Dataset Information

Probability-of-Superiority SEM (PS-SEM)-Detecting Probability-Based Multivariate Relationships in Behavioral Research.

ABSTRACT: In behavioral research, exploring bivariate relationships between variables X and Y based on the concept of probability-of-superiority (PS) has received increasing attention. Unlike the conventional, linear-based bivariate relationship (e.g., Pearson's correlation), PS defines that X and Y can be related based on their likelihood-e.g., a student who is above mean in SAT has 63% likelihood of achieving an above-mean college GPA. Despite its increasing attention, the concept of PS is restricted to a simple bivariate scenario (X-Y pair), which hinders the development and application of PS in popular multivariate modeling such as structural equation modeling (SEM). Therefore, this study addresses an empirical-based simulation study that explores the potential of detecting PS-based relationship in SEM, called PS-SEM. The simulation results showed that the proposed PS-SEM method can detect and identify PS-based when data follow PS-based relationships, thereby providing a useful method for researchers to explore PS-based SEM in their studies. Conclusions, implications, and future directions based on the findings are also discussed.

SUBMITTER: Li JC

PROVIDER: S-EPMC6008517 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Probability-of-Superiority SEM (PS-SEM)-Detecting Probability-Based Multivariate Relationships in Behavioral Research.

Li Johnson Ching-Hong JC

Frontiers in psychology 20180613

In behavioral research, exploring bivariate relationships between variables X and Y based on the concept of probability-of-superiority (PS) has received increasing attention. Unlike the conventional, linear-based bivariate relationship (e.g., Pearson's correlation), PS defines that X and Y can be related based on their likelihood-e.g., a student who is above mean in SAT has 63% likelihood of achieving an above-mean college GPA. Despite its increasing attention, the co ...[more]

PMID: 29951012

Similar Datasets

Project description:ContextInverse associations between 25-OH vitamin D levels and cardiovascular morbidity and mortality have been reported.ObjectivesOur goals were to 1) investigate effects of correcting inadequate D status on lipids, 2) determine whether free 25-OH D is better correlated with lipids than total 25-OH D.DesignA randomized, double-blind placebo-controlled trial was performed.SettingParticipants resided in the general community.ParticipantsAdults with inadequate D status were randomized to D3: 14 men, 12 women, age 60 ± 8 years (mean ± SD) or placebo: 12 men, 11 women: 59 ±12 years.InterventionResponses to 12-week oral vitamin D3 titrated (1000-3000 IU/d) to achieve 25-OH D levels ≥25 ng/mL were compared to placebo.Main outcome measuresMeasurements were 25-OH D (tandem mass spectometry), free 25-OH D (direct immunoassay), lipids (directly measured triglyceride, cholesterol, and subfractions; plant sterols and cholesterol synthesis precursors), and safety labs before and after 6 and 12 weeks D3 or placebo. Data were analyzed by repeated measures ANOVA and linear regression.ResultsVitamin D3 was titrated to 1000 IU/d in 15/26 (58%), to 2000 IU/d in 10, and 3000 IU/d in one patient. D3 had no effect on cholesterol or cholesterol subfractions except for trends for decreases in atorvastatin-treated patients (cholesterol, P = .08; low-density lipoprotein [LDL] cholesterol, P = .05). Decreased campesterol concentrations (P = .05) were seen with D3 but not placebo in statin-treated patients. Relationships between total 25-OH D and lipids were not detected, but inverse linear relationships were detected between free 25-OH D and triglycerides (P = .03 for all participants [n = 49], P = .03 in all statin-treated [n = 19], and P = .0009 in atorvastatin-treated [n = 11]), and between free 25-OH D and LDL cholesterol (P = .08 overall, P = .02 in all statin-treated, and P = .03 for atorvastatin-treated), and total cholesterol (P = .09 overall; P = .04 for all statin-treated, and P = .05 for atorvastatin-treated).ConclusionsVitamin D lipid-lowering effects appear limited to statin-treated patients and are likely due to decreased cholesterol absorption. Relationships between lipids and D metabolites were only detected when free 25-OH D was measured, suggesting the superiority of determining free 25-OH D levels compared to total 25-OH vitamin D levels when analyzing biologic responses.

Project description:BackgroundIn medical, social, and behavioral research we often encounter datasets with a multilevel structure and multiple correlated dependent variables. These data are frequently collected from a study population that distinguishes several subpopulations with different (i.e., heterogeneous) effects of an intervention. Despite the frequent occurrence of such data, methods to analyze them are less common and researchers often resort to either ignoring the multilevel and/or heterogeneous structure, analyzing only a single dependent variable, or a combination of these. These analysis strategies are suboptimal: Ignoring multilevel structures inflates Type I error rates, while neglecting the multivariate or heterogeneous structure masks detailed insights.MethodsTo analyze such data comprehensively, the current paper presents a novel Bayesian multilevel multivariate logistic regression model. The clustered structure of multilevel data is taken into account, such that posterior inferences can be made with accurate error rates. Further, the model shares information between different subpopulations in the estimation of average and conditional average multivariate treatment effects. To facilitate interpretation, multivariate logistic regression parameters are transformed to posterior success probabilities and differences between them.ResultsA numerical evaluation compared our framework to less comprehensive alternatives and highlighted the need to model the multilevel structure: Treatment comparisons based on the multilevel model had targeted Type I error rates, while single-level alternatives resulted in inflated Type I errors. Further, the multilevel model was more powerful than a single-level model when the number of clusters was higher. A re-analysis of the Third International Stroke Trial data illustrated how incorporating a multilevel structure, assessing treatment heterogeneity, and combining dependent variables contributed to an in-depth understanding of treatment effects. Further, we demonstrated how Bayes factors can aid in the selection of a suitable model.ConclusionThe method is useful in prediction of treatment effects and decision-making within subpopulations from multiple clusters, while taking advantage of the size of the entire study sample and while properly incorporating the uncertainty in a principled probabilistic manner using the full posterior distribution.

Dataset Information

Probability-of-Superiority SEM (PS-SEM)-Detecting Probability-Based Multivariate Relationships in Behavioral Research.

Publications

Probability-of-Superiority SEM (PS-SEM)-Detecting Probability-Based Multivariate Relationships in Behavioral Research.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets