highly significantp < 10 −5
This difference between groups at recruitment was highly significant (Kruskal–Wallis test, χ 2 (2) = 25.2; p < 10 −5 ), and represents a failure of randomisation due to our limited sampling.
This difference between groups at recruitment was highly significant (Kruskal–Wallis test, χ 2 (2) = 25.2; p < 10 −5 ), and represents a failure of randomisation due to our limited sampling.
However, even when members of the 2D control C-group are excluded, there is still a fairly significant Spearman correlation between S3D exposure and change in AC/A ratio ( N = 83, ρ = −0.25, p = 0.02).
There were marginally significant interactions between visit and age (visit*age-group: F (3,71) = 2.89, p = 0.041) and visit, age and TV-group (visit*age-group*TV-group: F (6,71) = 2.30, p = 0.044).
The main effect of TV-group approached significance ( F (2,105) = 2.72, p = 0.07; F (2,99) = 3.47; p = 0.035 without the 6 individuals who acquired correction), with the control group C having higher acuity than the other two groups.