In case of percent sulfate removal, variable A (pH) demonstrated negative relationship whereas B (reaction time) and C (FeCl3 dosage) gave us an idea about its positive relationship.
In case of percent sulfate removal, variable A (pH) demonstrated negative relationship whereas B (reaction time) and C (FeCl3 dosage) were found to be positively related.
Multivariable logistic regression modelling with backward removal of variables identified explanatory variables (i.e., socio-demographic, clinical and work status variables outlines above) that were significantly associated with work participation in the cancer and general population groups.
To investigate which variables were significantly correlated with the dependent variables, we used a backwards-removal procedure with all variables initially included in the model, and then sequentially removed variables with significance levels > 0.1.
Backward selection was performed on the initial multivariate model for the sequential removal of variables: in each step, the variable with the largest p value was removed until the model contained only statistically significant variables (two-sided p < 0.05).
For variable selection, we used backward stepwise removal of variables based on likelihood-ratio judgments.
The influence of each variable will be assessed using Wald tests, with stepwise removal of variables with p≥0.05.
The order of variable selection was determined by the χ statistic for each potential variable and the forward selection step could be followed by removal of variables in one or more backward elimination steps.
The models were refined by successive removal of variables showing no statistically significant contribution to the fit of either model, other than the selected cooking fuel variables, which were retained in the models irrespective of their statistical significance.
Eight weeks after frame removal, baseline variables (age, BMI, Smoking, Charlson comorbidity score, infection and high-/low-energy trauma) show no significant influence on patient-reported outcome (EQ5D-5L; P ≥ 0.26, Table 4).
Furthermore, the removal of variables not associated with population history strengthened the correlations between geographic and morphological distances considerably, especially for Model 3 (bipartite Asian origin).
