Fig 1.
Photograph of the arrangement of the 15 boxes before the first session.
Participants were asked to judge the weight of the mug, shown in the photo, before the experiment as an introduction to the lifting task. The experimenter sat behind the computer screen at the back and recorded the responses from the participant as they estimated the weight of the boxes one by one, starting with the medium box at the top right of the photo.
Table 1.
Weight and density of the boxes used as stimuli in the experiment.
Fig 2.
Photograph of the inside of the boxes and the added weights.
The photograph illustrates examples of how the box weights were adjusted in each of the three different sizes used in our experiment. Single 10g and 5g weights can be seen on top of the lid at the top of the photo.
Fig 3.
Average judgment of each box’s weight.
Points illustrate average point estimates and the colored fields boundaries are determined by the average upper boundary judgments and lower boundary judgments of the self-selected intervals. The colored fields become darker with each consecutive split of the intervals. Hence, the lightest color indicates the average width of the first interval (widest) and the darkest color indicates the average width of the fourth and final interval (most narrow). The aggregate results are stable across sessions, shown in the appendix with plots similar to Fig 3, but for each separate session. Importantly, there was great individual variability in the judgments, shown with figures for each participant, in the appendix.
Table 2.
Descriptive statistics of SWI-cases per participant for the interval conditions first and final mid, as well as the point judgments.
Table 3.
Descriptive statistics of weight judgments of a smaller box proportional deviation from the corresponding box that was one size larger.
Fig 4.
Average SWI-factor depending on weight and judgment method.
Separate lines for SWI-factor for small/medium (solid) and medium/large (dotted) comparisons. Whiskers indicate 95%-CI’s.
Fig 5.
Average SWI-factor depending on weight and judgment method for each judgment session.
Separate lines for SWI-factor for small/medium (solid) and medium/large (dotted) comparisons.
Fig 6.
Slope estimates for log of point judgments and log of mid4 in grams.
Main effects: the log of the true box weight, the type of judgment estimate (point vs mid4), and size (small, medium, large). The model also contains first-order interaction effects, and random effects of, log of true weight, type of judgment estimate, and size. The numerical model estimates can be found in S1 Table.
Fig 7.
Relative standard deviation of box judgments depending on measure and size of the box.
Whiskers indicate 95%-CI.