Exporting data in a time series format

ELAN
1

Hi,

We’re using ELAN to code some videos of caregivers and children sharing books together for 10 minutes. We have a number of tiers we have coded for, and here are a few examples - Tier 1 = turntaking: child, mother; Tier 2 = speech act: descriptive, wh-question etc; Tier 3 = speech act response: ignore child, move on the same topic etc). Alongside these videos, in an excel file we’ve also coded how engaged the child is on a five points scale throughout the 10 minutes clips, at 30 seconds intervals (therefore we have 20 data points for this engagement measure). Now we want to look at how the tiers we have coded relate to how engaged the child is. For example, it may be that the more WH-questions the mother asks, the more engaged the child is.

My question is: is there a way of exporting the ELAN data so that it gives you a csv file with time going down at a consistent sampling rate (e.g. 1 second sampling rate, or a 30 second sampling rate), with counts of particular tiers in another column. For example:

Time (in seconds) Speech Act: wh questions(i.e. how often a speech act occurs)
1 1
2 0
3 0
4 1

Therefore in this example two WH-questions occurred in 4 seconds.

I’ve experimented with exporting the data using file > export as > tab-delimineated text. However, the data is exported like this:

Begin Time End Time Duration SpeechActResponse SpeechAct
2.005 4.032 2.027 S-DESC
5.104 5.701 0.597 R-MOVE S-OTH

Any help is much appreciated!

Thanks,
Jamie

2

The short answer to your question is (as often is the case): no.
But there might be a way to achieve what you want. You could add a new tier and use the Tier->Create Regular Annotations… to create a sequence of adjacent annotations with the same duration (e.g. 1 sec or 30 sec, your sampling rate) for the entire duration of the media. Possibly number them via Tier->Label an Number Annotations…
You have to save the file and then you could try the Export Multiple Files As->Annotation Overlap Information… This allows to select one ‘reference’ tier (your sampling tier) and one or more other tiers for which information regarding overlap of annotations and of ‘first annotation after reference annotation’ is exported, tab-delimited. You’ll have to delete most of the columns in e.g. Excel, but I believe a few of the overlap columns will be interesting. Depending on how everything has been coded in your files, I only have a rough idea of that. Please consult the manual for the mentioned functions and/or just give it a try.

-Han