AMAyumi Mizuno

Ayumi Mizuno

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ayumi Mizuno has a total of 295 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Declining environmental predictability erodes reproduction in a tropical seabird

    by Santiago Ortega, Cristina Rodríguez, Ayumi Mizuno, Laura Arroyo, Hugh Drummond, Shinichi Nakagawa

  • Open Access

    Statistical fragility in the meta-analytic evidence base of animal cognition: a meta-meta-analytic reappraisal

    by Ayumi Mizuno, Pietro Pollo, Malgorzata Lagisz, Lauren Guillette, Masayo Soma, Shinichi Nakagawa

    How reliable is the meta-analytic evidence base in animal cognition? We conducted a pre-registered second-order meta-analysis of 28 published meta-analytical papers, comprising 48 eligible models and 5740 effect-size observations from primary studies. We assessed statistical power, Type M magnitude error, and Type S sign error at both primary-study and meta-analysis levels. Because these quantities depend on the assumed underlying effect, we compared the results based on uncorrected pooled effects, a regression-based bias correction, and a sensitivity analysis using a bias-robust fixed-effect model incorporating a within-study variance-covariance matrix (FE+VCV). At the primary-study level, estimated power was low under all three empirically estimated effect proxies: 17.4%, 9.0%, and 13.4%, respectively. Type M and Type S errors were also sensitive to the assumed effect and correction method. Power increased when larger externally specified effects were assumed, confirming that these quantities are conditional on the assumed effect. Meta-analysis-level results were generally more favourable but also sensitive to analytical choices. The evidence base was concentrated in experimental rodent studies of learning and memory, limiting field-wide generalisation. These findings support improving study design and reporting within the represented literature while expanding meta-analytic coverage across taxa and cognitive domains.

  • Open Access

    Music exposure reduces anxiety- and depression-like behavior in rodents: a systematic review and multilevel meta-analysis

    by Santiago Ortega, Anna Lenz, Erick Lundgren, Ayumi Mizuno, Sergio Poo Hernandez, Shinichi Nakagawa, Malgorzata Lagisz

    Music exposure may alter affect-related behavior in rodents, but findings vary among studies and little is known about effects on behavioral variability. We conducted a preregistered systematic review and multilevel meta-analysis of 20 experimental studies (298 effect sizes) of music exposure in laboratory rodents. We quantified changes in average anxiety- and depression-like behavior using the log response ratio (lnRR) and changes in between-animal variability using the log variability ratio (lnVR), while accounting for non-independence among effects. We also assessed reporting, methodological quality, and risk of bias using CRIME-Q. Music exposure was associated with about a 20% reduction in anxiety- and depression-like behavior relative to controls, but no overall change in behavioral variability. Average responses differed among behavioral assays, while exploratory analyses of variability suggested context-dependent effects: anxiety-like outcomes tended toward greater variability and depression-like outcomes toward lower variability, and behavioral assay and music meta-genre explained comparatively larger shares of heterogeneity in lnVR. Reporting of sample-size justification, allocation, blinding, attrition, and acoustic delivery was often incomplete in the primary studies. Evidence was limited to 20 studies and predominantly to young adult rodents and short-term exposures, constraining inference about variability and generalization beyond the laboratory contexts studied.

  • Open Access

    Mate choice copying in non-human animals: an update of two meta-analyses

    by Eduardo S. A. Santos, Aleksandra Milenović, Aneta Arct, Anna Lenz, Ayumi Mizuno, Cassidy Schneider, Christine Sosiak, Erick Lundgren, Hao Qin, Iwo Gross, et al.

    Mate choice copying, the use of social information from others' mating decisions, is a taxonomically widespread form of social learning that can shape sexual selection. Two meta-analyses concluded that social information has a moderate positive effect on mate choice; however, both are based on literature searches that are now at least 7 years old. We asked whether the overall average effect they reported remains supported. Using a multilingual, grey-literature-inclusive search across seven languages, we added 69 effect sizes from 29 studies and harmonised the two source metrics, giving 227 Hedges' g effect sizes from 87 studies and 172 log odds ratios from 69 studies across 33 species. Reanalysed with phylogenetically informed multilevel models, the overall effect remained positive and significant but was appreciably smaller [Hedges' g=0.45, 95% confidence interval (CI): 0.25 to 0.66, versus 0.58; odds ratio=1.81, 95% CI: 1.34 to 2.45, versus 2.71]. Heterogeneity was high, located within studies and among species rather than across the phylogeny. Small-study effects were strong on both scales, and correcting for them drew the mean towards zero. In some species, individuals do copy conspecifics' mate choices, but the effect is smaller, more variable, and more sensitive to publication bias than previously estimated.

  • Open Access

    One dataset, four meta-analyses: synthesising mean effects and heterogeneity within and among populations

    by Haley Lacza, Ayumi Mizuno, Coralie Williams, Malgorzata Lagisz, Shinichi Nakagawa

    Ecological syntheses (meta-analyses) usually ask ‘what is the average effect?’ But many questions also depend on whether outcomes become more or less variable and whether effects are predictable across contexts. We show how one dataset can support a workflow that synthesises both mean effects and within-group variability while also separating heterogeneity among study-level effects from residual heterogeneity among effect sizes within studies. Using an organic versus conventional crop-yield dataset, we analyse mean effects with the log response ratio (lnRR; Model 1) and within-group variability with the log variability ratio (lnVR) and log coefficient of variation ratio (lnCVR; Model 2), all calculated from the same summary statistics (means, s.d.s. and sample sizes). We then use multilevel location–scale meta-regression to test whether fertiliser type shifts expected lnRR and lnVR/lnCVR effects (Models 3 and 4) and whether residual and study-level heterogeneity differ between contexts. These heterogeneity components respectively inform consistency among represented comparisons and transferability to a new study. Analysing mean effects, within-group variability and context-dependent heterogeneity together can provide a more informative description of ecological responses than focusing on average effects alone.

  • Open Access

    prepR4pcm: An R package for preparing data and trees for phylogenetic comparative methods

    by Shinichi Nakagawa, Santiago Ortega, Ayumi Mizuno, Eduardo S. A. Santos, Malgorzata Lagisz, Bhavya Jain, Jimuel Jr Celeste, Sergio Poo Hernandez

    Phylogenetic comparative methods require species names in a trait dataset to match tip labels in a phylogenetic tree. Yet this apparently simple prerequisite is often one of the most fragile steps in a comparative workflow. Names may differ because of, for example, formatting, taxonomic revisions, synonyms, or spelling errors. If these differences are resolved informally, species can be lost from analyses, and the reasons for their loss can be difficult to reconstruct. Here, we present


  • Open Access

    Weak and species-specific morphological change in Atlantic Forest birds across two decades of warming

    by Eduardo Santos, Ayumi Mizuno, Santiao Ortega, Gustavo Burin

    Birds are widely expected to shrink as the climate warms, but few tropical studies span decades. In the central Amazon, mean body mass fell in all 77 species studied and a third of species lengthened their wings. We analysed 24 years (1995–2018) of morphology from 73 passerine species (12,571 live-capture records; ATLANTIC BIRD TRAITS) across the Brazilian Atlantic Forest, a region that warmed by about 0.25°C per decade. The records came from 47 research programmes whose sites changed over time, so we modelled contributor and municipality explicitly and checked by simulation that these models recover trends planted in the real sampling structure. Unadjusted models suggested a steep wing-length decline and increases in bill width and trait variability. Once sampling turnover was modelled, the wing decline was less than half as steep (−1.0% per decade, 95% credible interval spanning zero), and the bill-width and variability changes disappeared. Body mass declined by 0.9% per decade, close to the mean Amazonian rate of 1.0%, but its interval spanned zero (−2.8 to +1.0%). Responses varied strongly among species, and most species showed no change distinguishable from zero. Wings did not lengthen, and among individuals measured for both traits, mass tended to be retained relative to wing length. Neither trait was associated with interannual temperature anomalies. Atlantic Forest passerines thus show neither the consistent mass loss across species nor the wing elongation reported from the Amazon, although their average mass change is compatible with the Amazonian rate. Compiled trait databases can extend long-term monitoring to under-studied tropical regions, but only when sampling structure is modelled: ignoring it here would have more than doubled the estimated wing decline.

  • Open Access

    Mastering a simulation-based parametric method to obtain bias-corrected point estimates and sampling variance for effect sizes

    by Shinichi Nakagawa, Ayumi Mizuno, Coralie Williams, Santiago Ortega, Szymon M. Drobniak, Malgorzata Lagisz, Yefeng Yang, Alistair M. Senior, Daniel W. A. Noble, Erick Lundgren

    Meta-analyses require an effect-size estimate and its corresponding sampling variance from primary studies. For some effect size statistics, however, estimators of sampling variance are unavailable, requiring new derivations. Traditionally, such formulas are obtained using hand-derived Taylor expansions (the delta method), but this approach can be technically demanding and inaccessible to many applied researchers. Building on the idea of single-fit parametric resampling, we introduce the SAFE bootstrap: a Single-fit, Accurate, Fast, and Easy simulation recipe that replaces complex algebra with four intuitive steps: fit, draw, transform, and summarise. SAFE is a model-based parametric bootstrap applied to summary statistics, so its performance depends on how well the assumed sampling model approximates the sampling distribution of the reported statistics. Here, we focus on two-group effect sizes that can be parameterised from standard reported summaries, using Gaussian working models for continuous outcomes and binomial or multinomial models for discrete outcomes. Within this framework, SAFE yields model-based, bias-corrected point estimates and standard errors for familiar effect sizes, and readily extends to less common statistics. We first demonstrate SAFE with a simple example, then apply it to common effect sizes, such as the standardised mean difference and the log odds ratio, as well as several less common measures. With additional coding, SAFE can also accommodate zero values and small sample sizes, albeit with important caveats. Our tutorial, accompanied by R code supplements, aims both to clarify the logic of simulation-based estimation from summary statistics and to provide a practical workflow when only summary statistics are available.

  • Open Access

    A New Effect Size for Meta‐Analysis of Magnitude: lnM

    by Shinichi Nakagawa, Ayumi Mizuno, Coralie Williams, Malgorzata Lagisz, Rose E. O'DEa, Daniel W. A. Noble, Alistair M. Senior, Erick Lundgren, Santiago Ortega

    Meta‐analyses in ecology and evolution sometimes target the magnitude of differences between groups rather than their direction, for example, when the question concerns deviation from a biological optimum or divergence from a reference state. A common practice is to convert signed effects into magnitudes by taking absolute values, but this can induce upward distortion and non‐normal sampling distributions under standard meta‐analytic models. Here we introduce lnM, a log‐ratio effect size for the magnitude of difference between two groups, defined from one‐way ANOVA components. lnM is asymptotically normal, supports standard multilevel meta‐analysis and meta‐regression and applies to both ratio‐ and interval‐scale traits. Using theory, simulations and worked examples, we show when delta‐method and single‐fit bootstrap estimators perform well, and how lnM can be used to assess publication bias.


  • Open Access

    An item bank of reporting items and concepts in evidence synthesis reporting guidelines

    by Kyle Morrison, Edward R. Ivimey-Cook, Malgorzata Lagisz, Ayumi Mizuno, Manuela Santos Santana, Eduardo S. A. Santos, Alfredo Sánchez-Tójar, Zachary P. Simpson, Matthew James Grainger, Biljana Macura, et al.

    Evidence synthesis integrates findings from multiple research studies, and its value for decision making depends on complete, transparent reporting. Authors often rely on reporting guidelines, checklists, flow diagrams, or structured text guiding manuscript reporting, but they still struggle to implement what counts as clear, complete reporting. A central issue is that many guidelines address the same underlying concepts using different wording, making it unclear whether differences between guidelines are meaningful or merely semantic. To address this, we compiled an item bank of 2,249 reporting items from 59 unique evidence synthesis reporting guidelines (from 83 studies, 68 checklists), using a six-step process: literature searching and screening, item collation, readability analysis, concept extraction, and overlap assessment. Compiled items were relatively verbose (mean words/item = 19.9) and linguistically complex (Moving-Average Type-Token Ratio = 0.959), indicating limited reuse of consistent wording—likely making items harder to interpret and compare. Because items can contain multiple concepts, and concepts can recur across items and guidelines, we identified 403 unique reporting concepts. Guidelines also showed modest conceptual overlap (Jaccard Similarity Index = 0.166, n = 1,830 pairwise comparisons), suggesting some concepts were widely shared across reporting guidelines for different methodologies and disciplines, pointing to clear potential to harmonise reporting guidance and reduce duplicated effort. We provide all reporting items and concepts in the supplementary files. We further discuss trends and gaps in item use and wording across guidelines, outline applications of the collated item bank, and note limitations that point to future research. Together, our findings offer much-needed direction for developing and harmonising future reporting guidelines, an essential step toward more consistent and transparent reporting across evidence synthesis.