TIMEDUR is an analytic tool in CLAN that allows researchers to display how much time each participant speaks during an interaction, as well as the total length of pauses between speaker turns, and the amount of time participants speak in overlap with one another. This may, for example, be important for charting a general overview of the distribution of time at talk within an interaction. Furthermore, in combination with a tool such as GEM, varying patterns of talk distribution between different phases or activities within an encounter can be compared.
We will look at this, while referring to the Ex12_TIMEDUR .cha file in the workshop folder.
This is what you do:
- In the workshop folder there is a TIMEDUR exercise file. Open it.
- You will see that this transcript looks different from the other versions in the workshop package. Specifically, each speaker line has had a speaker ID entered. This is important, as the program will only output data from those lines where an ID has been added. All others are treated as pauses.
- Open the Command window
- Type timedur
- File in > choose the appropriate file > Done
- Run the command
The output file will give what resembles a spreadsheet, with along the top the list of participant
IDs, and combinations of IDs (FRA/ FRA-ROS / ROS-PPP). The combinations denote overlaps between annotated bullets. Down the side you will note the line numbers of the transcript or sections of transcript you have conducted the analyses on. In the spreadsheet, you will then be able to see, in msec, the amount of time the annotated bullet for each line of transcript denotes, as well as the duration of any overlapping talk.
At the bottom of the spreadsheet, the totals are given, again in msec, for each of the participants in the interaction.
Note: should one wish to export directly into an Excel format, add ‘+d’ into the command line.
Running Timedur for specified sections, phases or activities of data.
The above stages are used for whole files. You may, however, wish to specify particular sections, phases or activities (first half vs second half; greetings; introductions; diagnostic talk; focus-on-form; focus-on-meaning etc). One can do this by piping a particular sequence or collection of sequences to the TIMEDUR tool. For example, should I wish to see the general distribution of talk between DIM and GRI during the ‘laughter_sequence’ sequences addressed in Ex12, then I could run a command that outputs these sections, and runs them through the TIMEDUR tool. The command would look like this:
gem +schallenge_defence +d @ | timedur
Indeed, should I wish to chart the distribution of talk in the ‘other_repair’ sequences within the ‘laughter_sequences’ GEMs, then again, I could pipe these specifically to TIMEDUR thus:
gem +slaughter_sequence +d @ | gem +sother_repair +d | timedur
This will give the timings for only those ‘other_repair’ GEM sequences that are associated with the ‘laughter_sequence’ GEMs.