We consider three different cases for the amount of noise: a reference case with no noise (ω = 0), a medium-variance scenario with a medium amount of noise [ω = (0.5)], and a high-variance scenario with a high level of noise [ω = (1.5)].
For the three types of noises (Gaussian, salt and pepper, and speckle), we can observe that the proposed method is fairly robust against noises with medium variances, whereas for high variance noises the method presents acceptable performance since the BCR values is greater than 93%% for the majority of experiments.
Cohen [ 25] operationally defines effect sizes as follows: a small effect size is one that accounts for 2% (0.02) of the variance, a medium effect size accounts for 13%, (0.13) and a large effect size accounts for 26% (0.26).
Conservatively, that assumes a medium level of between-study variance, a statistical power of 0.80 and a significance level of α < 0.05.
In the second step, response efficacy, self-efficacy and perceived advantages of the safe behaviour were also significantly associated with using a stair gate and together explained 23% of the variance, indicating a medium effect size.
As seen in Table 3, with regard to MCS, a total of 17% of this variance was accounted for (R = 0.46, R = 0.17; F(13,1085) = 17.79; p < 0.001), with gender accounting for a small (f = .01) but significant 1% of variance (p < 0.05) and cognitive emotion regulation strategies were found to account for a medium (f = .19) and significant variance 16% (p < 0.01) in MCS.
However, in this study, we extracted 6 factors explaining 40.3 % variance, which is a medium result compared to other studies, for example: Hu et al. = 2 factors: 20 %; Arkkola et al. = 7 factors: 29.5 %; Shi et al. = 4 factors: 28.5 %; Lau et al. = 2 factors: 17.1 % [ 69– 73].
Analysis of the experimental data proceeded with the comparison of the fit of the local Zero-Sum-Multinomial (lZSM) and meta Zero-Sum-Multinomial (mZSM) distributions to the Control, low-variance (LV), medium-variance (MV) and high-variance (HV) conditions, separately.
However, the medium-variance case overestimates τ2 somewhat (estimated value is around 0.7, whereas the true value is 0.5).
In the realistic medium-variance case, the lines cross around pcut = 0.2 for the M8 scenario, and pcut = 0.4 for the M20 case (not shown).
The false-positive and false-negative curves cross at pcut = 0.2 for both M8 and M20 (not shown) scenarios in the medium-variance case.
