Modelling publication bias and p-hacking

P-hacking in meta-analyses: A formalization and new meta-analytic methods

As traditionally conceived, publication bias arises from selection operating on a collection of individually unbiased estimates. A canonical form of such selection across studies (SAS) is the preferential publication of affirmative studies (i.e., those with significant, positive estimates) versus nonaffirmative studies (i.e., those with nonsignificant or negative estimates). However, meta-analyses can also be compromised by selection within studies (SWS), in which investigators "p-hack" results within their study to obtain an affirmative estimate. Published estimates can then be biased even conditional on affirmative status, which comprises the performance of existing methods that only consider SAS. We propose two new analysis methods that accommodate joint SAS and SWS; both analyze only the published nonaffirmative estimates. First, we propose estimating the underlying meta-analytic mean by fitting "right-truncated meta-analysis" (RTMA) to the published nonaffirmative estimates. This method essentially imputes the entire underlying distribution of population effects. Second, we propose conducting a standard meta-analysis of only the nonaffirmative studies (MAN); this estimate is conservative (negatively biased) under weakened assumptions. We provide an R package (phacking) and website (metabias.io). Our proposed methods supplement existing methods by assessing the robustness of meta-analyses to joint SAS and SWS.

The Influence of Alpha Frequency on Temporal Binding across the Senses: Response to the Special Focus

The papers collected in this Special Focus, prompted by S. Buergers and U. Noppeney [The role of alpha oscillations in temporal binding within and across the senses. Nature Human Behaviour, 6, 732-742, 2022], have raised several interesting ideas, arguments, and empirical results relating to the alpha temporal resolution hypothesis. Here we briefly respond to these, and in the process emphasize four challenges for future research: defining the scope and limitation of the hypothesis; developing experimental paradigms and study designs that rigorously test its tenets; decomposing the scalp-level signal and isolating underlying neural circuits; and bringing uniformity to the current diversity of analysis and statistical methods. Addressing these challenges will facilitate the progression from merely correlating alpha frequency with various perceptual phenomena to establishing whether and (if so) how alpha frequency influences sensory integration and segregation.

Alpha Oscillations and Temporal Binding Windows in Perception-A Critical Review and Best Practice Guidelines

An intriguing question in cognitive neuroscience is whether alpha oscillations shape how the brain transforms the continuous sensory inputs into distinct percepts. According to the alpha temporal resolution hypothesis, sensory signals arriving within a single alpha cycle are integrated, whereas those in separate cycles are segregated. Consequently, shorter alpha cycles should be associated with smaller temporal binding windows and higher temporal resolution. However, the evidence supporting this hypothesis is contentious, and the neural mechanisms remain unclear. In this review, we first elucidate the alpha temporal resolution hypothesis and the neural circuitries that generate alpha oscillations. We then critically evaluate study designs, experimental paradigms, psychophysics, and neurophysiological analyses that have been employed to investigate the role of alpha frequency in temporal binding. Through the lens of this methodological framework, we then review evidence from between-subject, within-subject, and causal perturbation studies. Our review highlights the inherent interpretational ambiguities posed by previous study designs and experimental paradigms and the extensive variability in analysis choices across studies. We also suggest best practice recommendations that may help to guide future research. To establish a mechanistic role of alpha frequency in temporal parsing, future research is needed that demonstrates its causal effects on the temporal binding window with consistent, experimenter-independent methods.

Telomere Length in Patients with Gestational Diabetes Mellitus and Normoglycemic Pregnant Women: a Systematic Review and Meta-analysis

We performed a systematic review and meta-analysis of studies assessing telomere length in blood leukocytes or mononuclear cells in women with gestational diabetes mellitus (GDM) and normoglycemic pregnant women (NPW) and their infants. The review protocol was registered in PROSPERO (CRD42022300950). Searches were conducted in PubMed, Embase, LILACS, CNKI, and Wang Fang, from inception through November 2022. The primary outcomes were maternal and offspring telomere length. The Newcastle-Ottawa Scale was used to assess the quality of included studies. Random-effect meta-analyses were applied to estimate standardized mean differences (SMDs) and their 95% confidence interval (CI). The meta-analysis of four studies showed no significant maternal telomere length difference (SMD = -0.80, 95% CI: -1.66, 0.05) in women with GDM compared to NPW. In the sensibility analysis omitting one study with a small sample of women, the telomere length becomes significantly reduced in women with GDM (SMD = -1.10, 95% CI: -2.18, -0.02). GDM patients had increased glucose (SMD = 0.28, 95% CI: 0.09, 0.46) and glycosylated hemoglobin than NPW (SMD = 0.62, 95% CI: 0.23, 1.01) while total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglycerides did not display differences between women with and without GDM. There was no significant difference in cord blood telomere length in offspring from women with GDM and NPW (SMD = 0.11, 95% CI: -0.52, 0.30). Cord blood insulin levels (SMD = 0.59, 95% CI: 0.33, 0.85) and birthweight (SMD = 0.59, 95% CI: 0.39, 0.79) were higher in offspring from pregnant women with GDM than in those from NPW. There were no significant differences in maternal and offspring telomere length between pregnancies with and without GDM.

Infinite diameter confidence sets in Hedges’ publication bias model

Meta-analysis, the statistical analysis of results from separate studies, is a fundamental building block of science. But the assumptions of classical meta-analysis models are not satisfied whenever publication bias is present, which causes inconsistent parameter estimates. Hedges’ selection function model takes publication bias into account, but estimating and inferring with this model is tough for some datasets. Using a generalized Gleser–Hwang theorem, we show there is no confidence set of guaranteed finite diameter for the parameters of Hedges’ selection model. This result provides a partial explanation for why inference with Hedges’ selection model is fraught with difficulties.