A surprising result from university classrooms
Universities are spending increasing amounts of time deciding which digital teaching tools deserve investment. Artificial intelligence, narrated presentations, online courses and electronic assessment platforms all promise to make teaching more efficient.
But a new study from Nigerian federal universities suggests that the strongest apparent relationship with business education delivery may come from a familiar assessment tool rather than the most technologically sophisticated option.
Published in Frontiers in Education on 9 October 2026, the research examined responses from 306 participants in South-South Nigeria. It compared four types of educational software and their associations with the delivery of business education content.
The strongest individual relationship involved Google Forms. The study also found positive relationships for cognitive artificial intelligence, narrated PowerPoint lectures and Skillshare when each was examined separately.
However, the results changed when the researchers analysed all four technologies together. Only Google Forms and Skillshare retained statistically significant unique associations in the joint model.
That distinction is important: the findings describe relationships in survey data, not experimental proof that one application causes better teaching.
What the researchers investigated
The research team, led by Edet E. Okon, focused on four software technologies used in the context of business education.
The first was cognitive artificial intelligence, a broad category of AI systems that can assist with reasoning, information processing and adaptive learning tasks.
The second was Voice on PowerPoint, abbreviated VOPPT, referring to presentation slides accompanied by recorded or narrated explanations.
The third was Skillshare, an online learning platform associated with practical, project-oriented instruction.
The fourth was Google Forms, which can be used to distribute questionnaires, administer quizzes, collect responses and organise assessment information.
These tools serve different functions, so treating them as interchangeable examples of educational technology would obscure meaningful differences.
The researchers asked whether each technology was associated with the reported delivery of business education content and whether the associations remained after considering the technologies jointly.
The analysis was informed by Activity Theory, which views tools as part of a wider system involving people, institutional rules, teaching objectives and available resources.
How the study was designed
The investigators used a correlational research design rather than an intervention or randomized trial.
The study population consisted of 306 respondents, and the researchers reported studying the entire identified population.
Data were collected through a structured questionnaire and analysed using simple and multiple linear regression.
Simple regression examined each software technology individually, while multiple regression considered all four in the same statistical model.
The distinction matters because educational institutions that use one digital tool may also use others. Apparent relationships in separate analyses can overlap.
The researchers assessed statistical significance at the five-percent level and reported correlation coefficients, coefficients of determination and regression results.
The outcome concerned business education content delivery as measured in the study. It was not a direct experimental measurement of student examination performance or graduate employment.
Because respondents supplied questionnaire information, common reporting patterns and perceptions may have influenced the observed associations.
Google Forms showed the strongest individual relationship
Of the four technologies, Google Forms produced the strongest association with business education content delivery in the individual analyses.
The researchers reported a correlation coefficient of 0.804 and an R-squared value of 0.646.
An R-squared value of 0.646 means the simple regression model accounted for approximately 64.6 percent of the observed variation in the study’s content-delivery measure.
It does not mean Google Forms improved teaching by 64.6 percent or that the software caused 64.6 percent of learning outcomes.
Google Forms is commonly used for quizzes, feedback, attendance information and organising assessment responses.
Those functions may align closely with the administrative and instructional activities measured by the questionnaire.
However, the study cannot tell whether effective institutions adopt Google Forms, whether the software helps their work or whether both reflect another underlying factor.
That uncertainty is a reason to interpret the strong correlation carefully rather than as a definitive purchasing recommendation.
What the other three tools showed
Voice on PowerPoint had an individual correlation coefficient of 0.626, with an R-squared value of 0.392.
Skillshare produced a correlation coefficient of 0.578 and an R-squared value of 0.335.
Cognitive AI showed a correlation coefficient of 0.574 and an R-squared value of 0.330.
All three individual relationships were reported as statistically significant, with p-values below 0.001.
These results indicate that respondents reporting greater application of these technologies also tended to report stronger business education content delivery.
They do not establish that narrated presentations, Skillshare or AI independently raised academic achievement.
The individual R-squared values should not be added together. They are calculated in separate models and may reflect overlapping explanatory information.
The more revealing comparison is therefore the model that considered all four technologies simultaneously.
Why the combined model changes the interpretation
When the four technologies were entered together, the model accounted for 65.6 percent of the variation in the measured content-delivery outcome.
The researchers reported a multiple correlation of 0.810 and an adjusted R-squared value of 0.652.
Google Forms retained a strong statistically significant unique association, with a standardized coefficient of 0.808 and p-value below 0.001.
Skillshare also retained a significant unique association, with a coefficient of 0.185 and p-value of 0.004.
By contrast, cognitive AI had a coefficient of minus 0.107 and a p-value of 0.164, while narrated PowerPoint had a coefficient of minus 0.062 and a p-value of 0.439.
Neither of those two coefficients was statistically significant in the combined model.
This does not show that AI or narrated slides harm teaching. Their negative coefficients were not statistically distinguishable from zero in that specification.
Instead, the findings suggest that much of their individual association with content delivery overlapped with information captured by the other technologies.
The result illustrates why an impressive single-variable correlation can be misleading when several related factors operate together.
What the statistics do and do not mean
The study’s headline figure of 65.6 percent describes explanatory fit within a regression model built from the collected data.
It is not a measure of the percentage of students who learned successfully or a prediction that introducing four tools would improve instruction by that amount.
Similarly, a significant p-value indicates that the observed statistical association would be unlikely under a particular null hypothesis and modelling assumptions.
It does not establish practical importance, economic value or a causal mechanism.
The regression coefficients also depend on the definitions and measurement scales used in the questionnaire.
Strong associations can arise when participants who rate one aspect of their educational environment positively also rate related aspects positively.
Without experimental assignment, direct classroom observation and independent learning assessments, the direction of the relationship remains uncertain.
The safest conclusion is that the survey identified a pattern worth investigating through more rigorous educational evaluations.
Why assessment software might stand out
Business education often combines conceptual teaching with practical exercises, assignments and feedback.
A simple response-collection tool can be closely connected to these daily teaching processes.
Lecturers can use electronic forms to gather answers, identify misunderstandings and organise routine assessment tasks.
Students may encounter such tools repeatedly across a course, making their presence particularly visible in perceptions of instructional delivery.
However, the study did not experimentally compare classes using Google Forms with otherwise identical classes using paper-based assessments.
Nor did it isolate which particular features of the platform were most influential.
The findings should therefore prompt questions about assessment design, feedback quality and workflow rather than an assumption that one branded product is uniquely effective.
Comparable systems could plausibly perform similar functions, but that possibility was not tested in the research.
The role of AI deserves a closer look
Artificial intelligence attracts considerable attention because it can help generate examples, explain concepts and support administrative tasks.
Yet the Nigerian survey did not find a statistically significant unique association for cognitive AI after the other three technologies were included.
One interpretation is that AI’s apparent contribution overlapped with broader digital teaching practices.
Another possibility is that the survey’s definition of cognitive AI did not capture differences between specific applications or their quality of use.
The study cannot determine which explanation is correct.
Importantly, a nonsignificant coefficient is not proof that AI has no educational value.
It means the available data did not provide sufficient evidence of an independent relationship in the combined model.
Evaluating AI tools properly would require clearer measures of how they are used, comparisons of teaching practices and independent evidence of student outcomes.
Lessons for university decision-makers
The findings suggest that institutions should evaluate digital teaching investments by their instructional functions rather than their novelty.
Assessment tools, multimedia presentations, online learning resources and AI systems solve different problems.
A university might need better feedback processes before it needs more sophisticated automated content generation.
Alternatively, a department struggling with accessibility or large class sizes might have very different priorities.
Technology selection should be guided by learning objectives, lecturer skills, connectivity and student access.
Training is particularly important because software cannot improve teaching if staff lack time or support to use it effectively.
Institutions should also assess privacy, accessibility, cost and ongoing maintenance before adopting a platform.
The research supports thoughtful integration of technology, but it does not establish a universal ranking of tools for every university.
Why the findings matter beyond Nigeria
The study addresses federal universities in South-South Nigeria, where digital infrastructure and institutional resources can shape teaching practice.
Other countries face related questions about how to invest limited education budgets.
South African universities, for example, also need to balance innovation with reliable access, teaching quality and student support.
However, the Nigerian findings should not be assumed to apply directly to South African institutions or other education systems.
Student populations, teaching practices, available infrastructure and software adoption may differ substantially.
The broader transferable lesson is methodological: compare technologies jointly, distinguish correlations from effects and measure the educational outcomes that matter.
Those principles are relevant wherever institutions are deciding how to use digital tools responsibly.
Limitations and the next research questions
The study was correlational, making causal claims inappropriate.
Its sample was confined to 306 respondents in a specific Nigerian higher education setting.
The questionnaire-based outcome captured reported content delivery rather than independently verified changes in learning performance.
Regression analysis can adjust for included variables, but it cannot eliminate all confounding or measurement bias.
The joint results may also be sensitive to overlap among technology-use measures and the precise specification of the model.
Future research could randomly assign teaching approaches, track student assessment outcomes and compare costs alongside educational benefits.
Researchers could examine whether the effectiveness of a tool depends on class size, lecturer training, subject content or student access.
Studies across multiple countries would also help determine whether the observed relationships generalize beyond the original setting.
These steps would move the evidence from descriptive associations toward stronger guidance for educational policy.
A practical message beneath the numbers
The researchers found that all four technologies were positively associated with reported business education content delivery when considered separately.
But in the joint analysis, only Google Forms and Skillshare retained statistically significant unique relationships.
The combined model explained 65.6 percent of the measured variation, a substantial statistical association that must not be mistaken for a causal effect.
Rather than showing that the newest technology is always the best, the study highlights the importance of what a tool actually helps lecturers do.
Better evidence is still needed to determine which digital interventions improve student learning and under what conditions.
For universities, the most useful question may be not which software sounds most advanced, but which teaching problem it is designed to solve.
Source Information
Study: Application of software technologies and delivery of business education content in federal universities.
Authors: Edet E. Okon, Odey John Odo, Patience U. Ushie, Patrick E. Oche, Francis A. Effah and colleagues.
Journal: Frontiers in Education, Digital Learning Innovations.
Published: 9 October 2026.
Methods: Correlational questionnaire study of 306 respondents in federal universities in South-South Nigeria; simple and multiple regression.
