Fig 1.
Stimuli and their physical parameters.
(A) All stimuli used in Experiment 1 and 2. Screw caps were removable and could either be plain or attached with a string and a wooden bead. Small set objects are in the back, big set objects in the front row and equal-density objects in in-between. (B) Physical parameters of stimuli in (A). Small and big set objects were constant in volume and thus their density increased with increasing mass. This increase is more prominent in the small set. Objects from the equal density set were constant in density. The 230 g stimulus of the big set also had a density of 0.39 g/cm3 and was also included in the analysis of the equal density set.
Fig 2.
Heaviness estimates and model fits for the blindfolded (A), the poor (B), the medium (C) and the full vision condition (D). Estimates are averaged over individual heaviness scores with error bars denoting 95% confidence intervals. Lines represent the model fit.
Fig 3.
Density weights in Experiment 1.
Fitted density weights for the poor, the medium and the full vision condition. Weights are averaged over individual participants with error bars denoting the 95% confidence interval of between-participant variability. * p < 0.05.
Fig 4.
Heaviness estimates averaged over individual heaviness scores with error bars denoting 95% confidence intervals of between- participant variability and lines being the model fit. (A, B, C): Heaviness estimates for different density sets as a function of mass with a different panel for every grip type (A: string, B: precision grip, C: enclosure).
Fig 5.
Density weights in Experiment 2.
Fitted density weights for the precision grip and the enclosure condition. Weights are averaged over individual participants with error bars denoting the 95% confidence interval of between-participant variability.
Fig 6.
(A) Average PSE values for the small (purple) and the big set (red) with 95% confidence intervals. (B) Observed against predicted illusion strength for both grip types. Lighter gray squares indicate the enclosure condition. Open symbols denote individuals. Diagonal error bars for the aggregated data mark the error of the differences between observed and predicted values and have to be compared to the identity line.