Finnish Learning Questionnaire Passes Its Italian Test in Healthcare Students

A questionnaire developed in Finland to probe how university students experience teaching and learning has now been carefully translated, culturally adapted, and statistically tested for Italian healthcare students, according to a new study published in BMC Medical Education. The instrument, known as HowULearn 2, originated at the University of Helsinki and is designed to capture students’ perceptions of the teaching-learning environment and their own approaches to studying. The Italian team, led by Shaniko Kaleci of the University of Modena and Reggio Emilia, set out to determine whether this research-based tool could function reliably in a different language and academic culture, a question that matters enormously for educators who want to compare student experience across borders or use the feedback to improve their own courses.

The stakes are higher than they might first appear. Healthcare students face a distinctive cluster of academic, clinical, and psychosocial stressors, from demanding coursework to the emotional weight of clinical placements, and these pressures can erode both well-being and academic performance. Instruments like HowULearn 2 are meant to give educators and researchers a structured window into how students perceive their learning conditions, with the ultimate goal of supporting self-regulated learning and improving teaching. But a questionnaire that works in Helsinki cannot simply be run through a translation program and deployed in Modena. Words shift meaning, academic conventions differ, and items that seem clear in one culture can become ambiguous in another. That is why formal adaptation and validation studies are the unsung workhorses of educational research.

The Italian team approached the task in stages, each designed to answer a specific psychometric question. First came face validity, which asks whether the items look and read as clear and relevant to the people who actually fill them in. Forty students assessed the translated questionnaire, and the results were encouraging: 52 of the 61 items, or 85.2 percent, met the predefined clarity criterion. Items that fell short were flagged and revised, a process that turns the raw translation into something that reads naturally to Italian undergraduates rather than as a stiff foreign import.

Next, the researchers convened an eight-member expert panel to judge content validity, meaning whether the items adequately cover the construct the questionnaire is supposed to measure. The panel’s verdict was strong: the scale-level content validity index computed as an average, abbreviated S-CVI/Ave, reached 0.91, a value conventionally interpreted as excellent. In practical terms, the experts agreed that the Italian version preserved the conceptual breadth of the original Finnish instrument, covering the experiences of teaching and learning that the Helsinki team intended to capture.

The heart of the study was structural validity, the question of whether the statistical relationships among the 61 items reveal a coherent underlying architecture. The researchers administered the questionnaire to undergraduate healthcare students, inviting 351 and receiving data from 345. Data quality was remarkable: 343 participants, or 99.42 percent, had six or fewer missing item responses, and 294 complete cases were available for the factor analysis. To explore the structure, the team used exploratory factor analysis with principal-axis factoring and oblique Promax rotation, a combination that allows the extracted factors to correlate with one another, which is sensible when the dimensions of a learning environment are expected to be related rather than independent.

Before extracting any factors, the team checked whether the data were suitable for this kind of analysis at all. The Kaiser-Meyer-Olkin measure of sampling adequacy came in at 0.923, far above the usual 0.60 threshold and indicating that the items share enough common variance to justify factoring. Bartlett’s test of sphericity was overwhelmingly significant, with a chi-square of 11,533.725 on 1,830 degrees of freedom and a p-value below 0.001, confirming that the correlation matrix was not a statistical artifact. To decide how many factors to retain, the researchers ran Horn’s parallel analysis with 1,000 random replications using the 95th-percentile criterion, a modern and widely recommended alternative to the older eigenvalue-greater-than-one rule. This analysis supported a seven-factor solution.

Here the study reveals an interesting tension between statistics and theory. Although the parallel analysis pointed to seven factors, the researchers retained a five-factor solution, describing it as a theory-driven exploratory solution consistent with the intended dimensions of the questionnaire. In psychometric practice, purely statistical criteria can overextract factors that reflect item-level quirks rather than meaningful constructs, so researchers often weigh the empirical evidence against the conceptual design of the instrument. The Italian team’s choice signals that the five intended dimensions of HowULearn 2 remained interpretable and coherent in the Italian data, even as they acknowledge that the seven-factor statistical signal deserves further scrutiny in future work.

Reliability, the consistency of the measurements, was assessed in two complementary ways. Internal consistency, which measures how well items within each subscale hang together, was evaluated with Cronbach’s alpha. The coefficients ranged from 0.785 to 0.930 across the subscales, spanning the acceptable-to-excellent range, and the overall alpha for the full questionnaire reached 0.952, indicating that the instrument as a whole measures a strongly unified construct. Test-retest reliability, which asks whether the questionnaire gives stable results over time, was examined in 63 participants who completed the instrument twice, four weeks apart. Intraclass correlation coefficients using the ICC(3,1) model ranged from 0.607 to 0.838, a spread the authors characterize as moderate to good. For a questionnaire probing subjective experiences of teaching and learning, some fluctuation over a month is expected, and these values suggest the Italian version is stable enough for research and feedback purposes.

The authors are appropriately measured in their conclusions. They report that the Italian HowULearn 2 showed satisfactory face and content validity, high sampling adequacy, and a theoretically interpretable five-factor exploratory solution, with acceptable-to-excellent internal consistency and moderate-to-good test-retest reliability. At the same time, they call for further independent and multicentre studies, including confirmatory factor analysis, to assess the stability and generalisability of the measurement structure. This is an important caveat: exploratory factor analysis, by its nature, discovers structure in a specific sample, and only confirmatory methods applied to fresh samples in different institutions can establish that the structure holds. The single-site design, drawn from one Italian university, also means the findings await replication elsewhere before the Italian version can be considered fully portable across the country’s diverse tertiary education landscape.

What makes this study quietly significant is what it enables next. With a validated Italian HowULearn 2, Italian educators in medicine, nursing, dentistry, and other healthcare programs can now gather systematic, research-grounded feedback on how their students experience teaching and learning, and use that feedback to foster self-regulated learning and refine their courses. Cross-national comparisons with Finnish and other administrations of the instrument become possible in principle, opening the door to studying how cultural and institutional contexts shape student experience. And for the students themselves, whose academic and clinical lives are shaped by the environments their universities create, a reliable tool for listening to their perspective is a small but meaningful step toward education that adapts to the learners rather than the other way around. The study, conducted in accordance with institutional ethical standards and the Helsinki Declaration, received no specific external funding, and all participating students gave informed consent, a fittingly careful foundation for an instrument devoted to understanding how students learn best.

Subject of Research: Cross-cultural adaptation and psychometric validation of the HowULearn 2 student learning questionnaire for Italian tertiary healthcare education

Article Title: Adaptation and validation of HowULearn 2 student questionnaire for use in Italian tertiary education

Article References: Kaleci, S., Costi, S., Bellini, P., Generali, L., Lotti, A., Sena, P., Chester, J., Amaducci, G., & Mecugni, D. (2026). Adaptation and validation of HowULearn 2 student questionnaire for use in Italian tertiary education. BMC Medical Education. https://doi.org/10.1186/s12909-026-10521-2

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10521-2

Keywords: HowULearn 2, questionnaire validation, psychometrics, medical education, healthcare students, structural validity, exploratory factor analysis, Cronbach's alpha, test-retest reliability, content validity, Italian higher education, self-regulated learning