Accuracy and Precision of Fixed and Random Effects in Meta-Analyses of Randomized Control Trials for Continuous Outcomes

Meta-analyses of treatment effects in randomized control trials are often faced with the problem of missing information required to calculate effect sizes and their sampling variances. Particularly, correlations between pre- and posttest scores are frequently not available. As an ad-hoc solution, researchers impute a constant value for the missing correlation. As an alternative, we propose adopting a multivariate meta-regression approach that models independent group effect sizes and accounts for the dependency structure using robust variance estimation or three-level modeling. A comprehensive simulation study mimicking realistic conditions of meta-analyses in clinical and educational psychology suggested that the prevalent imputation approach works well for estimating the pooled effect but severely distorts the between-study heterogeneity. In contrast, the robust meta-regression approach resulted in largely unbiased fixed and random effects. Based on these results recommendations for meta-analytic practice and future meta-analytic developments are provided.

Power Approximations for Meta-Analysis of Dependent Effect Sizes

Meta-analytic models for dependent effect sizes have grown increasingly sophisticated over the last few decades, which has created challenges for a priori power calculations. We introduce power approximations for tests of average effect sizes based upon the most common models for handling dependent effect sizes. In a Monte Carlo simulation, we show that the new power formulas can accurately approximate the true power of common meta-analytic models for dependent effect sizes. Lastly, we investigate the Type I error rate and power for several common models, finding that tests using robust variance estimation provide better Type I error calibration than tests with model-based variance estimation. We consider implications for practice with respect to selecting a working model and an inferential approach.

Listening Ears or Reading Eyes: A Meta-Analysis of Reading and Listening Comprehension Comparisons

In this study, a meta-analysis of reading and listening comprehension comparisons across age groups was conducted. Based on robust variance estimation (46 studies; N = 4,687), the overall difference between reading and listening comprehension was not reliably different (g = 0.07, p = .23). Reading was beneficial over listening when the reading condition was self-paced (g = 0.13, p = .049) rather than experimenter-paced (g = −0.32, p = .16). Reading also had a benefit when inferential and general comprehension rather than literal comprehension was assessed (g = 0.36, p = .02; g = .15, p = .05; g = −0.01, p = .93, respectively). There was some indication that reading and listening were more similar in languages with transparent orthographies than opaque orthographies (g = 0.001, p = .99; g = 0.10, p = .19, respectively). The findings may be used to inform theories of comprehension about modality influences in that both lower-level skill and affordances vary comparisons of reading and listening comprehension. Moreover, the findings may guide choices of modality; however, both audio and written options are needed for accessible instruction.

The Long-Term Effectiveness of Internet-Based Interventions on Multiple Health Risk Behaviors: Systematic Review and Robust Variance Estimation Meta-analysis

Background Smoking tobacco, poor nutrition, risky alcohol use, and physical inactivity (SNAP) behaviors tend to cluster together. Health benefits may be maximized if interventions targeted multiple health risk behaviors together rather than addressing single behaviors. The internet has wide reach and is a sustainable mode for delivery of interventions for multiple health behaviors. However, no systematic reviews have examined the long-term effectiveness of internet-based interventions on any combination of or all SNAP behaviors in adults aged 18 years or older. Objective This systematic review examined, among adults (aged ≥18 years), the effectiveness of internet-based interventions on SNAP behaviors collectively in the long term compared with a control condition. Methods The electronic databases Medline, PsycINFO, Embase, CINAHL, and Scopus were searched to retrieve studies describing the effectiveness of internet-based interventions on ≥2 SNAP behaviors published by November 18, 2019. The reference lists of retrieved articles were also checked to identify eligible publications. The inclusion criteria were randomized controlled trials or cluster randomized controlled trials with adults examining an internet-based intervention measuring the effect on ≥2 SNAP behaviors at least 6 months postrecruitment and published in English in a peer-reviewed journal. Two reviewers independently extracted data from included studies and assessed methodological quality using the Quality Assessment Tool for Quantitative Studies. A robust variance estimation meta-analysis was performed to examine the long-term effectiveness of internet-based interventions on all 4 SNAP risk behavior outcomes. All SNAP outcomes were coded so they were in the same direction, with higher scores equating to worse health risk behaviors. Results The inclusion criteria were met by 11 studies: 7 studies measured the effect of an internet-based intervention on nutrition and physical activity; 1 study measured the effect on smoking, nutrition, and physical activity; and 3 studies measured the effect on all SNAP behaviors. Compared with the control group, internet-based interventions achieved an overall significant improvement across all SNAP behaviors in the long term (standardized mean difference –0.12 [improvement as higher scores = worse health risk outcomes], 95% CI –0.19 to –0.05; I2=1.5%, P=.01). The global methodological quality rating was “moderate” for 1 study, while the remaining 10 studies were rated as “weak.” Conclusions Internet-based interventions were found to produce an overall significant improvement across all SNAP behaviors collectively in the long term. Internet-based interventions targeting multiple SNAP behaviors have the potential to maximize long-term improvements to preventive health outcomes.

Overconfidence bias and stock market volatility in Ghana: testing the rationality of investors in the Covid-19 era

PurposeThe paper tests the overconfidence bias and volatility on the Ghana Stock Exchange (GSE) during the pre-Covid-19 pandemic and Covid-19 pandemic period.Design/methodology/approachThe study employs pairwise Granger causality to test the presence of overconfidence bias on the Ghana stock market as well as GARCH (1,1) and GJR-GARCH (1, 1) models to understand whether overconfidence bias contributed to volatility during pre-Covid-19 pandemic and Covid-19 pandemic period. The pre-Covid-19 pandemic period spans from January, 2019 to December, 2019, and Covid-19 pandemic period spans from January, 2020 to December, 2020.FindingsThe paper finds a unidirectional Granger causality running from weekly market returns to weekly trading volume during the Covid-19 pandemic period. These results indicate the presence of overconfidence bias on the Ghana stock market during the Covid-19 pandemic period. Finally, the conditional variance estimation results showed that excessive trading of overconfident market players significantly contributes to the weekly volatility observed during the Covid-19 pandemic period.Research limitations/implicationsThe empirical findings demonstrate that market participants on the GSE exhibit conditional irrationality in their investment decisions during the Covid-19 pandemic period. This implies investors overreact to private information and underreact to available public information and as a result become overconfident in their investment decisions.Practical implicationsFindings from this paper show that there is evidence of overconfidence bias among market players on the GSE. Therefore, investors, financial advisors and other market players should be educated on overconfidence bias and its negative effect on their investment decisions so as to minimize it, especially during the pandemic period.Originality/valueThis study is a maiden one that underscores investors’ overconfidence bias in the wake of a pandemic in the Ghanaian stock market. It is a precursor to the overconfidence bias discourse and encourages the testing of other behavioral biases aside what is understudied during the Covid-19 pandemic period in Ghana.

Neural Network-Based Electric Vehicle Range Prediction for Smart Charging Optimization

Range prediction is a standard feature in most modern road vehicles, allowing drivers to make informed decisions about when to refuel. Most vehicles make range predictions through data- or model-driven means, monitoring the average fuel consumption rate or using a tuned vehicle model to predict fuel consumption. The uncertainty of future driving conditions makes the range prediction problem challenging, particularly for less pervasive battery electric vehicles (BEV). Most contemporary machine learning-based methods attempt to forecast the battery SOC discharge profile to predict vehicle range. In this work, we propose a novel approach using two recurrent neural networks (RNNs) to predict the remaining range of BEVs and the minimum charge required to safely complete a trip. Each RNN has two outputs which can be used for statistical analysis to account for uncertainties; the first loss function leads to mean and variance estimation (MVE), while the second results in bounded interval estimation (BIE). These outputs of the proposed RNNs are then used to predict the probability of a vehicle completing a given trip without charging, or if charging is needed, the remaining range and minimum charging required to finish the trip with high probability. Training data was generated using a low-order physics model to estimate vehicle energy consumption from historical drive cycle data collected from medium-duty last-mile delivery vehicles. The proposed method demonstrated high accuracy in the presence of day-to-day route variability, with the root-mean-square error (RMSE) below 6% for both RNN models.

COVID-19 Vaccine Intentions Within the Medicare Population

COVID-19 vaccine intentions by older adults reflect individual care seeking behavior and medical system trust and broader systemic cultural shifts related to vaccine hesitancy. The purpose of this paper is to examine the October wave of the rapid response panel survey fielded by the Centers for Medicare and Medicaid Services (CMS) to track and monitor the effects of the pandemic within the Medicare population. With a sample size of 9686 Medicare beneficiaries, the calculated statistics use replicate weights to adjust for the complex survey sample design and balanced repeated replication using Fay’s adjustment of 0.3 for variance estimation. When asked about the likelihood of getting the COVID-19 vaccine if one were available, 58 percent of the Medicare population definitely or probably intended to get the vaccine, 16 percent expressed they would probably or definitely not, and 26 percent were not sure. Black or Hispanic Medicare beneficiaries were significantly more likely to express they would probably not or definitely not get the vaccine than White, non-Hispanic Medicare beneficiaries. Distrust of what government says about the vaccine and concern about the safety or side effects were the most common reasons for not intending to get the vaccine. Those expressing intentions to not get the COVID-19 vaccine in the October 2020 survey wave were more likely to lack access to the internet, which is a potential systematic barrier if they changed their intentions following the FDA approvals of the COVID-19 vaccines and more information became available in the winter and spring of 2021.

Estimation of Domain Means from Business Surveys in the Presence of Stratum Jumpers and Nonresponse

Misclassified frame records (also called stratum jumpers) and low response rates are characteristic for business surveys. In the context of estimation of the domain parameters, jumpers may contribute to extreme variation in sample weights and skewed sampling distributions of the estimators, especially for domains with a small number of observations. There is limited literature about the extent to which these problems may affect the performance of the ratio estimators with nonresponse-adjusted weights. To address this gap, we designed a simulation study to explore the properties of the Horvitz-Thompson type ratio estimators, with and without smoothing of the weights, under different scenarios. The ratio estimator with propensity-adjusted weights showed satisfactory performance in all scenarios with a high response rate. For scenarios with a low response rate, the performance of this estimator improved with an increase in the proportion of jumpers in the domain. The smoothed estimators that we studied performed well in scenarios with non-informative weights, but can become markedly biased when the weights are informative, irrespective of response rate. We also studied the performance of the ’doubled half’ bootstrap method for variance estimation. We illustrated an application of the methods in a real business survey.

The Effects of Exercise on Cognitive Function in Older Adults With Different Types of Dementia: A Meta-Analysis

Combating dementia is a public health priority, and exercise training is one promising strategy for dementia prevention. However, its efficacy in promoting cognitive outcomes in different types of dementia remains unknown. We conducted a systematic review (N = 27) and meta-analysis (N = 24) of randomized controlled trials with cognitive function as a primary or secondary outcome. We aimed to assess the effect of exercise interventions on the cognitive function of older adults (>60 years) diagnosed with different types of dementia. We synthesized data from 2,441 older adults with dementia. Eleven trials included older adults with multiple types of dementia, eight with Alzheimer's disease, six with unspecified types of dementia, and two with vascular cognitive impairment. We performed random-effects models using robust variance estimation (RVE) and tested potential moderators using the approximate Hotelling-Zhang test (HTZ). Results suggest a small effect of exercise on cognitive function for all-cause dementia (g = 0.18; 95% CI: 0.04, 0.33; p = 0.016); however, the effects did not differ by type of dementia. Moderation analyses showed that trials that did not specify participants' severity of dementia, applied individual-level randomization, and had higher intervention adherence demonstrated larger exercise effects on cognitive function for all-cause dementia. We conclude that exercise promotes small improvements in the cognitive function of older adults with all-cause dementia. More research including different types of dementia is needed if we hope to determine the precise effects of exercise for each type of dementia.