Two experiments were done to examine the predictability of the items: an event predictability study and a referent predictability study.
In the event predictability study participants viewed the pairs of pictures. They rated the second event in each picture both on how related it was to the first event (using a 0-7 scale) and how predictable it was based on the first event (using a 0-7 scale). The ratings were converted to z-scores for each participant and item, and the means of those z-scores for the critical stimuli items are presented here. There are two lists, which correspond to the lists of linguistic stimuli presented elsewhere.
In the rating study participants viewed the first picture in each pair and were asked to select the referent they thought was most likely to be talked about next. The data presented here represent the average of the participants’ selection of the goal referent for that item. As participants had to choose between the goal and the source referent, the inverse of this score is the likelihood to choose the source.