The I-squared statistic has become hugely popular in meta-analysis. Motivated by the recent statement that ‘Informally, it can be considered to reflect the extent to which confidence intervals displayed in a forest plot do not overlap with each other’, this short note formalises this perspective. The insights provided by our analysis lend further weight to the interpretation of the I-squared statistic as a measure of inconsistency. This perspective reminds us that visual inspections of forest plots are vitally important when assessing between-study variation.

This study conducts a systematic review and meta-analysis to evaluate the impact of Pakistan’s Benazir Income Support Programme (BISP), the country’s largest unconditional cash transfer initiative, on poverty reduction and women’s empowerment. Using robust variance estimation (RVE) meta-analysis to account for effect-size dependencies and heterogeneity, the study synthesises quantitative evidence from evaluations of BISP. The findings show a modest but statistically significant positive effect on poverty alleviation. In contrast, there is no statistically significant evidence that BISP improves women’s empowerment, although individual studies report some gains in mobility and decision-making. High heterogeneity across studies indicates substantial variation in outcomes across contexts and measures. The findings highlight the limitations of relatively low-value cash transfers in generating transformative improvements in poverty and women’s empowerment. They suggest that complementary interventions and broader structural reforms may be necessary for cash transfers to produce more substantial and sustained improvements in household welfare and women’s agency.

Meta-analysis is often viewed as a gold standard for synthesizing available evidence, but it faces challenges as evidence becomes available gradually and studies may be disseminated selectively and at different times. When publication timing is systematically related to study results, meta-analytic conclusions at any given time can be biased; for example, studies with statistically significant or larger effects may be published earlier than those with null or smaller effects. This phenomenon, commonly referred to as time-lag bias, is a key reason why early meta-analytic conclusions may later weaken, reverse, or become more heterogeneous as additional evidence accrues. To address the limitations of existing definitions and detection approaches, which are largely tied to monotone, effect-size-driven dissemination, we broaden the concept of time-lag bias beyond the traditional yet simple notion that “positive” studies are published earlier. Specifically, we view time-lag bias more generally as a manifestation of time-dependent evidence availability, whereby the timing of study dissemination can distort meta-analytic conclusions at different stages of evidence accrual. We illustrate this broader perspective through several real-world case studies, including classical time-lag bias, inverse time-lag bias, non-monotone temporal patterns, and anomalous evidence localized within specific time windows. These examples show that time-lag bias can arise through multiple dissemination mechanisms and may exaggerate, attenuate, or even reverse pooled estimates over time. They highlight the importance of explicitly accounting for the temporal dimension of evidence accumulation and motivate the routine use of diagnostic and sensitivity strategies to assess robustness to selective dissemination over time.

Public doubt about the usefulness of antidepressants appears to have grown, even as scientific agreement about their efficacy has become more settled. This paper examines the role that figures associated with critical psychiatry have had in promoting this doubt. It traces the long entanglement of antidepressants and their critics over the last eighty years, culminating in a 2022 review of the serotonin hypothesis, which, though it made no claims about efficacy, was widely interpreted as showing that antidepressants do not work. Drawing on Oreskes and Conway’s concept of ‘manufactured doubt’, the paper explores how certain philosophical commitments have shaped these actors’ views, at times impeding their engagement with new evidence. In doing so, it marks a distinction between productive and unproductive doubt. Whilst external scrutiny has played a significant role in shaping psychiatry, some positions are held so rigidly that the resulting critiques obscure the scientific consensus and risk undermining psychiatric care.

Meta-analytic structural equation modeling (MASEM) is a flexible modeling framework that combines the strengths of meta-analysis and structural equation modeling. Despite the increasing popularity of the MASEM technique, its usage and scope are limited to relatively simple meta-analytic data structures and moderator hypotheses. Within-study variability in moderator variables and dependence in primary studies’ correlation estimates tend to be handled using ad hoc methods. In this study, we demonstrate use of multilevel multivariate modeling with robust variance estimation (RVE) for handling within-study dependence when pooling studies’ correlation estimates and we propose how to incorporate categorical moderator values directly into the structural equation model for moderators with values varying within studies. We outline the proposed approach using three stages: (a) specification of the structural equation model using moderator variable categories, (b) pooling correlation matrix elements using the three-level multivariate model with RVE, and (c) testing the moderator hypotheses through either model comparison or direct estimation of structural equation model’s path differences. We illustrate the approach using three examples to demonstrate its flexibility across different MASEM models. This framework broadens the scope of moderator hypotheses that can be tested using MASEM and enhances the methodological rigor for handling dependencies in studies’ correlation estimates for MASEM.

Emerging evidence suggests that dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis, particularly cortisol, may contribute to the cognitive deficits commonly observed in psychotic disorders. We conducted the first combined systematic review and quantitative synthesis of associations between cortisol and cognition in psychosis. Twelve studies met inclusion criteria (940 participants; 74.68% cases; 39.57% female; mean age = 31.83 years). Of these, nine studies (822 participants; 71.04% cases; 35.96% female; mean age = 32.72 years) contributed data to three-level meta-analyses examining associations between basal cortisol and composite cognition, attention, working memory and processing speed (AWP), executive functioning (EF), and verbal learning and memory (VM). Cortisol sampling time significantly moderated the relationship between cortisol and EF (F = 10.98, p = .004), with a significant negative association emerging between afternoon/evening cortisol and EF (r = −0.27, 95% CI [−0.42, −0.12], p = .01). No other robust cortisol–cognition associations were observed. However, sensitivity analyses revealed a negative association between morning cortisol and AWP following exclusion of an influential study. These findings do not support a general association between basal cortisol and cognitive functioning in psychosis. Instead, they suggest that cortisol–cognition relationships may be both domain- and sampling-time-specific, with the strongest evidence linking elevated afternoon/evening cortisol to poorer executive functioning. Further research using more precise assessments of cortisol dynamics is needed to clarify the role of HPA-axis dysregulation in cognitive impairment across psychotic disorders.

Viral interference between SARS-CoV-2 and other respiratory virus infections has been implicated, but symptom-based testing in observational studies may bias estimations. We used individual participant data from ten cohort studies conducted during pandemic years 2020–2024 to study virus interference. Respiratory samples were PCR-tested for viruses including SARS-CoV-2. The primary analysis was based on samples collected from cohort participants regardless of respiratory symptoms. A secondary analysis included only symptom-based testing data. We estimated odds ratios for co-occurrence of SARS-CoV-2 with other respiratory viruses. The primary analysis included data from six cohorts (n = 1 023). In total 205 of 2 465 samples were SARS-CoV-2 positive (8.3%). Co-detections with SARS-CoV-2 occurred less frequently than expected under the assumption of zero interference (OR: 0.72; 95% CI: 0.40–1.25) for samples positive for another virus. The secondary analysis (five cohorts, n = 1 715) included 2 088 samples from symptomatic subjects, with 506 SARS-CoV-2 infections (24.2%). The OR of SARS-CoV-2 detection was 0.16 (95% CI: 0.10–0.25) for samples positive for another virus. Minimizing selection bias, our results still point towards negative viral interference with SARS-CoV-2, but effect-estimates are weaker and non-significant. Apart from selection bias, the stronger effects found in symptomatic individuals could also indicate that symptomatic infection enhances virus interference.

This meta-analysis examines emotional Stroop interference in bilinguals, a literature characterised with inconsistent findings on age of acquisition (AoA), emotional valence, and L2 proficiency. Following PRISMA 2020 guidelines, a systematic search across six electronic databases identified 10 studies (12 independent effect sizes, N = 450). Random-effects meta-analysis yielded a non-significant pooled effect characterized by extreme heterogeneity. Mixed-effects moderator models identified age of L2 acquisition as the critical boundary condition: early-acquired bilinguals exhibited reliable interference, whereas late-acquired bilinguals did not. Proficiency assessment method and valence eligibility were non-significant, suggesting that superficial methodological differences do not account for the literature’s variability. Formal stimulus-type analysis was precluded by a near-absence of emotion-label studies. These findings indicate that emotional Stroop interference in bilinguals is conditional on neurodevelopmental timing, supporting neurodevelopmental models of integrated language–emotion networks. The results underscore the need for standardised paradigms and more diverse bilingual samples in future primary research.

Fibromyalgia syndrome (FMS) is a chronic condition characterised by widespread musculoskeletal pain, fatigue, sleep disturbances, emotional symptoms and cognitive impairment. Pain catastrophising, a maladaptive cognitive-affective response to pain, has been consistently associated with greater symptom burden and poorer pain-related outcomes in FMS. Multicomponent pain therapy (MPT), an interdisciplinary approach that combines several therapeutic modalities, may address the multifaceted symptom burden of FMS. This study aimed to synthesise the available evidence on whether MPT was associated with reductions in pain catastrophising among adults with FMS from a transdisciplinary perspective. A systematic review and meta-analysis were conducted in accordance with Cochrane Collaboration guidance and the PRISMA statement; the protocol was registered in PROSPERO (CRD42024596489). Randomised controlled trials (RCTs), including pilot RCTs, were eligible if they assessed pain catastrophising in adults diagnosed with FMS. Data were extracted and analysed using Comprehensive Meta-Analysis software, and subgroup analyses were conducted according to intervention duration. The certainty of evidence was assessed using GRADE. Thirteen RCTs comprising 1,314 participants were included. At post-intervention, MPT was associated with a medium reduction in pain catastrophising (SMD = −0.668; 95% CI: −0.953 to −0.383; p < .001), with moderate heterogeneity (I2 = 43.5%). At the 3-month follow-up, 6 studies comprising 472 participants yielded a larger pooled estimate (SMD = −0.957; 95% CI: −1.543 to −0.370; p = .001), with moderate heterogeneity (I2 = 49.4%). Medium-duration interventions (11–20 sessions) yielded the largest subgroup estimate. Funnel-plot asymmetry was detected, although trim-and-fill-adjusted analyses remained in favour of MPT. Overall, MPT may reduce pain catastrophising in adults with FMS, and medium-duration interventions yielded the largest subgroup estimate. However, the certainty of evidence was low at post-intervention and very low at the 3-month follow-up; therefore, the magnitude and stability of these findings remain uncertain.

Developing accurate, dynamic feeding formulas for growing-finishing pigs via meta-analysis of large datasets can effectively reduce feed waste from ad libitum feeding and curb environmental pollution from nutrient excretion in feces and urine. This study aims to obtain a dynamic feed formula suitable for growing-finishing pigs through model fitting and to verify its effectiveness. We utilized data from 523 treatments across 112 studies to establish the relationship between daily digestible energy intake and body weight in growing-finishing pigs using three fitting models: linear mixed-effects (LME), Bayesian ridge regression (BRR) and random forest regression (RFR) models. After analysis, LME is the optimized dynamic model. To verify its validity, we selected 16 healthy Duroc × Landrace × Yorkshire crossbred castrated male pigs with an average initial body weight of about 85 kg and randomly assigned them to two groups. The Con group was allowed ad libitum feeding, while the Meta group used the LME model to restrict feeding. After 42 days, the growth performance between the two groups were no significant difference. The Meta group exhibited significantly lower feed-to-gain and average daily feed intake (P < 0.05), serum low-density lipoprotein cholesterol and aspartate aminotransferase levels (P < 0.05), and an increased abundance of beneficial gut microbial communities. The LME model, a precise and dynamic feeding formula derived from meta-analysis, effectively curbs feed intake while enhancing nutrient utilization and feed efficiency. It also mitigates environmental pollution and positively modulates gut microbiota composition.

Cannabis use is a known risk factor for adverse mental health outcomes, including psychosis, which is frequently associated with aggression. However, the specific impact of cannabis use on the risk of violence remains unclear. This meta-analysis, conducted according to PRISMA guidelines, synthesized evidence from 63 peer-reviewed studies (N = 265,079) identified via MEDLINE, EMBASE, and PsycINFO searches up to November 2024 to clarify the link between cannabis use and violence, either as a perpetrator or as a victim. Random-effects models revealed that cannabis users are at a significantly higher risk of perpetrating violent acts. This association was observed in both psychiatric patients (OR 2.49, 95% CI 1.72–3.61) and the general population (OR 2.05, 95% CI 1.74–2.41), and remained significant across both longitudinal (OR 1.17, 95% CI 1.07–1.28) and cross-sectional (OR 2.37, 95% CI 1.66–3.40) study designs. There was an especially large effect for violence involving a criminal conviction in both general (OR 3.75, 95% CI 2.54–5.53) and psychiatric (OR 4.21, 95% CI 2.86–6.22) populations. Furthermore, cannabis use was significantly associated with an increased risk of becoming a victim of violence (OR 1.49, 95% CI 1.38–1.60), a risk more pronounced in females and mixed-gender cohorts compared to males. This meta-analysis indicates an increased risk of violence, especially that resulting in a criminal conviction, among cannabis users. It also highlights the potential of cannabis use as a target for interventions to decrease violence, and underscores the need for routine assessment of cannabis use among psychiatric patients.

Parent training programs (PTPs) aim to improve parenting skills and reduce children’s behavioral and emotional problems. Given that recent research has extended PTPs’ application from externalizing difficulties to internalizing symptoms – problem domains that frequently overlap in children – evaluating their effects together can clarify PTPs’ broader impact. This meta-analysis synthesized findings from 34 randomized controlled trials involving children aged 3–12 years to examine the efficacy of PTPs and identify factors influencing treatment outcomes. Across studies, PTPs significantly reduced both internalizing (g = 0.28) and externalizing (g = 0.37) problems, yielding effect sizes comparable to those reported in prior parent training research. Importantly, PTPs also demonstrated significant cross-domain effects, indicating that programs targeting externalizing problems may also reduce internalizing problems, and vice versa. Moderator analyses indicated that only children’s clinical diagnoses significantly predicted externalizing outcomes. These findings suggest that PTPs may be implemented flexibly across varying delivery conditions. Despite small overall effects, heterogeneity, and publication bias, the findings support the effectiveness and adaptability of PTPs. Future research should employ mechanism-informed, scalable designs to enhance accessibility and optimize treatment efficiency.

Meta-analyses in social science fields face multiple methodological challenges arising from how primary research studies are designed and reported. One challenge is that many primary studies report multiple relevant effect size estimates, leading to a data structure with dependent observations. Another is selective reporting bias, which arises when the availability of study findings is influenced by the statistical significance of results. Although many selective reporting diagnostics and bias-correction methods have been proposed, few are suitable for meta-analyses involving dependent effect sizes. Among available methods, step-function selection models are conceptually appealing and have shown promise in previous simulations, but tend to understate uncertainty in parameter estimates if dependent effects are ignored. We study methods for estimating step-function models from data involving dependent effect sizes, focusing specifically on estimating parameters of the marginal distribution of effect sizes and accounting for dependence using cluster-robust variance estimation or bootstrap resampling. We describe two estimation strategies, demonstrate them by re-analyzing data from a previous synthesis on ego depletion effects, and evaluate their performance through an extensive simulation study under both univariate and multivariate selection processes. Simulation findings indicate that selection models provide low-bias estimates of average effect size and clustered bootstrap confidence intervals provide acceptable coverage levels for meta-analyses with at least 30 studies. Although unadjusted estimates are more efficient when selective reporting is absent or minimal, the proposed methods provide a robust and useful approach for evaluating the potential biases arising when selective reporting is present in meta-analyses with dependent effects.