efficient was used to test the correlation between these two variables). There were a number of event categories that were all associated with Speaker, Hearer and Mutual gaze that significantly correlate with problem solving success. These included interactions related to progressing the investigation (Z = 3 P<0.004 Speaker and Hearer, Z = 4 P<0.001 Mutual), feedback to partner (Z = 2 P<0.024) and changes associated with emotions negative (Z= 3 P<0.06). Here we see that planning activities were associated with eye gaze. However, we have labelled this category not as planning, but as strategies employed to progress the investigation.
With respect to Hearer and Mutual gaze, the three key areas of interaction which were common to both concern discussions about firstly the pucks (Z = 2 P<0.017 Hearer gaze, Z = 2 P<0.045 Mutual gaze), secondly around dialogue which progressed understanding (Z = 3 P<0.01 Hearer gaze, Z = 2 P<0.01 Mutual gaze) and thirdly, exchanges classified as positive emotional support which were only associated with problem solving success for Mutual gaze (Z = 3 P<0.004).
Gesturing was always prompted through discussion of the movement of the pucks. It was practically impossible for the partners to exchange ideas about the collisions without moving their hands and arms in some way. (Z = 3 P<0.001 Speaker gesture, Z = 1 P<0.001 Hearer gesture and Z = 4 P<0.0001 Mutual gesture). Another type of movement of the hands occurred for all conditions when the pairs were providing each other with Mutual support (Z = 2 P<0.02 Speaker gesture, Z = 4 P<0.0001 Hearer gesture and Z = 4 P<0.001 Mutual gesture). When the partners were approaching difficult issues, such as disagreement or confusion, admitting something was wrong or one partner did not understand what the other one was doing, again there was hand movement. These types of actions have been described in our event category system as a negative emotional response and occurred for the Speaker gesture (Z = 3 P<0.006), Hearer gesture (Z = 3 P<0.007) and Mutual gesture (Z = 3 P<0.001). Mutual gesturing also correlated with problem solving success with respect to progressing the investigation (Z = 4 P<0.0001). This was also the case of Hearer gesture associated with strategies to progress the investigation (Z = 4 P<0.0001).
There were significant differences with respect to gender and gaze. In fact it was the Speaker in the male pairs who looked at his partner more when talking about the motion of the ice pucks than the other gender groups (F = 4 P<0.05; one way anova of gender vs category of talk with respect to Speaker gaze). This meant in effect that the speaker looked at the hearer while the hearer was looking at the computer screen. It was, however, the females who used a mutual gazing strategy when talking about the ice pucks (F = 4 P<0.05). With respect to Progressing Understanding; it was the female Speakers who gazed significantly more in this condition than the other gender groups (F = 1 P<0.05). They also gazed more than the other groups when discussing the workings of the computer program (F = 3 P<0.01).
The significant differences occurred with the mixed gender pairs who gazed significantly less than the single sex pairs when Progressing the Investigation (F = 3 P<0.05) and also providing positive emotional support (F = 1 P<0.01) to one another. In the mixed pairs, it was the Speaker who did most of the gesturing since there was less Hearer gesturing (F = 4 P<0.05) and Mutual gesturing (F = 2 P<0.02) in this gender condition. With all gender pairings the Speaker controlled most of the problem space and there was no significantly difference by gender for Speaker gesture with respect to the seven event categories. Where differences were found it was the male Hearers who gestured more about the motion of the ice pucks that the mixed pairs (F = 1 P0.015) and Mutual gesturing was more common with the female pairs. This type of mutual interaction was more significantly when the females (F = 1 P<0.05) were planning (F = 2 P<0.001) providing feedback to their partner and also engaged in discussions about the ice pucks (F = 2 P<0.02).
The types of inter subject events that prompt eye gazing and gesture fall into seven main groupings. Two of which were primarily concerned with the subject domain and the workings of the computer program. These have been described as "Motion of Ice Pucks" and the "Computer Program" in our event category system. We have found that gesture associated with talk about the motion of the ice pucks has a significant correlation with problem solving success. It was very difficult for the subjects to discuss collisions without using their hands to