Skip to main content
. 2019 Dec 23;8(4):754–769. doi: 10.1556/2006.8.2019.66

Table 4.

Results of the moderated mediation analysis (Model 63 – see Figure 4a) testing the effects of run length (IV), Δ probability (mediator), final color and gambling status (moderators), BSSS, BIS-11, BSSS × gambling status, and BIS-11 × gambling status (covariates) on alternation choice

Logistic regression summary −2 LL Model LL p Cox–Snell Nagelkerke
357.1832 74.3422 <.001 0.2073 0.2800
Model B z p CI
Constant −0.5034 −1.5491 .1214 [−1.1404, 0.1335]
Δ Probability 0.0571 3.6187 .0003 [0.0262, 0.0880]
Run length 0.8928 2.0061 .0448 [0.0205, 1.7651]
Δ Probability × final color −0.1216 −5.5521 .0000 [−0.1645, −0.0787]
Final color 1.6794 4.5850 .0000 [0.9615, 2.3974]
Run length × final color −1.1975 −2.3038 .0212 [−2.2164, −0.1787]
Gambling status −0.0625 −0.1606 .8724 [−0.8249, 0.6999]
Run length × gambling status 0.0257 0.0481 .9616 [−1.0205, 1.0719]
BSSS −0.1030 −0.3027 .7621 [−0.0773, 0.0566]
BIS-11 0.0007 0.0327 .9739 [−0.0408, 0.0422]
BSSS × gambling status −0.0156 −0.3278 .7431 [−0.1091, 0.0778]
BIS-11 × gambling status −0.0196 −0.6978 .4853 [−0.0745, 0.454]
Conditional direct effect of run length on alternation choice at values of the moderator(s)
Final color Gambling status Effect p CI
Red Non-problem gamblers .8928 .0457 [0.0205, 1.7651]
Red Problem gamblers .9185 .0391 [0.0496, 1.7875]
Black Non-problem gamblers −.3047 .5074 [−1.2056, 0.5962]
Black Problem gamblers −.2790 .5634 [−1.2255, 0.6674]
Conditional indirect effect(s) of run length through Δ probability (mediator) on alternation choice at values of moderator(s)
Final color Gambling status Effect CI
Red Non-problem gamblers 0.4673 [0.0488, 1.3982]
Red Problem gamblers 0.3175 [−0.0674, 1.1556]
Black Non-problem gamblers 0.1270 [−0.3476, 0.6961]
Black Problem gamblers 0.2962 [−0.1681, 0.8525]

Note. BSSS: Brief Sensation Seeking Scale; BIS: Barratt Impulsiveness Scale 11; CI: confidence interval; LL: log likelihood.