EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 3.0

Statistical Reporting Inconsistencies in Mediation and Moderation Analyses of Aging Perceptions and Depression Among Older Adults: A Comment on Song and Cui [Letter]

Rashid Menhas

Psychology Research and Behavior Management · 2026-11 · Volume 19 · 1-2

DOI: 10.2147/prbm.s669659 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Psychology Research and Behavior Management, 2026-11. Verfasst von Rashid Menhas. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 3.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

I read with interest the longitudinal study by Song and Cui on negative self-perceptions of aging (SPA), cognitive function, socioeconomic status (SES), and depression among older Chinese adults. 1 Several discrepancies between the reported statistical tables and their interpretation warrant attention because they affect the identification of mediation and moderation results.My observations are based exclusively on published article without re-analysis of the underlying data.First, the mediation estimates received conflicting pathway labels.In Table 4, the pathway "Negative SPA → Depression" is reported as b = 0.0322 (95% CI, 0.0286-0.0363),whereas "Negative SPA → Cognitive function → Depression" is reported as b = 0.0004 (95% CI, 0.0002-0.0009).In contrast, the Results section describes b = 0.0322 as evidence of the mediating effect of cognitive function and b = 0.0004 as the total effect.Therefore, the narrative conflicts with the pathways identified in Table 4. Furthermore, the single-arrow pathway is not explicitly defined as either a direct or total effect, and the effect scale of the bootstrap estimates is unclear in a study with a binary depression outcome.The authors should specify what each reported estimate represents and, where estimated, identify the corresponding direct, indirect, and total effects, confidence intervals, and effect sizes.The nonsignificant cognitive function-depression association reported in Model 3 does not, by itself, refute a bootstrapped indirect effect; the concern is the contradictory identification of the reported effects.Second, the reported moderation term differs from the stated hypothesis.Hypothesis 3 and the Analytical Strategy specify an interaction between negative SPA and SES in predicting depression.Yet, the final column of Table 3 labels the interaction as "Cognitive function × SES" (OR = 0.998), while the Results interpret this estimate as the negative SPA × SES interaction and report p = 0.981.These interaction terms test different propositions; therefore, the reported null result cannot be interpreted unambiguously until the fitted term is identified.The authors should establish which interaction was estimated and align the model table, accompanying Results, and conclusion with that specification.Table 3 also labels both of its final columns "Model 3", although the Results call the final specification Model 4; correcting this identification would further support transparent interpretation.Third, the sample and bootstrap-cluster counts were not reconciled.The Methods report 2583 older adults and 8,559 observations from the 2011, 2014, and 2018 waves, whereas the Results describe 5000 bootstrap replications based on 2853 clusters, explicitly characterized as individual-level clusters.The authors should specify whether the 2,853 clusters refer to the full available dataset, the final analytic sample, or a differently defined resampling population; explain their relationship to the reported 2,583 participants; and document how the person-wave observations and multiple-imputation procedures relate to the bootstrap sample.This is an unresolved reporting discrepancy, not evidence that the resampling procedure was computed incorrectly.

Zitieren

APA

Menhas, R. (2026). Statistical Reporting Inconsistencies in Mediation and Moderation Analyses of Aging Perceptions and Depression Among Older Adults: A Comment on Song and Cui [Letter]. Psychology Research and Behavior Management, Volume 19, 1-2. https://doi.org/10.2147/prbm.s669659

Vancouver

Menhas R. Statistical Reporting Inconsistencies in Mediation and Moderation Analyses of Aging Perceptions and Depression Among Older Adults: A Comment on Song and Cui [Letter]. Psychology Research and Behavior Management. 2026;Volume 19:1-2. doi:10.2147/prbm.s669659

BibTeX
@article{Menhas2026,
  author = {Menhas, Rashid},
  title = {Statistical Reporting Inconsistencies in Mediation and Moderation Analyses of Aging Perceptions and Depression Among Older Adults: A Comment on Song and Cui [Letter]},
  journal = {Psychology Research and Behavior Management},
  year = {2026},
  volume = {Volume 19},
  pages = {1--2},
  doi = {10.2147/prbm.s669659}
}
Kostenlos starten →