Belz and colleagues present GREAT, a seven-item clinical instrument for predicting electroconvulsive therapy response in unipolar depression, with promising discriminatory validity (AUC 0.841). We identify three methodological gaps – absent calibration, limited sample representativeness and unquantified incremental value – that must be addressed before GREAT can guide individual clinical decisions.

Multilevel regression and poststratification (MRP) is widely used to estimate opinion in small subgroups and to adjust unrepresentative surveys. Yet, even flexible MRP models contain errors generated by non-response and model misspecification. We propose a principled, data-driven method to leverage observable errors on auxiliary quantities with known marginal distributions—for example, election outcomes—to improve estimates of policy attitudes. Our method leverages the correlation between auxiliary variables and outcomes of interest to calibrate MRP estimates to these known marginal distributions. We illustrate our approach using a pre-election poll measuring support for an abortion referendum. We find that the method reduces county-level error by nearly two-thirds relative to traditional MRP. We also show how our calibration approach can be used to generate estimates for smaller nested geographies, such as precincts, even in the absence of poststratification data at this level. Our approach provides a framework for fully incorporating known population data to improve estimates of public opinion in small subgroups, providing scholars another tool to study representation.

Advances in LED lighting technologies enable increasingly complex light regimes, providing greater insight into plants’ responses to dynamic light – such as seasonality and fluctuating conditions – rather than the traditional discrete (i.e., on/off) lighting. However, current methods of programming such regimes are time-consuming and/or limited to 1–2 wavebands. Robust methods are therefore needed to accurately programme multichannel/waveband LED lighting systems. We present a multistep, multidimensional algorithm to accurately programme multi-waveband LED lights. This algorithm accounts for non-linearity between intensity settings and measured light quantity output, as well as optical crosstalk between channels of different wavebands. It outperforms methods that treat waveband channels as independent variables, allowing users to more accurately programme multichannel light regimes. This will allow the community to probe plant responses to dynamically changing light spectra. We have made this algorithm available as an R package, LightFitR (installable from CRAN with ‘install.packages(“LightFitR”)’.

Evaluation of clinical prediction models across multiple clusters, whether centers or datasets, is becoming increasingly common. A comprehensive evaluation includes an assessment of the agreement between the estimated risks and the observed outcomes, also known as calibration. Calibration is of utmost importance for clinical decision making with prediction models, and it often varies between clusters. We present three approaches to take clustering into account when evaluating calibration: (1) clustered group calibration (CG-C), (2) two-stage meta-analysis calibration (2MA-C), and (3) mixed model calibration (MIX-C), which can obtain flexible calibration plots with random effects modeling and provide confidence interval (CI) and prediction interval (PI). As a case example, we externally validate a model to estimate the risk that an ovarian tumor is malignant in multiple centers (N = 2489). We also conduct a simulation study and a synthetic data study generated from a true clustered dataset to evaluate the methods. In the simulation study, MIX-C and 2MA-C (splines) gave estimated curves closest to the true overall curve. In the synthetic data study, MIX-C produced cluster-specific curves closest to the truth. Coverage of the PI across the plot was best for 2MA-C with splines. We recommend using 2MA-C with splines to estimate the overall curve and 95% PI and MIX-C for cluster-specific curves, especially when the sample size per cluster is limited. We provide ready-to-use code to construct summary flexible calibration curves, with CI and PI to assess heterogeneity in calibration across datasets or centers.

The impact of misalignment errors, specifically yaw and pitch deviations, on dihedral reflectors’ scattering responses is studied for millimeter-wave polarimetric multiple-input multiple-output automotive radars. Through simulations and experiments at 77 GHz, it is demonstrated that significant radar cross-section (RCS) variations of up to 30 dB can occur within small misalignment ranges (0$^{\circ}$–2$^{\circ}$). The findings emphasize that larger dihedral dimensions can amplify sensitivity to misalignment in some specific misalignment scenarios, offering trade-offs between reflection strength and robustness to misalignment errors. The study also explores near-field effects, revealing notable discrepancies between the dihedral near- and far-field scattering response in misalignment scenarios. A polarimetric calibration method is applied to show how polarimetric channel phase response is affected under such conditions, achieving stable results in specific configurations (e.g., dihedral at 0$^{\circ}$ under yaw misalignment angle). This study addresses key challenges in calibration accuracy, including the high sensitivity of RCS to small angular misalignments, the trade-offs between reflector dimensions and robustness, and the influence of near-field effects in practical setups.

Radiocarbon dating is a widely used method in archaeology and earth sciences, but the precision of calibrated dates from single radiocarbon measurements can be difficult to understand. This study investigates the precision of calibrated radiocarbon dates depending on the uncertainties of the measurement and the details of the calibration curve. Using data for the Holocene epoch and the IntCal20 calibration curve, over 1,000,000 hypothetical radiocarbon measurements were calibrated and analyzed. The study shows that high-precision measurements can yield calibrated date ranges from less than 50 years to more than 200 years (at the 95.4% probability) depending on the specifics of the calibration curve. This research may serve as a tool for planning future studies and assessing whether high-precision measurements are beneficial for proposed case.

Plague and famine are two of the worst killers in human history. Both struck the Czech lands in the Middle Ages not long after each other (the famine of 1318 CE and the plague of 1348–1350 CE). The aim of our study was to try to relate the mass graves found in the vicinity of the Chapel of All Saints with an ossuary in the Kutná Hora–Sedlec site to these two specific events. For this purpose, we used stratigraphic and archaeological data, radiocarbon dating, and Bayesian modeling of 172 calibrated AMS ages obtained from teeth and bones of 86 individuals buried in the mass graves. Based on the stratigraphic and archaeological data, five mass graves were interpreted as famine graves and eight mass graves were interpreted as plague graves. Using these data and the calibration of the radiocarbon results of the tooth-bone pairs of each individual, we constructed the Bayesian model to interpret the remaining mass graves for which no contextual information was available (eight mass graves). In terms of Bayesian model results, the model fits stratigraphic data in 23 out of 34 cases and in all seven cases based on calibration data. To validate the model results on archaeologically and stratigraphically uninterpreted data, ancient DNA analysis is required to identify Yersinia pestis.

Deep geological repositories are critical for the long-term storage of hazardous materials, where understanding the mechanical behavior of emplacement drifts is essential for safety assurance. This study presents a surrogate modeling approach for the mechanical response of emplacement drifts in rock salt formations, utilizing Gaussian processes (GPs). The surrogate model serves as an efficient substitute for high-fidelity mechanical simulations in many-query scenarios, including time-dependent sensitivity analyses and calibration tasks. By significantly reducing computational demands, this approach facilitates faster design iterations and enhances the interpretation of monitoring data. The findings indicate that only a few key parameters are sufficient to accurately reflect in-situ conditions in complex rock salt models. Identifying these parameters is crucial for ensuring the reliability and safety of deep geological disposal systems.

Sudden annual rises in radiocarbon concentration have proven to be valuable assets for achieving exact-year calibration of radiocarbon measurements. These extremely precise calibrations have usually been obtained through the use of classical χ2 tests in conjunction with a local calibration curve of single-year resolution encompassing a rapid change in radiocarbon levels. As the latest Northern Hemisphere calibration curve, IntCal20, exhibits single-year resolution over the last 5000 years, in this study we investigate the possibility of performing calibration of radiocarbon dates using the classical χ2 test and achieving high-precision dating more extensively, examining scenarios without the aid of such abrupt changes in radiocarbon concentration. In order to perform a broad analysis, we simulated 171 sets of radiocarbon measurements over the last two millennia, with different set lengths and sample spacings, and tested the effectiveness of the χ2 test compared to the most commonly used Bayesian wiggle-matching technique for temporally ordered sequences of samples such as tree-rings sequences, the OxCal D_Sequence. The D_Sequence always produces a date range, albeit in certain cases very narrow; the χ2 test proves to be a viable alternative to Bayesian wiggle-matching, as it achieves calibrations of comparable precision, providing also a highest-likelihood estimate within the uncertainty range.

When large achievement tests are conducted regularly, items need to be calibrated before being used as operational items in a test. Methods have been developed to optimally assign pretest items to examinees based on their abilities. Most of these methods, however, are intended for situations where examinees arrive sequentially to be assigned to calibration items. In several calibration tests, examinees take the test simultaneously or in parallel. In this article, we develop an optimal calibration design tailored for such parallel test setups. Our objective is both to investigate the efficiency gain of the method as well as to demonstrate that this method can be implemented in real calibration scenarios. For the latter, we have employed this method to calibrate items for the Swedish national tests in Mathematics. In this case study, like in many real test situations, items are of mixed format and the optimal design method needs to handle that. The method we propose works for mixed-format tests and accounts for varying expected response times. Our investigations show that the proposed method considerably enhances calibration efficiency.

A distinction is proposed between measures and predictors of latent variables. The discussion addresses the consequences of the distinction for the true-score model, the linear factor model, Structural Equation Models, longitudinal and multilevel models, and item-response models. A distribution-free treatment of calibration and error-of-measurement is given, and the contrasting properties of measures and predictors are examined.

The conventional method of measuring ability, which is based on items with assumed true parameter values obtained from a pretest, is compared to a Bayesian method that deals with the uncertainties of such items. Computational expressions are presented for approximating the posterior mean and variance of ability under the three-parameter logistic (3PL) model. A 1987 American College Testing Program (ACT) math test is used to demonstrate that the standard practice of using maximum likelihood or empirical Bayes techniques may seriously underestimate the uncertainty in estimated ability when the pretest sample is only moderately large.

Overconfidence plays a role in a large number of individual decision biases and has been considered a ‘meta-bias’ for this reason. However, since overconfidence is measured behaviorally with respect to particular tasks (in which performance varies across individuals), it is unclear whether people generally vary in terms of their general overconfidence. We investigated this issue using a novel measure: the Generalized Overconfidence Task (GOT). The GOT is a difficult perception test that asks participants to identify objects in fuzzy (‘adversarial’) images. Critically, participants’ estimated performance on the task is not related to their actual performance. Instead, variation in estimated performance, we argue, arises from generalized overconfidence, that is, people indicating a cognitive skill for which they have no basis. In a series of studies (total N = 1,293), the GOT was more predictive when looking at a broad range of behavioral outcomes than two other overestimation tasks (cognitive and numeracy) and did not display substantial overlap with conceptually related measures (Studies 1a and 1b). In Studies 2a and 2b, the GOT showed superior reliability in a test–retest design compared to the other overconfidence measures (i.e., cognitive and numeracy measures), particularly when collecting confidence ratings after each image and an estimated performance score. Finally, the GOT is a strong predictor of a host of behavioral outcomes, including conspiracy beliefs, bullshit receptivity, overclaiming, and the ability to discern news headlines.

Emotion recognition in conversation (ERC) faces two major challenges: biased predictions and poor calibration. Classifiers often disproportionately favor certain emotion categories, such as neutral, due to the structural complexity of classifiers, the subjective nature of emotions, and imbalances in training datasets. This bias results in poorly calibrated predictions where the model’s predicted probabilities do not align with the true likelihood of outcomes. To tackle these problems, we introduce the application of conformal prediction (CP) into ERC tasks. CP is a distribution-free method that generates set-valued predictions to ensure marginal coverage in classification, thus improving the calibration of models. However, inherent biases in emotion recognition models prevent baseline CP from achieving a uniform conditional coverage across all classes. We propose a novel CP variant, class spectrum conformation, which significantly reduces coverage bias in CP methods. The methodologies introduced in this study enhance the reliability of prediction calibration and mitigate bias in complex natural language processing tasks.

This study suggests that there may be considerable difficulties in providing accurate calendar age estimates in the Roman period in Europe, between ca. AD 60 and ca. AD 230, using the radiocarbon calibration datasets that are currently available. Incorporating the potential for systematic offsets between the measured data and the calibration curve using the ΔR approach suggested by Hogg et al. (2019), only marginally mitigates the biases in calendar date estimates observed. At present, it clearly behoves researchers in this period to “caveat emptor” and validate the accuracy of their calibrated radiocarbon dates and chronological models against other sources of dating information.