Dataset Information

Multivariate Higher-Order IRT Model and MCMC Algorithm for Linking Individual Participant Data From Multiple Studies.

ABSTRACT: Many clinical and psychological constructs are conceptualized to have multivariate higher-order constructs that give rise to multidimensional lower-order traits. Although recent measurement models and computing algorithms can accommodate item response data with a higher-order structure, there are few measurement models and computing techniques that can be employed in the context of complex research synthesis, such as meta-analysis of individual participant data or integrative data analysis. The current study was aimed at modeling complex item responses that can arise when underlying domain-specific, lower-order traits are hierarchically related to multiple higher-order traits for individual participant data from multiple studies. We formulated a multi-group, multivariate higher-order item response theory (HO-IRT) model from a Bayesian perspective and developed a new Markov chain Monte Carlo (MCMC) algorithm to simultaneously estimate the (a) structural parameters of the first- and second-order latent traits across multiple groups and (b) item parameters of the model. Results from a simulation study support the feasibility of the MCMC algorithm. From the analysis of real data, we found that a bivariate HO-IRT model with different correlation/covariance structures for different studies fit the data best, compared to a univariate HO-IRT model or other alternate models with unreasonable assumptions (i.e., the same means and covariances across studies). Although more work is needed to further develop the method and to disseminate it, the multi-group multivariate HO-IRT model holds promise to derive a common metric for individual participant data from multiple studies in research synthesis studies for robust inference and for new discoveries.

SUBMITTER: Mun EY

PROVIDER: S-EPMC6582193 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multivariate Higher-Order IRT Model and MCMC Algorithm for Linking Individual Participant Data From Multiple Studies.

Mun Eun-Young EY Huo Yan Y White Helene R HR Suzuki Sumihiro S de la Torre Jimmy J

Frontiers in psychology 20190612

Many clinical and psychological constructs are conceptualized to have multivariate higher-order constructs that give rise to multidimensional lower-order traits. Although recent measurement models and computing algorithms can accommodate item response data with a higher-order structure, there are few measurement models and computing techniques that can be employed in the context of complex research synthesis, such as meta-analysis of individual participant data or integrative data analysis. The ...[more]

PMID: 31244727

Similar Datasets

Project description:BACKGROUND:Data repositories have the potential to play an important role in the effective and safe sharing of individual-participant data (IPD) from clinical studies. We analysed the current landscape of data repositories to create a detailed description of available repositories and assess their suitability for hosting data from clinical studies, from the perspective of the clinical researcher. METHODS:We assessed repositories that enable storage, sharing, discoverability, re-use of the IPD and associated documents from clinical studies using a pre-defined set of 34 items and publicly available information from April to June 2018. For this purpose, we developed an indicator set to capture the maturity of the repositories' procedures and their suitability for the hosting of IPD. The indicators cover guidelines for data upload and data de-identification, data quality controls, contracts for upload and storage, flexibility of access, application of identifiers, availability of metadata, and long-term preservation. RESULTS:We analysed 25 repositories, from an initial set of 55 identified as possibly relevant. Half of the included repositories were generic, i.e. not limited to a specific disease or clinical area and 13 were launched in the last 8 years. The sample was extremely heterogeneous and included repositories developed by research funders, infrastructures, universities, and editors. All but three repositories do not apply a fee for uploading, storage or access to data. None of the repositories completely demonstrated all the items included in the indicator set, but three repositories (Dryad, Drum, EASY) met - fully or partially - all items. Flexibility of data-access modalities appears to be limited, being lacking in half of the repositories. CONCLUSIONS:Our evaluation, though often hampered by the lack of sufficient information, can help researchers to find a suitable repository for their datasets. Some repositories are more mature because of their support for clinical dataset preparation, contractual agreements, metadata and identifiers, different modalities of access, and long-term preservation of data. Further work is now required to achieve a more robust and accurate system for evaluation, which in turn may encourage the sharing of clinical study data. TRIAL REGISTRATION:Study protocol available at https://zenodo.org/record/1438261#.W64kW9Egrcs .

Project description:AimsTo investigate cross-sectional and longitudinal associations between personality and smoking, and test whether socio-demographic factors modify these associations.DesignCross-sectional and longitudinal individual-participant meta-analysis.SettingNine cohort studies from Australia, Germany, the United Kingdom and the United States.ParticipantsA total of 79 757 men and women (mean age = 50.8 years).MeasurementsPersonality traits of the five-factor model (extraversion, neuroticism, agreeableness, conscientiousness and openness to experience) were used as exposures. Outcomes were current smoking status (current smoker, ex-smoker and never smoker), smoking initiation, smoking relapse and smoking cessation. Associations between personality and smoking were modelled using logistic and multinomial logistic regression, and study-specific findings were combined using random-effect meta-analysis.FindingsCurrent smoking was associated with higher extraversion [odds ratio per 1 standard deviation increase in the score: 1.16; 95% confidence interval (CI) = 1.08-1.24], higher neuroticism (1.19; 95% CI = 1.13-1.26) and lower conscientiousness (95% CI = 0.88; 0.83-0.94). Among non-smokers, smoking initiation during the follow-up period was predicted prospectively by higher extraversion (1.22; 95% CI = 1.04-1.43) and lower conscientiousness (0.80; 95% CI = 0.68-0.93), whereas higher neuroticism (1.16; 95% CI = 1.04-1.30) predicted smoking relapse among ex-smokers. Among smokers, smoking cessation was negatively associated with neuroticism (0.91; 95% CI = 0.87-0.96). Socio-demographic variables did not appear to modify the associations between personality and smoking.ConclusionsAdult smokers have higher extraversion, higher neuroticism and lower conscientiousness personality scores than non-smokers. Initiation into smoking is associated positively with higher extraversion and lower conscientiousness, while relapse to smoking among ex-smokers is associated with higher neuroticism.

Dataset Information

Multivariate Higher-Order IRT Model and MCMC Algorithm for Linking Individual Participant Data From Multiple Studies.

Publications

Multivariate Higher-Order IRT Model and MCMC Algorithm for Linking Individual Participant Data From Multiple Studies.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets