The notice concerns the article “A Longitudinal Investigation of the Repetitive Non-Suicidal Self-Injury Questionnaire (R-NSSI-Q) in South Korean Adolescents: Addressing Test-Retest Reliability and Predictive Validity,” originally published on 12 February 2026. The correction was authored by Hwa Jung Lee, Dong Hun Lee, Ye Jin Kim, Sihyeong Kim and Rassamee Chotipanvithayakul. The paper is part of a growing effort to improve psychological assessment for adolescents, particularly when researchers need to distinguish a stable pattern of behavior from a temporary response to stress. Because self-injury is a highly sensitive subject, accurate measurement is not a cosmetic concern: numerical errors can influence how readers interpret the strength of relationships between symptoms, risk factors and later outcomes.

The R-NSSI-Q is designed to assess repetitive non-suicidal self-injury, a term used for deliberate self-inflicted injury without the intention of dying. The distinction from suicidal behavior is clinically important, although the two can coexist and both require serious attention. A questionnaire cannot diagnose an individual by itself. Rather, it provides a structured way to collect information that can support research, screening and professional assessment. In a longitudinal investigation, participants are assessed over time instead of at only one moment. This allows researchers to ask two different technical questions: whether the instrument produces reasonably consistent results when the underlying characteristic has not substantially changed, and whether earlier scores are statistically associated with later measurements or outcomes.

The first question is addressed through test-retest reliability. In practical terms, researchers administer the same measure on separate occasions and examine the degree to which the scores correspond. A high association suggests that the questionnaire is capturing a relatively stable construct rather than generating largely random results. It does not mean that every adolescent will give identical answers, nor does it prove that the instrument is valid for every purpose. People’s experiences can change, and responses may be affected by memory, mood, context or willingness to disclose. Test-retest reliability is therefore one component of measurement quality. It indicates temporal consistency, while validity asks whether the instrument is actually measuring the psychological feature it claims to measure.

The second question, predictive validity, concerns whether scores obtained at one point are related to relevant measurements later on. Researchers commonly evaluate this relationship using correlation and regression analyses. A correlation coefficient summarizes the direction and strength of an association between two variables, typically ranging from −1 to +1. A coefficient near zero indicates little linear association, while values closer to either extreme indicate stronger linear relationships. A regression coefficient, by contrast, estimates how an outcome changes in relation to a predictor within a specified statistical model, often while accounting for other variables. Neither statistic automatically demonstrates causation. A questionnaire score that predicts a later outcome may reflect shared influences, unmeasured factors or the broader context in which both measurements were collected.

That statistical framework explains why the missing punctuation mattered. Decimal notation is part of the identity of a reported coefficient. A correlation of .28 is a modest positive association; a value of 28 would be impossible under the usual correlation scale. Similarly, a regression coefficient of .07 communicates a small estimated change, whereas 7 suggests an entirely different scale and interpretation. The correction states that several decimal points were inadvertently omitted during typesetting in Tables 8 and 9 and in the corresponding prose. The publisher specifically notes that all affected correlation and regression coefficients were below 1 in magnitude and should have included a leading decimal point. The error was introduced in presentation, not in the calculations or analysis.

The authors and publisher emphasize that the mistake does not affect the study’s results, interpretation or conclusions, and that the original article has been updated. This distinction is central to understanding what a correction means in scientific publishing. Journals issue corrections when the published record contains an error that should be fixed, even if the error does not invalidate the research. A typesetting problem can be relatively minor operationally but important for transparency, because readers may download a version of an article, extract values into a review or use a table in later work. Updating the article and documenting the change helps ensure that future readers see the intended numbers and can trace why the correction was issued.

The paper’s institutional affiliations also underline the cross-disciplinary nature of the work. Four authors are affiliated with the Traumatic Stress Center in the Department of Education at Sungkyunkwan University in Seoul, Republic of Korea, while Rassamee Chotipanvithayakul is affiliated with the Department of Epidemiology at Prince of Songkla University’s Faculty of Medicine in Hat Yai, Thailand. The collaboration brings together school and mental-health research with epidemiological expertise. That combination is valuable for longitudinal measurement studies, which must consider not only the wording and reliability of questionnaire items but also study design, statistical uncertainty, population characteristics and the limits of applying findings from one group to another.

The correction also highlights a broader challenge in adolescent mental-health research: the need to balance numerical precision with human context. Statistical coefficients can summarize patterns across a sample, but they cannot capture the full reasons an adolescent may experience distress or engage in self-injury. Nor can a reliable questionnaire replace a careful clinical conversation, safeguarding procedures or access to appropriate support. The value of a validated measure lies in making research observations more systematic and comparable, allowing scientists to examine whether patterns persist over time and whether interventions reach young people who need them. For that purpose, even a leading zero is consequential, because trustworthy evidence depends on both sound analysis and faithful communication of its results.