Abstract:State-Trace Analysis (STA) is a methodology for investigating the number of latent variables. Recently, a quantitative STA technique based on conjoint monotonic regression and double bootstrap method (STA-CMR) has been proposed. More discussion is needed on the type I error and the statistical power of this technique, as it adopts null hypothesis significance testing (NHST) to draw statistical inference. Because the results of STA are comparable with analysis of variance (ANOVA) in a three-factor experiment with linearity assumption, it is necessary to compare STA-CMR with ANOVA accordingly. This study investigated the type I error and the statistical power of STA-CMR and ANOVA in specific linear and nonlinear models using simulated data. Results demonstrated that both techniques were effective in the linear models, where ANOVA had a greater statistical power and STA-CMR had a more rigorous control of type I error. In the nonlinear models, although STA-CMR worked just as well as in the linear models, ANOVA completely lost its effectiveness. Besides, we found that the estimated type I error rate of STA-CMR was always smaller than the preset significance level in both linear and non-linear models. We suggest that the suppressed type I error rate may be caused by the bootstrap procedure, but the exact causes need further investigation. In conclusion, despite the suppressed type I error rate, STA-CMR can be a useful tool for determining the number of latent variables, particularly in non-linear models.

The Applicability of the Conventional Statistic Analysis Methods in Minimization

The Type Ⅰ Error and Power of the Conventional Statistic Analysis Methods in Minimization

A note on extending Scheffé's modified multiple-comparison procedure to other analysis situations

Statistical Analysis of CCM.M-K1 International Comparison Based on Monte Carlo Method.

Nonparametric Response-Adaptive Randomization for Continuous Responses

Shrinkage-based regularization tests for high-dimensional data with application to gene set analysis

Inference on Average Treatment Effect under Minimization and Other Covariate-Adaptive Randomization Methods

Power Analysis and Sample Size Planning in ANCOVA Designs

MODE: Minimax Optimal Deterministic Experiments for Causal Inference in the Presence of Covariates

The method of minimization for random allocation in clinical trials: introduction and its SAS program

The Balance Comparison among Simple Randomization, Stratified Blocked Randomization and Minimization

Statistical Inference for Covariate-Adaptive Randomization Procedures

Comparison of Type I Error and Statistical Power Between State Trace Analysis and Analysis of Variance

Testing for Treatment Effect in Covariate-Adaptive Randomized Clinical Trials with Generalized Linear Models and Omitted Covariates

An Empirical Comparison of Parametric and Permutation Tests for Regression Analysis of Randomized Experiments

Exploring Statistical Methods in Clinical Trials When Some of the Patients are Treated against the Randomization

Asymptotic properties of covariate-adaptive randomization

Testing two variances for superiority/non-inferiority and equivalence: Using the exhaustion algorithm for sample size allocation with cost

Reducing multivariate independence testing to two bivariate means comparisons

A Viable Method for Goodness-of-fit Test in Maximum Likelihood Fit

An efficient asymptotic approach for testing monotone proportions assuming an underlying logit based order dose-response model