Experiment Draft 1

*Just to note, this is completely a draft, and I am not even fully decided on the topic of the experiment*

The experiment I am thinking about doing right now is something in the realm of measuring conversational involvement, or some specific aspects of conversational involvement within different types of communication. Conversational involvement is a measure of how cognitively and behaviorally engaged participants are in a conversation. It measures specific nonverbal behaviors along five dimensions: immediacy, expressiveness, interaction management, altercentrism, and social anxiety. In prior work on conversational involvement, a number of behaviors are identified to strongly correlate between high and low involvement in conversations. Some of these behaviors include general proxemic attentiveness, forward lean, relaxed laughter, coordinated speech, number of silences and latencies in communication, number of object manipulations, facial animation, vocal warmth, and amount of random body movement. Some of these can be easily quantitatively measured, like silences / pauses, forward lean, amount of relaxed laughter. I want to explore the conversational involvement during different online interactive communications vs. in person communications - and I want to see the effect of having an initial in person conversation with controlled high or low conversational involvement on subsequent online interactive communications.    

An initial draft of the experiment would involve video tapping the short conversations on set topics between two people in person. One person would be the study subject, and the other would be a 'tester'. The 'tester' would vary their conversational involvement between trials from high to low amounts by changing their behavior according to the conversational involvement metrics. The subjects would later be recorded engaging in a subsequent video chat conversations or audio conversations with the tester. The video of them can be analyzed to extract their level of conversational involvement during these sessions. The subjects would also be recorded in another chat with a random tester they had not met before, and the conversational involvement of these two trials would be compared. Also the different conversational involvements of the subjects in the second conversation would be compared against each other depending on whether their initial interaction with the tester had low or high conversational involvement on the part of the tester.