Profile avatar
wpball.com
Interested in working out how to reduce unfair and avoidable differences in health. Mostly youth mental health. Chancellor's Fellow at Robert Gordon University | Social Epidemiologist | Pro Cycling | Rstats | Causal Inference 🇮🇪🏴󠁧󠁢󠁳󠁣󠁴󠁿🍉
2,341 posts 3,104 followers 1,104 following
Regular Contributor
Active Commenter
comment in response to post
Which interventions though? If there's an intervention with evidence of a causal effect of X on Y*, we should be able to use causal methods to identify populations who will see the most benefit. *I would hope we have something like this before intervening!
comment in response to post
I always wonder why people would approach this issue using prediction methods. What's the research question that requires a prediction to be made? What would we do with the prediction once it's made?
comment in response to post
There's no point in stopping to think about whether we SHOULD do the thing because we CAN do the thing!
comment in response to post
It's a double whammy because it's also an average measure.
comment in response to post
(We hope)
comment in response to post
Just like this research!
comment in response to post
All of my DAGs terminate at the 'visit pub' node
comment in response to post
Thank you! It's easy to feel demoralised when you look at the big media outlets - what you are doing definitely brings me some hope.
comment in response to post
Wow, that's a particularly egregious example of the harms of genetic essentialism in research.
comment in response to post
Bizarre logic, but I assume at least some of that is a defense mechanism against feeling helpless.
comment in response to post
I am not surprised 😋 They should be Virchow's statement...
comment in response to post
Out of interest, who was being dismissive?
comment in response to post
I was advised never to use a question in an article title if the answer was no. What a ridiculous concept.
comment in response to post
What kind of evidence is required to prove a mediation relationship?
comment in response to post
Winning strategy!
comment in response to post
I'm glad you're not ending in Swansea itself - Mumbles will be a nice treat to end on!
comment in response to post
comment in response to post
Sorry @andrew.heiss.phd but I also stubbornly use the %>% operator. I promise I'm not a chatbot.
comment in response to post
Lenin*
comment in response to post
comment in response to post
John Ioannidis is the Napoleon Bonaparte of research waste.
comment in response to post
You just know that they're the worst when their self-written bio mentions how many papers their name is on
comment in response to post
I've encouraged these conversations between my kids and my grandmother, who grew up in rural County Limerick pre-electrification(!) The difference between 3 generations of childhood is quite amazing really.
comment in response to post
Based on previous performances, you'd back Ayuso at his previous best level to at least stay with Yates or Gee (maybe not this Carapaz?). It was probably the best decision at the time. If they went all in on Del Toro & lost more time with Ayuso, they'd still be in a tough spot now.
comment in response to post
After yesterday, I'm sure there will be teams who will want to attack on the Mortirolo and put as many Kj into Del Toro as possible.
comment in response to post
Yes, please do sort out the NHS data - but don't just give it straight to AI companies who only want to extract profit from a public service. Better to promote 'public good' research instead.
comment in response to post
Not only is it preventable, but it appears to be a deliberate effort.
comment in response to post
All the best Ellie
comment in response to post
👍
comment in response to post
Continue using it from github? github.com/anspiess/pro... Not sure in this case, but there are lots of technical reasons that packages can be removed from CRAN that don't affect the functionality.
comment in response to post
Congratulations!
comment in response to post
Some of the co-workers might try to pinch your lunch though - beware!
comment in response to post
60% of the time, it works all the time!