For statistical inference we used bootstrapping (100,000 resamples) to estimate whether 0 was within or outside the 95% bias-corrected and accelerated bootstrap confidence intervals.
We used percentile bootstrapping with 1,000 resamples to estimate statistical significance for correlations based on professor and relationship scores as these scores violated the independence of observations assumption.
The reliability and validity of the measurement model were assessed by confirmatory factor analysis (CFA) [ 63- 65] by using the SmartPLS 2.0 software with the bootstrap resampling method (1000 resamples) to estimate the parameters of the research model.
We used a maximum-likelihood approach to infer the model parameters and a nonparametric bootstrap resampling approach to estimate confidence intervals and confidence regions.
Furthermore, the potential effect of the number of chromosomes on genome heterogeneity was assessed for both organisms, with the use of a bootstrap resampling procedure to estimate the test statistics corresponding to the contingency table "Chromosome × chromosome region".
Given the computational burden, we applied the bootstrap resampling technique to estimate the standard deviation and the confidence interval (CI) for the optimal cut-point for some selected scenarios with sample size N = 100 or N = 150.
For each of 11 markers that did not support the red-fruited clade based on MEGA results, consensus trees were generated by Phylip ver. 3.69 using Seqboot to produce 100 datasets by bootstrap resampling, Dnadist to estimate genetic distances using the Jukes-Cantor method, Neighbor to produce unrooted NJ trees and Consense to compute a consensus tree by the majority-rule consensus tree method [ 67].
The article applies the classic t test to check whether the specific input combination is feasible, and whether any constraints are binding; next, it applies bootstrapping (resampling) to test the estimated gradients in the KKT conditions.
Testing for mediation involves estimating whether the path coefficient between health satisfaction and each mediator multiplied by the path coefficient between each mediator and the global BREF score is significantly different from 0. Bootstrap resampling was employed to estimate the confidence interval on the path products for each of the multiple mediators simultaneously [ 45, 47].
To compensate for this, we perform model averaging through non-parametric bootstrapping (resampling with replacement) to estimate the confidence in various network features learned [ 24].
