#childlang friends!
Milestone:📣 @bliplab's ✨heroic✨ multilingual team have passed the 300 transcription mark!
We used an onscreen wordless strybookbook to collect over 400 parent-child conversations on video calls. Look at this amazing progress! T1 and T2 almost complete!
@bliplab We described our #LockdownScience methods in this paper, where we talk about conducting the study online, and the strategies we used to keep our 142 parents engaged over the three time points (Retention rates: 96% to T2; 92% to T3). 💪🏼💪🏽💪🏾 frontiersin.org/articles/10.33…
@bliplab Our multilingual families in SG provide wonderful data about patterns of #translanguaging across the course of a story. In this example, you can see has lots of English at the start and end, and lots of Mandarin Chinese in the middle, but the pattern is constantly shifting.
@bliplab To look at our data in this way, we've developed a special way of transcribing - a protocol and set of tools called BELA. The BELA transcriptions allow us to mark out bits of speech where the words originate in different languages, or have an unclear origin 🤔
@bliplab Big thanks to Corpus all-star cast!
BLIPLab Manager: @FeitingW
BELA Program Lead: Le Tuan Anh
Corpus Manager: Victoria Chua
w/
Serafina Fong
Wong Jin Yin @ShamalaSundaray
Eshwaaree Yogarrajah
Nur Sakinah Mohd Salleh
Shaza Amran
Annabel Loh
Vinitha Selvarajan
+ our team of 30 RAs!
The wordless storybook was designed with #MultilingualismInMind
It is open access and free to download with CC-BY-NC 4.0 Licence 🎉
Great tips for helping participants enjoy your studies - and a pathway to better data!
Another tip is to think about whether your task’s structure leads participants to expect certain kinds of questions or actions. @VanjaKovic and I call this “experimental pragmatics” 1/
@VanjaKovic Once, we ran a study where TWO THIRDS of our participants responded weirdly😱 Turns out that they read the instructions, but the task itself led them to expect something different - They thought we made a mistake in the instructions, so they ignored them! osf.io/gkz2w
@VanjaKovic The pattern was really clear - our participants were trying to help us by doing 'the right thing' even though it was not what we asked them to do
The Experimenter-as-Idiot Fallacy: They thought we were idiots, and tried to 'correct' our mistake🤦🏻♀️
Ah I really enjoyed being Discussant for today’s symposium at #SPR2022@TheRealSPR
It’s a joy to bring these three projects together and talk about the challenges of research using high dimensional neuro data and some pathways to doing better! 1/
@VanjaKovic led us off with an intro to the three great projects being brought together for this syposium, highlighting their focus on building open communities to advance research
@angelasoskic shared the background behind the ARTEM-IS project — reminding us that even within one rather narrow study type (N400s) there can be as many analysis pipelines as there are papers!
Very excited to share this work led by @hannahlgoh with @LucaOnnis2
Can bilinguals’ of a language with one type of retroflex (in fricatives and affricates) learn to hear a different kind of retroflex (in stop consonants)? Can they transfer their perceptual skills?
What’s a retroflex? A speech sound that is pronounced with the tip of the tongue flipped over a bit, like in this diagram (the purple line). In some languages there are phoneme contrasts where changing the shape of your tongue like this makes a difference in what word you say!
Languages like Hindi and Tamil have contrasts between “d” sounds made with these two tongue shapes.
We can write them in the IPA like this:
/d/ alveolar
/ɖ/ retroflex
English only has /d/, and many English speakers find it hard to hear differences between /d/ and /ɖ/
Join us at #SIPS2022 tomorrow to help make rock solid tools for transparent neuroscience!
We believe the right kind of tools can make methods reporting smoother and less error-prone.
We’ll introduce the ARTEM-IS Web-App (beta) for you to try, and we can work on improving it! 🧵
Accurate methods reporting is super important because different processing pathways can lead to different results, as we’ve shown in one multiverse analysis of N400s 🤔
(see how the Related version Reversed comparison has different outcomes in different pathways)
If we want to improve #reproducibility and #replicability in our field, we need to make sure we can accurately report what has been done - which path was taken through the Garden of Forking Paths?
Lots of reporting guidelines currently exist but…
So pleased to see our first report of translanguaging in child-directed speech as a poster at #WILD2022!
Since our Corpus team couldn't travel to San Sebastian, our fab lab-mate Han will be standing by to answer questions during the poster session at 16:40 CEST) today👇and also..
..in a bit of an experiment, I'll also be standing by for any Q&A about methods, transcription protocols, how on earth we managed 92% retention rate for the 143 children in the sample (8m to 4y), over three picture-narration tasks spanning 4-6 months!
This poster is really exciting for me because it was a study we rolled out during lockdown. At the time we were planning an ambitious series of home visits, but suddenly we weren't able to visit any homes! So we redesigned our entire protocol for online participation.
In a few hot minutes I have the great pleasure to deliver a plenary lecture at ‘New Horizons in Education’ for the University of Belgrade. I’ll be talking about “De-Centering the Professor” with a focus on diversity and access 😃 1/
It’s exciting because I really take seriously the idea that teachers are quite different from the vast majority of the students in our classes, so I want to share some strategies for making the class better for people who aren’t US 2/
So here’s the experimental bit! I don’t speak Serbian, and although the conference languages are both English AND Serbian, I know that there will be some attendees who have stronger Serbian language skills than English…
3/