gap and failure rates: - 2006/07–2012/13: COM underperformance relative to MCQ averaged 20 ± 11% (max 31%). In 2012/13, 39% failed COM vs 12% failed MCQ. COM failure rates across years ranged 23–47%; MCQ 9–19%. Pilot and subsequent years (Table 1 summaries, COM means ± SD; A vs B): - 2012/13 (pre-VBL both): A 55±26, B 53±23; T=0.707, p=0.48, η²=0.002 (no significant difference). - 2013/14 (VBL at A only): A 63±25, B 51±26; U=8269, p=0.001, η²=0.03. At Location A, mean rose by 8% and median by 11% vs prior year; distributional tests confirmed significant improvement after VBL (U=6875, p=0.021). - 2014/15 (VBL both): A 74±23, B 57±26; U=8185.5, p=0.001, η²=0.11. At Location B, mean/median rose by ~6% vs 2013/14; pre/post VBL significant (U=23219.5, p=0.023). - 2015/16: A 77±24, B 67±26; U=7201, p=0.001, η²=0.35; increase driven by more First-class marks (≥70%) at B, raising cohort mean. - 2016/17: A 67±29, B 69±26; U=11077.5, p=0.79, η²=0.002 (no significant difference). Drop at A linked to fewer VLE visits among lower-achieving students. Grades and engagement (Figs. 3–4): - 2013/14: ~47% of Location A achieved grade A in COM (≈62% higher than Location B), with ~38% higher VLE visits at A; COM grade F proportion at B was ~50% higher than A. - 2014/15 onward: VLE visit frequency increased at both locations after VBL integration; A generally had higher visit counts than B. Among lower achievers (grades C–F), VLE engagement fell markedly between 2015/16 and 2016/17 (mean decrease ≈28%); 26% of this bracket failed COM, depressing A’s cohort mean in 2016/17. - Overall, introduction and integration of VBL corresponded to sustained improvements in COM marks and higher proportions of “good degrees” (2:1 and First) in COM. Video usage patterns (Fig. 7): - Mock exam videos were most popular; “Mock exam Q1” averaged ~2000 views/year for ~350-student cohorts, indicating repeated revision viewing and preferential engagement with assessment-aligned videos. Student preferences and correlations (Figs. 5–6; Table 2): - Survey response rates: 2015/16 N=335, n=117 (35%); 2016/17 N=352, n=80 (23%). - ~97% used the VLE; ~90% used videos; lower engagement with books/eBooks compared to VLE/VBL. Qualitative feedback highlighted value of access/ease, step-by-step pacing, short duration (≤15 min), integration into lessons, and tutor voice. - Spearman correlations show VLE–Video positive and significant (e.g., 2016/17 ρ≈0.59**); negative correlations with Book (e.g., 2016/17 VLE–Book ρ≈−0.49**, eBook–VLE ρ≈−0.29**), indicating reliance on tutor-provided digital resources over traditional textbooks. Overall: Data indicate VBL contributed significantly to improved skills-based assessment performance (COM), heightened engagement via VLE, and increased higher-grade attainment, particularly when videos were integrated and promoted within the curriculum.

Findings address the research questions by demonstrating that VBL enhanced first-year business students’ achievements in statistics, most notably in the skills-based COM exam. The 2013/14 pilot showed immediate, significant gains at the experimental location versus control, and subsequent integration at both locations yielded sustained improvements and higher proportions of top grades. Patterns of VLE usage and positive VLE–Video correlations suggest that structured, assessment-aligned videos increase engagement and translate into better performance. The design characteristics that supported engagement included short, logically structured, step-by-step videos narrated by the tutor and directly tied to workshops and mock assessments. The temporary attenuation of gains in 2016/17 at one location was associated with reduced VLE engagement among lower-achieving students, reinforcing the importance of active promotion and continued integration of VBL to maintain benefits. The results underscore that VBL is most effective when tightly integrated with classroom instruction, clearly linked to assessments, and supported by the tutor’s presence, thereby reducing anxiety and fostering learner autonomy and ownership.

The longitudinal analysis shows that integrating tutor-created, curriculum-aligned VBL into a first-year Business Statistics module improves student engagement and achievement, particularly in computer-based, skills-focused assessments. Effective VBL requires more than hosting videos on a VLE; it necessitates deliberate integration with teaching sessions, explicit links to assessments, concise and navigable design (≤15 minutes), and active promotion. VBL enhances accessibility, repeatability, and learner autonomy, helping to overcome maths anxiety and providing consistent support across cohorts. Future research should strengthen causal inference by linking individual background (e.g., prior maths exposure) with granular engagement analytics (identity-based viewing logs, repeat counts, time-on-task) and assessment outcomes, enabling modelling of dose–response (e.g., views needed to reach pass/grade thresholds) and investigating which video features most influence performance.

- Student background confounding: The study could not conclusively control for prior mathematics/statistics exposure (e.g., A-levels), so improved achievement may partially reflect pre-existing differences. - Identity/engagement linkage: Anonymity prevented identifying individuals, tracking repeat video views, total watch time, or linking specific engagement behaviours to assessment outcomes. Consequently, dose–response relationships (e.g., number of views required to reach pass/grade bands) could not be established. - Survey linkage: Anonymous questionnaires could not be linked to individual achievement profiles, limiting analysis of how preferences and behaviours relate to performance. - Generalisability: Conducted within a single university’s business school context; although multi-year and multi-site, external validity to other institutions or disciplines may be limited without replication.