James Borders

How many trainees were supported by your award? Enter your answer 14.What was the reason given for your grant’s cancelation? Enter your answer 15.Is there anything else you would like to share with the Committee? Enter your answer 16.Are you comfortable with the Committee publicly sharing the information you provided? If you select “ask me first,” please be sure you have provided an email address or phone number. Yes No Ask me first

🧪 Attention those with terminated grants, including for NSF, NASA, EPA, NIH, etc:

Let Congress know about the impact of terminations via this form: democrats-science.house.gov/grantcancela...

Only Dem staff on the House Science Committee will have access (unless if you explicitly state otherwise).

Interested in learning about data sharing? Dr. Micah Hirsch and I will be hosting a free webinar through CSDisseminate.
🗓 June 3
⏰ 4-5pm EST
🔗 Register: www.eventbrite.com/e/open-data-...

Hello! We have finally made the leap to @bsky.app. We look forward to sharing our lab updates on #cough and #swallowing #clinical research here!

Hi @lisa-everton.bsky.social - could you please add @airwaylab.bsky.social to this list? Thanks!

It's time for BU Giving Day 2025! ♥️ If you're interested in supporting the Stepp Lab this year, the link to donate is in our bio. We greatly appreciate any and all contributions, and we are immensely grateful for all of the support you all give us and our work year-round! ✨

Yes!

We’ve got some exciting news: OpenCSD and CSDisseminate are joining forces and becoming 1 working group! Our mission is still the same: to spread the word about #OpenScience in communication sciences and disorders. But now we’re doing it together.
www.csdisseminate.com/Announcement...
#SLP #AuD #CSD

Thanks for sharing - let me know if you're able to get in touch with your PO. I'm hoping to get some clarity on my council which was initially scheduled for late January.

Are you NIDCD by chance? I see your research focuses on autism and literacy.

Thanks for the update - do you know if we're allowed to contact POs?

That's a good sign. Have you been able to get in contact with your PO at all?

Was your council date previously scheduled (and now moved to a later date because of the freeze) by chance? Trying to figure out if they're now rescheduling councils that didn't meet.

During my maternity “leave” last sem, I helped write 3 trainee fellowship applications to NIH & NSF. Probably ~50 hrs of my time, certainly hundreds of theirs. All have reviews occurring in ~month. In some cases it affects their ability to stay in my lab. This grant review shutdown is devastating.

Yeah academia is all about persistence and dealing with rejection. Just gotta keep moving forward

Congrats! Every revision is a chance to grow and learn (and it never ends). I’m always skeptical of stories of an acceptance without revisions. Science thrives on feedback.

Performing high‐powered studies efficiently with sequential analyses Running studies with high statistical power, while effect size estimates in psychology are often inaccurate, leads to a practical challenge when designing an experiment. This challenge can be address...

Anyone know of a behavioral clinical trial that used a sequential analysis to potentially stop early for futility or efficacy? (eg onlinelibrary.wiley.com/doi/full/10....)

Bonus points if it's publicly funded. I'm particularly interested in how one would justify this to a grant agency.

And provide a free website interface with R code to generate your own synthetic datasets - csdsynthetic.shinyapps.io/synthetic_da...

We hope this study will encourage discussion on the merits and barriers of data sharing in CSD, as well as the utility of synthetic data.

In this study, we found that synthetic data largely maintained the statistical properties across studies representing the ASHA Big 9 domains.

We conclude with a proposed framework for sharing different types of data

Synthetic data involves creating an artificial dataset that doesn't represent real participants but maintains the dataset's statistical properties.

These methods have been around for >30 years and are used across academic fields and government agencies like the US census bureau.

Reproducibility is a core principle of science, but requires sharing data. Unfortunately, data sharing is uncommon in the field of CSD.

Concerns largely center around maintaining participant confidentiality, particularly when consent cannot be obtained.

Excited to share a new preprint with @slpaustin.bsky.social & Elaine Kearney where we examine how synthetic data can be used to share sensitive data in CSD while maintaining participant confidentiality & statistical properties of the dataset.

A brief 🧵...

osf.io/dutga

First application on the job market submitted! Fingers crossed 🤞

I guess this goes beyond whether one *can* use Bayes theorem in this example; and more so whether they *should*

Thanks! Super interesting. And using values from different studies makes sense? Or should they come from the same study?

My gut feeling is that we shouldn't be doing this (for a myriad of reasons), but interested if the math checks out. I've only ever thought of Bayes theorem in the context of a regression model with actual data.

#Bayesians- just read an article where they use Bayes theorem with values from the literature to determine 'risk' of an event (e.g., the prior is baseline prevalence and LR is odds of the event). Is this legit? @tjmahr.com @solomonkurz.bsky.social @rmcelreath.bsky.social

doi.org/10.1044/2022...

Excerpt from this blog post: hbiostat.org/proj/covid19...

Bayesian sequential designs almost seem too good to be true. I can't think of a case where we wouldn't use this type of design in clinical trials to save $, resources, and patient's time.

Slowly working my way through some of Frank Harrell's resources. Amazed at his ability to communicate complex topics in such a clear and accessible manner.

Appreciate it! And huge thanks for all your work - it's so accessible and has been incredibly helpful to my learning. I've lost count of the number of times I've shared your blog posts and translations of various books to colleagues.

Basically I want to be able to say that I can detect a 1-unit change in this mean response with X sample size. But when you simulate data for an ordinal model, most packages require you to conceptualize changes on the log-odd scale (for example).