ACLEW-DARCLE transcription training
The majority of our corpus data uses the ACLEW- DARCLE (Analyzing Children's Linguistic Experience around the World- Daylong Audio Recordings of Children's Language Environment) coding style. We use ELAN to do so.
***IF YOU HAVE AN APPLE M1 (or later silicon chip) MAC, download and use ELAN version 6.2. anything later will be super laggy and drive you insane.
It's a training scheme that allows us to share data with other researchers and compile very large, compatible datasets that can be analyzed in the same way!
It takes some time to learn, though. You'll read through 10 tutorials and then do several rounds of a "Gold Standard Test", where you have to reach a very high bar of agreement with another coder until you can annotate files for real. Be patient, and don't stress if you fail a lot of rounds before you pass each age group! That's normal!
Tutorials
To get started, read all the way through these tutorials. You can follow along using your first test file if you'd like by following the instructions under
Make a new folder with your initials here (on the sharepoint site, *until we can get you access to the secure drives only*:
OR IF YOU HAVE BLABSHARE ACCESS make the folder here instead:
smb://sox4.university.harvard.edu/Fas-Phyc-PEB-Lab/Seedlings/ACLEW/Training_English
Once you have read through all of the tutorials, then you can start annotating and testing yourself on the Gold Standard files, which will test your ability to annotate and transcribe files with a very low margin of error.
Gold Standard Test
To start, copy this file into your "coder_XX" folder:
(OR on blab-share)
smb://sox4.university.harvard.edu/Fas-Phyc-PEB-Lab/Seedlings/ACLEW/Training_English/ACLEW_Training_Templates/0-7/2337-GS_Training_Round_1-.eaf
Double click to open it once it's been copied, and then go to File -> Save As, and save with your initials after the final hyphen.
Attach the .wav file from the ACLEW_Training_Templates/0-7 folder according to the directions in the tutorials.
Using the tutorials as a guide, you are going to annotate three *audio* files. Each file corresponds to a kid from three different age groups, so you can get a sense of the differing annotations that go with each age bracket (canonical/non-canonical babbling, lexical, multi-word). Each file has five minutes selected to annotate, and each training round is one minute of the five minute segment.
Begin with the 0-7 month age group, round 1 (the first minute is marked out to code).
Begin annotating by following the instructions laid out in the ACLEW DAS tutorials. Remember the first step is to simply listen to the one minute section you are about to annotate.
Once you feel comfortable you have completed a round be sure to save your work one last time and follow the directions in Slide 12 of Tutorial 10 (Export as tab delimited file!!!) to export your file to a .txt file. Use this web app to check to see if you passed the round.
If you pass the comparison, you will move on to the next *age group* and repeat this process, e.g. move on to 8-18 mo round 1.
If you do not pass the comparison, you will move to the next round in the same age group and repeat this process. For each age group, there are 5 potential rounds (five one-minute chunks) for you to complete. Remember, these are difficult and the standard is very high. The point is not to pass quickly but to get better as you go!
The gold standard checker only checks the length of your annotation segments and the tiers with a closed vocabulary (xds, vcm, lex, mwu). Please also practice the MinCHAT transcription (tutorial 8, writing down what everyone is saying!) on each round. Once you pass an age group, let a lab manager know and she will spot check your transcriptions and remind you of anything that needs to be fixed.
Spanish Training Rounds
Files are located in Seedlings/ACLEW/Training_Spanish
DO NOT ANNOTATE - Do only segmentation, xds for not CHI, and vcm, lex, mwu for CHI tier.
Begin with 8-18 (there is not a Gold Standard test for the 0-7 month file).
Check with the same web app as the English Training. For each age group, there are 5 potential rounds for you to complete. Good luck!!
Last updated