For all pairs (X, Y) of methods, the tested null hypothesis was H0: the mean error provided by "method X" is lower than the mean error provided by "method Y", against the alternative hypothesis Ha: mean errors of "method X" and "Y" are equal.
In practice, method 4 is well suited to include many controls in the first phase, due to its use of register data, but we can conclude that the empiric standard errors of method 4 is rather constant when increasing the number of first-phase subjects compared to the initial scenario 1.
Even at a range of 3500 m, the averaged horizontal/depth/aspect errors of Method-I are only 0.5 m, 25 m, and 0.5 rad (28.7°), compared to the Method-IV errors of 0.9 m, 45 m, and 1.1 rad (63.1°).
Thus, by analyzing histogram plot, minimum and maximum errors of methods are attained and the error with the most distribution (the most probable) is also obtainable. Figure 13 shows the histograms of the errors attained by six methods.
The empirical standard errors of methods are given in Figure 5. Complete data analysis has the lowest standard errors, while complete cases and parametric imputation also tend to be low.
Second, the excellent precision (a measure of random error) of Method II in this example ignores the above-mentioned systematic error in ΣS0 determinations.
The measurement results showed that: the relative error of method of LFMCL distance measurement based on monostatic system was less than 0.1%.
Error of method: for intra-examiner reliability, measurements were compared via coefficient of reliability and kappa statistics.
The following tests were used: Error of method: for intra-examiner reliability, measurements were compared via coefficient of reliability and kappa statistics.
Future research should report reliability, validity and measurement error of methods, apply transparent data analysis approaches and include defensible sampling methods to ensure that the findings are generalizable.
Within this conceptual framework, it has been also correctly observed that the prediction error of methods based on BIA of body compartments is the sum of many errors, namely the impedance measurement error, the regression error, the intrinsic error of the reference method, the electric volume model error, and the biological variability among subjects [ 15].
