Devon Greyson

Thanks! I’ll go bug some epidemiologists about this.

Right. Just wondering about relative validity of wastewater surveillance (obv only where there is municipal sewer) vs interview-based estimates vs other approaches I’m unaware of to try to establish this range.

Assuming that some % of milder measles cases are going unreported in recent outbreaks, what would be a valid way to estimate total cases in a region experiencing an outbreak?

I missed that Statistique Canada had released 2024 life expectancy.

Life expectancy in Canada and US in 2024 has virtually returned to 2019 pre-pandemic levels (US 0.2y higher, CA 0.06y lower).

www150.statcan.gc.ca/n1/daily-quo...

Oh, no, I am just uneducated, and it's Arabic-Indic. IDK how you change it back, but that's what it is.

Watch it be some made up language from a fantasy novel or popular YA scifi series mystifying us, haha.

I literally assumed this was AI by the nonsensical watch face numbers before I realized who had posted it. #NotAStroke

It is not but I guess she wanted to quote it anyway.

Towards the study of world misinformation | HKS Misinformation Review What if nearly everything we think we know about misinformation came from just a sliver of the world? When research leans heavily on online studies from a few wealthy nations, we risk drawing global c...

This paper argues that misinformation research needs to move beyond studies of a few wealthy countries and online samples to include diverse, real-world information environments worldwide in order to produce more accurate and effective insights. I agree!
misinforeview.hks.harvard.edu/article/towa...

Part of being a community-engaged researcher is treating communities w respect. This includes being forthright abt the limitations of academic research, e.g., it takes a LONG TIME to get to publication, to get rsch funding we have to do something *new* in scholarly terms, not just good for people.

My off-the-top-of-head ideas:
1. Refine the RQ to add novelty & still address their underlying issue (maybe add a 2nd more novel Q);
2. Use this topic as case to test/further theory;
3. Add novel methods to make methodological contribution;
4. Possibly a comparative study if that would add novelty.

^By "not much scholarly novelty" I mean the topic, research question, and method they suggest to you have all been done before with other, likely-similar communities (perhaps in slightly different geographies) in the past.

(I have my own ideas, but am curious what I may not have thought of!)

An interesting question arose in conversation last week and I'm curious about thoughts from other academics/social scientists:

What do you do when a community partner (esp a marginalized community you have built a relationship with) wants to do a study that doesn't have much scholarly novelty?

😱

Graphic featuring details of project. Pictured is Dr. Ogilvie wearing a black turtleneck, Dr. Marquez wearing a black button-down shirt with red dots, and Dr. Sekirov wearing a pink sweater.

Although Canada has made significant progress through screening and early treatment, recurrence after standard care continues to affect roughly 10 per cent of patients. The complex interplay between HPV, the immune system and the microbiome of the genital tract is still not fully understood. The Terry Fox Research Institute is investing $2 million to help change that. With this grant, a multidisciplinary team led by Drs. Gina Ogilvie, Inna Sekirov and Citlali Marquez will work to uncover how the body interacts with HPV and pinpoint those most at risk of cervical cancer, paving the way for personalized and preventative follow-up care.

"The knowledge gained from this research will help improve how we identify and monitor patients at higher risk, leading to more precise and effective follow-up care." - Dr. Gina Ogilvie, University of British Columbia.

March 4 is Intl. HPV Awareness Day: Ending cervical cancer is within reach, but only if we close the remaining gaps.

That's why TFRI is investing $2 million in a BC-based research team working to decode how the body and HPV interact to stop cervical cancer.

Learn more: tinyurl.com/mukh5fan

In 2025 things changed @ NIH. It's possible my colleagues and I over-represent areas of research heavily targeted (e.g., vaccination, gender & women's health, health equity) for elimination by US regime, but overall US rsch $ was way down last year & foreign subgrants halted at least temporarily.

15% even seems optimistic based on recent competitions (not all proj grant, some team/op/catalyst). Have we seen any documentation yet regarding any changes in appln volume since US largely got out of the funding biz (at least in certain area).

Not all tools are right for all jobs. A LLM trained to spit out standard outlines, format writing, or audit it for hallmarks of clarity, might be helpful In scientific or business writing. An LLM generating novels, though, will inherently be derivative, missing the creativity human minds can bring.

A decent starting point to assessing potential of an “AI” technology is to ask its purpose: is it aiming to improve consistency of a task & rapidity of algorithmic change? Might work—these are things mechanization can be good for. OTOH, if it just aiming to save $ by eliminating humans, question it.

Some tech under the “AI” umbrella, such as ML to screen for tumors on radiology scans, appears to have great potential. Some, such as prescription chatbots, are an attempt to cut corners and highly vulnerable to error and manipulation. It is essential to differentiate between these

And for those unfamiliar with Chatman's "small worlds" theory, this is an accessible application of it to academia: ojs.lib.uwo.ca/index.php/el...

All to say: I don't have the energy these days to get worked up about hyperbolic academia spats.
My plan (& what I recommend to those who ask) is to take the useful kernels in such essays, like the importance of focusing training on higher-level work not the easily-automated, and move on.

Yes, a lot of that article is effectively ragebait. Virtually no one serious about literature-based research methods would agree that genAI does lit reviews better than a *good* human lit review. Maybe this statement speaks more to the poor quality of lit reviews the author encounters?

I myself have given career advice to recent public health grads to ensure they have the people and/or strategic skills necessary in the field to make themselves non-replaceable by ever bigger & faster computational models.

The idea that "AI" can do "social science" research better than humans is limited to computational work w human oversight, at best.

Now, if your bread & butter is doing things by hand that machines can do faster, in *any* field, you will need to evolve.

Re: The "Academics need to wake up about AI" piece going around: This is an excellent example of people in a small world generalizing their experience to everyone in a given community.

Okay, hearing it's not just me. Hopefully just a temporary glitch, although not ideal right before/at Proj Grant due date.

#CIHR folks, I've had 2 Co-A applns recently on which I entered my CCV/MSC/consent & saved, but it didn't seem to "stick." Luckily the NPAs were on top of things both times & reached out. On 1 I had to save & reload 4 times to make the CCV# stick. Is something up w #ResearchNet or is this me?

A screenshot of Rachel Dratch from SNL. She's a white woman with short dark hair and blue eyes, and she's making an awkward expression that implies, "Sorry I have to tell you this and ruin your good time. But also I'm kind of not sorry about it." I've added a subtitle that reads, "That's AI."

an updated debbie downer sketch but it's friends showing her funny videos and she has to respond like this to every single one like this:

Important for everyone to realize: the anti-vaxx activist RFK Jr did indeed TRADEMARK “MAHA”

Before he became HHS Secretary, he declared $100k earned from “a licensing agreement to use Maha brand marks”

He wanted to put it on supplements, vitamins, essential oils to make $$$

He’s a grifter

1/2