New self-efficacy scale measures nursing students’ medication calculation confidence

Medication errors remain one of the most persistent threats to patient safety in hospitals worldwide, and the earliest point of intervention is often the moment a nursing student first prepares a syringe, calculates a pediatric dose, or double-checks a vial label before administration. A newly published methodological study in BMC Nursing introduces a purpose-built instrument designed to measure exactly that critical moment in professional formation: the Medication Preparation and Calculation Self-Efficacy Scale, or MPCSES. Developed by a team of nursing researchers at Marmara University in Istanbul, Türkiye, the scale offers educators a validated way to gauge how confident, capable, and decision-ready students feel when facing the full sequence of medication management, from interpreting a prescription to drawing up the correct volume and administering it safely. The work arrives at a time when nursing programs across the globe are being pressed to demonstrate not just what their students know, but what they believe they can do under real clinical pressure, a distinction that decades of psychological research have shown to be decisive in predicting actual performance.

The theoretical foundation of the new instrument rests on Albert Bandura’s self-efficacy theory, the influential framework holding that a person’s belief in their own capability to execute the behaviors required for a specific performance is a stronger predictor of that performance than knowledge alone. In the medication context, this matters because dose calculation is not a purely mathematical exercise. A student may be able to solve a formula on paper yet freeze when confronted with a confusing concentration expression on a drug label, an unusual weight-based prescription for an infant, or the ambient stress of a busy ward. The researchers built their initial pool of items by combining a systematic review of the literature with the constructs described in Bandura’s theory, aiming to capture the multidimensional character of medication self-efficacy: the cognitive demands of calculation, the psychomotor and technical demands of preparation, and the clinical judgment required to notice when something is wrong. The first draft of the scale contained 55 items, each designed to reflect a concrete situation a nursing student is likely to encounter.

Before any statistical testing, the team turned to a panel of 15 experts in nursing fundamentals and related fields, employing the Delphi technique, a structured consensus process in which experts rate the relevance and clarity of each item across successive rounds. The content validity of the instrument was then quantified using content validity indexes, a standard metric in instrument development. At the item level, the index values ranged from 0.87 to 1.00, comfortably above the commonly cited threshold of 0.78 for retaining items, while the scale-level average index reached 0.97, an unusually high figure that indicates near-unanimous expert agreement that the items genuinely represent the construct of interest. This step matters more than it may appear from the outside: a scale that measures something adjacent to, but not exactly, the intended construct can mislead educators into targeting the wrong deficits, wasting curricular time and potentially leaving genuine gaps in safety training unaddressed.

With content validity established, the researchers moved to the heart of psychometric evaluation: construct validity, the question of whether the scale actually measures the latent structure it claims to measure. A total of 539 nursing students from two universities participated, deliberately divided into independent samples so that exploratory factor analysis and confirmatory factor analysis could be performed on separate datasets, a design considered best practice because it prevents a model from being built and validated on the same data. Exploratory factor analysis, which searches for hidden patterns of correlation among items without imposing a prior structure, revealed a clean three-factor solution accounting for 63.72 percent of the total variance, a strong figure in social science measurement. After the exploratory stage and a round of confirmatory-model-based refinement, the original 55-item scale was pared down to 29 items distributed across the three factors, trimming redundancy while preserving the conceptual coverage that the expert panel had endorsed.

The confirmatory stage subjected the refined three-factor model to rigorous statistical scrutiny, and the fit indices tell a reassuring story for anyone versed in structural equation modeling. The ratio of chi-square to degrees of freedom came in at 2.735, within the conventional acceptance range of below 3, while the Comparative Fit Index and the Tucker-Lewis Index reached 0.918 and 0.911 respectively, both above the 0.90 benchmark for acceptable fit. The Goodness-of-Fit Index, the Adjusted Goodness-of-Fit Index, the Normed Fit Index, and the Incremental Fit Index all landed in the 0.904 to 0.913 band, and the Standardized Root Mean Square Residual of 0.055 fell well under the 0.08 threshold. The Root Mean Square Error of Approximation, arguably the most demanding of the fit statistics, was 0.077, just inside the commonly used 0.08 cutoff. Taken together, these numbers indicate that the hypothesized three-factor structure is a plausible and statistically defensible representation of how medication-related self-efficacy is organized in students’ minds.

The three factors retained in the final model allow educators to assess distinct but related dimensions of the medication process, rather than collapsing everything into a single undifferentiated confidence score. Convergent validity, the degree to which items designed to measure the same factor actually correlate strongly with one another, was supported by composite reliability values between 0.937 and 0.941 and average variance extracted values between 0.646 and 0.697, both exceeding the classic 0.50 criterion proposed by Fornell and Larcker. Discriminant validity, the evidence that the three factors are genuinely separate rather than a single factor masquerading as three, was confirmed by the Fornell-Larcker criterion, which requires each factor’s average variance extracted to exceed the squared correlations between that factor and the others. In practical terms, a nursing program using the scale will be able to distinguish a student who calculates doses flawlessly but hesitates in the physical preparation of medications from one who handles the technical steps confidently yet struggles with the arithmetic, and tailor remediation accordingly.

Reliability, the consistency of the measurement across items and across time, proved equally robust. Cronbach’s alpha, the most widely used index of internal consistency, ranged from 0.918 to 0.949 across the three factors and reached 0.965 for the overall scale, values that indicate exceptional coherence among the items. Test-retest stability was examined using the intraclass correlation coefficient, which came in at 0.89 with a 95 percent confidence interval spanning 0.79 to 0.95, demonstrating that the instrument yields stable readings when the underlying construct has not changed. Beyond internal statistics, the researchers also tested criterion-related validity by comparing scale scores with performance on a five-item medication calculation test. The correlation was moderate, positive, and statistically significant, with r equal to 0.412 and a p value below 0.001. The moderate rather than strong magnitude is itself informative: it confirms that self-efficacy is related to, but not identical with, actual calculation skill, precisely the conceptual separation that Bandura’s theory predicts and that justifies measuring both constructs independently.

The clinical significance of a validated self-efficacy instrument in this domain is difficult to overstate. Medication administration is consistently ranked among the highest-risk activities performed by nurses, and calculation errors involving weight-based pediatric dosing, intravenous infusion rates, and unit conversions are implicated in a meaningful share of preventable harm. Yet most existing educational assessments target knowledge or arithmetic ability, leaving the affective and confidence-based dimensions unmeasured. A scale that reliably captures self-efficacy gives educators a diagnostic lens onto students’ readiness before they enter clinical placements, and it can support longitudinal research tracking whether specific teaching interventions, such as simulation-based practice, high-fidelity scenarios, or structured calculation laboratories, produce durable gains in confidence that translate into safer behavior. The instrument also opens the door to cross-institutional and eventually cross-cultural comparisons, since a common validated metric is a prerequisite for meaningful benchmarking between nursing schools.

The study was conducted with explicit ethical safeguards, receiving approval from the Marmara University Faculty of Health Sciences Non-Invasive Clinical Research Ethics Committee, with institutional permissions secured at both participating universities. Participation was voluntary, written informed consent was obtained from all students, and the research was carried out in accordance with the Declaration of Helsinki. The authors, Aynur Kaynar Şimşek, Ayşe Nur Karagöz, Gülten Karahan Okuroğlu, and Şule Ecevit Alpar, all affiliated with the Department of Fundamentals of Nursing at Marmara University, report no competing interests and received no external funding for the work. The article was published as an open-access resource in BMC Nursing on 8 September 2026, meaning that educators and researchers worldwide can examine, replicate, and adapt the full methodological details without subscription barriers, a transparency measure that strengthens confidence in the reported psychometric findings.

As nursing education continues its global shift toward competency-based frameworks and measurable outcomes, instruments like the MPCSES represent the quiet infrastructure that makes such reforms possible. A 29-item scale that can be administered in minutes, that discriminates between the knowledge, skill, and judgment components of medication management, and that has survived exploratory and confirmatory factor analysis, expert consensus validation, and rigorous reliability testing, is a genuine addition to the assessment toolkit. For the 539 students whose responses shaped it, and for the countless nursing students who will complete it in the years ahead, the scale’s deeper promise is systemic: safer ward rounds, fewer silent near-misses at the medication trolley, and an educational system capable of detecting and repairing confidence gaps before they become clinical errors.