Description (median, (25 75% percentiles)); comparisons using Mann–Whitney and chi-squared tests; multivariate analysis using a step-by-step logistic regression model.
Searchlight approaches (Kriegeskorte et al., 2006) are multivariate analysis methods that step through all voxels of interest in sequence (in our case, all voxels in a given network) and examine voxel values in a 'searchlight' region of interest surrounding the current voxel.
For multivariate analysis, a two-step pathway analysis was used to provide interpretable results that could accommodate variables with linear or curvilinear relationships.
We conducted our multivariate analysis in two steps, first adjusting for age and gender, and then adjusting for the additional covariates of ethnicity, marital status, employment and tenancy, such that the relative contribution of these two sets of covariates to the strength of the association between CMI and IPV could be examined.
As a final step, multivariate analysis is conducted to test the research hypothesis.
In a final step, multivariate analysis using a regression model on EFS, progressively removing variables in case of P>0.05, was performed, according to a Cox model.
In the multivariate analysis, the Cox forward step wise model revealed that the performance of an invasive diagnostic procedure outside the tumour centre was an independent prognostic factor for local recurrence and metastasis while inadequate surgical margins was an independent adverse prognostic factor for local recurrence and tumour-related death.
The main experimental factors affecting the complexation and the extraction of metals (pH, PAN concentration, salt addition and extractant solvent and disperser solvent volume) were optimized using a multivariate analysis consisting of two steps: a Plackett-Burman design followed by a Circumscribed Central Composite Design (CCCD).
We constructed a two-step multivariate analysis for evaluation of independent associations with GLS.
Significant variables on the univariate analysis were analyzed by multivariate analysis using a forward step-wise logistic regression model.
