# Table 3 Logistic regression with Bayesian variable selection explaining the adoption of SPBs following cyber abuse victimisation accounting for the method of abuse experienced by the victim ($$N = 746$$). Check marks denote coefficients included in the models

Coefficient $$\beta$$ SD 2.5% 97.5% Pr($$\beta$$ $$\ne$$ 0) Modela
HPDb HPDb %c 1 2 3 4 5
Constant − 2.85 0.45 − 3.77 − 1.98 100 $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$
Method 5 1.11 0.21 0.69 1.52 100 $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$
Impact 1.57 0.18 1.21 1.92 100 $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$
OVR − 0.81 0.23 − 1.24 − 0.38 99 $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$
Age 0.03 0.01 0.01 0.05 94 $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$ $$\checkmark$$
Method 4 0.18 0.27 0.00 0.77 39   $$\checkmark$$   $$\checkmark$$
Gender 0.09 0.17 0.00 0.51 29    $$\checkmark$$ $$\checkmark$$
Method 1 0.04 0.12 0.00 0.39 17      $$\checkmark$$
Method 3 − 0.02 0.09 − 0.30 0.04 12
Employment 0.01 0.08 − 0.04 0.21 9
Method 2 − 0.01 0.07 − 0.22 0.00 9
Race − 0.01 0.06 − 0.18 0.02 8
Posterior Probability of the model (%) 24.3 13.2 9.1 6.7 4.9
1. aBest 5 models (cumulative posterior probability = 59.1%)
2. bHighest Posterior Density
3. cProbability of inclusion