Computer logs are often analyzed with the following techniques to investigate specific hypotheses about the data:
- Histogram of how many messages are sent each day. This shows the ebb and flow of interaction during a semester, for instance.
- Histogram of how many threads of different lengths. This is an indicator of the depth of discussions.
- Coding of messages (or sentences or complet statements). This might be used to compare a control group to an experimental condition. A well-defined set of coding categories must be defined, e.g., for messages that are off-task, social, new ideas, responses, etc. Coding is usually done by at least two people independently and their results are compared with an inter-rater reliability ratio.
- Social network analysis maps the number of messages (e.g., email) from each subject to each other, indicating who are central and who are socially isolated interactors. This might be used to show differential influence of central vs isolated actors on each other.