Post-hoc analyses of this interaction revealed a marginally significant influence of the QSS slope trend estimates for the abstract ( emtrend = 0.84, SE = 0.42, t = 1.99, p = .050) but not for the realistic ( emtrend = −0.33, SE = 0.42, t = −0.78, p = .439) 3D landmark visualization conditions.
Additionally, the results revealed that the QSS score was a significant predictor ( β = 0.41, SE = 0.16, z = 2.66, p = .008), while the interaction with the landmark visualization condition approached significance ( β = −0.28, SE = 0.14, z = −1.94, p = .052) for the reconstruction of the landmarks sequence.
The close-to-significant interaction in the main model, the post-hoc analyses on QSS slope trend estimates, and the visual inspection of Figure 8 show that realistic landmark visualization is most helpful in landmark sequence memory for participants with low spatial abilities, and this difference between the conditions diminishes with increased spatial abilities. 3.3.3.