@drkatedevlin
how about you give the talk anyway - do it online and record it so those who were stopped from hearing you can catch up. I suspect you might get quite a big audience…
Applicants for fellowships: Please preprint your papers. I'm sitting reviewing a proposal from a newly minted PhD. 2 first-authored papers are 'under review' and I can't find them on the web. I could evaluate proposal better if I could read them.
@wendy_bickmore
Sorry Wendy, but this is just wrong. People are dealing with major life events; some are sick or have sick relatives; many are trying to work with kids at home or creating online teaching materials. Most are exhausted. Don't shame them for not reviewing.
Heh. Just been called an 'old battleaxe' by a man who I asked to move when he talked on his mobile in quiet coach. Feel quite proud of the title. He moved, to general approval of rest of carriage.
As I wade through a load of research proposals, the one thing many people don't grasp is that most people on the committee who will judge your work don't have specific expertise in your area. Write for them.
I've just posted my course on data simulation for complete beginners on Youtube: . Idea is to give intuitive grasp of basic concepts - insights from simulated data into p-hacking and power. Complete with pandemic hair.
Just had request for data from 2011, which confirms all the advice I've been given:
- You will not remember old passwords
- Old hard drives won't work any more
- You will have many files called 'final'
- Keep your scripts as well as data
- Store open data for your future self!
Having watched the Cummings statement, I have to say I was rather disappointed with the media, who just had their prepared questions and did not think on their feet and respond to some of the new information. So here's the questions I think need to be asked (thread)
Submitted complaint to
@BBC
. Fully expect to get fobbed off, but this, to my mind, is an important question about the
@BBCNaga
case.
BBC say they get 200,000 complaints per annum, (over 500 per day) so why was this one acted on? attn
@krishgm
@theAliceRoberts
i’ve always found this reaction odd. I mean, if you believe in a supernatural being, then he’s not going to be constrained by the laws of physics. It’s only improbable if he’s *not* supernatural
I shouldn't really try and run a complicated power analysis first thing in the morning. I've just got a message to say my required sample size is 11796875000000000.
Finding it increasingly difficult to review papers as I just don't believe much of the research I read, given the relentless emphasis on positive findings. Even when studies are well-powered, and predictions were clear, null results are just skipped over.
New free Open Source book, with Paul Thompson, on Evaluating What Works. Intended as a primer on methods for evaluating interventions for those with little background in statistics and research methods. Examples are from speech and language therapy.
We need a word for the thing that you can't throw away because it looks like an important component of something, but you have no idea what it is, so it lives in your home in a drawer for 20 yrs
Autism male brain study retracted. Kudos to authors for documenting an honest error so clearly: problems with analysis script led to systematic error. Good reason for making all data and scripts open to help catch this kind of thing.
Peer review is much maligned. My usual reaction:
On receipt: who are these idiots they have used as reviewers?
Day 1: I guess I'll make a list of their points and how to respond to them
Day 2: Well, yes, this point is a good one. Erm, and this one...
Day N: Paper is much better
Respect for those doing this giant RCT. Maybe it's time to accept that there's little transfer for cognitive skills: if you want to learn something, you have to study it.
4,000 10-year-olds were randomly assigned to learn chess in a school-randomized trial across 100 scools in the UK. Pre-registered analysis showed no effect on math, reading, or science tests. Yet again, correlation doesn't equal causation
@JohnPeterJerrim
well, I hoped to submit my last paper on retirement day, but editorialmanager had other plans. Took 30 mins to build a pdf and mangled one of the figures in the process.
After rinsing and repeating have accepted defeat.
As part of my campaign to get everyone to learn about stats and replicability by simulating data, I have posted some slides intended to be v basic introduction, with emphasis on power/p-hacking
i usually reply politely to emails requests for internships, but I’m at a loss when someone writes that they have skills in ‘coming up with hypotheses that I always succeed to support’. Just depressing that they’ve been trained to think that’s a selling point
Good morning. Our bodies have not changed much in 200,000 years. But our minds and behaviour has, radically. Darwin, Hamlet and me on how we became modern, from my new book, the
#BookofHumans
Bad news: attempt to trim hair with husband's clippers an unmitigated disaster. Considering buying a burqua.
Good news: no public appearances scheduled for foreseeable future.
Not putting sample size in your Abstract is a particularly serious problem for papers behind paywall, as I often don't know if I'm interested or not without N
@PeterStefanovi2
While I appreciate everyone trying to nail down the Tories on this point, a deeper issue is, if we had 40 new hospitals, who the hell is going to staff them, given that the gov seem to have no interest in making a career in medicine or nursing attractive.
C'mon
@guardianscience
. You can do better than this recycled clickbait. Headline says 'can detect', lede says 'could'; Study with 2 groups of 38 and more variables than subjects, and you don't even link to article.
I thnk I'm going to spend the rest of the evening watching the counter go up on this petition: it's more soothing than watching the PM telling me what I want
This is an amazing thread. Someone buy this woman a drink! I just love everything she says - hope a sensible journal editor will see this thread and offer to publish it. We need people like
@kirstyjean
in science.
Today I received the worst review of a manuscript I’ve ever had. The TL;DR: great study, shame the results were negative. Here’s why that’s a crappy attitude - a thread (1/n)
Reviewing a paper with all data/scripts well organised on OSF: puts me in a v good mood. Not only do I have confidence in data, and can see how careful authors are, but can also learn hugely from seeing the scripts. Will try to emulate!
@SusanEbbels
@deevybee
Yes we're in the middle of publishing a whole raft of null results NCTs. I'm increasingly convinced there's no alternative to long-term good quality preschool education.
Ahem. Oxford University is a signatory of DORA and should not be posting advertisements that give 'publications in high impact journals' in a job specification.
Sometimes I think the people who do the predatory journal emails are just seeing how far they can go in craziness. This one really brightened up my day.
I'm not amused by Dr Birx. She, and Fauci, continually respond as if Trump has said something reasonable that just needs a bit of qualification or context. This will cost lives.
She should put on her sweetest smile and say 'that's complete and utter nonsense, Mr President'
Get a load of Dr. Birx's demeanor after Trump tells her, "I would like you to speak to the medical doctors to see if there's any way that you can apply light and heat to cure [coronavirus]."
Apologies for anyone I have emailed this morning. I try to avoid emailing at weekends, but was under the impression it was Monday.
Husband just put me straight.
I was shocked. Like most people I thought Anova automatically adjusted for N comparisons. It does that, but only for levels within a factor, not for N factors/interactions. I couldn't find a stats textbook that told you this.
@SRHAstraea
It's astounding how high the false positive rate it. I am still digesting this after reading Dorothy Bishop's blog post a few weeks ago
@deevybee
Just a note to look after yourself!
Stress of disrupted routine is exhausting. Happened when our building was evacuated cos of asbestos 3 yr ago. Apart from anxiety re future, you can't do stuff on autopilot - everything takes mental effort. So don't push yourself too hard.
@rolandmcs
i remember watching this when it first aired. I was astounded that so many people sided against Rushdie, and especially disappointed with Shirley Williams. Christopher Hitchens is much misse
I am getting so many (anonymous) emails with people who want me to check certain authors or papers that I cannot possibly follow up.
So many names...
And so much hidden pain among honest scientists about these dishonest coworkers.
I need help with this. It is just too much.
New preprint with
@AbalkinaAnna
with evidence that a paper mill has infiltrated an established Wiley journal - Journal of Community Psychology
(thread)
Dear Researchers
If you fail to replicate your own published finding, it's helpful if you state that somewhere publicly.
Otherwise, if you only tell people privately, then to rest of world it looks as if we are suppressing evidence if we don't cite it.
#tryingtodotherightthing
@Chris_Bishop
it’s been done, at least in UK in written version. Laurie Taylor column in Times Higher reported weekly from University of Poppleton, where Maureen, the secretary, was only competent person. It was utter genius. Worth checking if any of available online
Sick to the back teeth of
@APA
journals giving no options but to purchase PDF when you locate an article on the web. Other publishers are able to work out if your institution is signed up, or at least offer a button to sign in.
@APA
needs to get its act together.
Fed up with journals that rely solely on automated systems to communicate. A recent exchange:
J: Will you review this article?
me: Yes, if code & data are available.
J: 'Recently, we invited you to review Manuscript x. You have yet to respond'
me: Piss off.
Schools are to join a child mental health trials to gather evidence about what helps to keep youngsters well. One test will use mindfulness, relaxation techniques & breathing exercises, a second will offer classes to understand mental health & disorders.
Exact parallel to what whistleblowers are told when revealing research fraud: we have to keep it confidential, or we’ll undermine trust in science.
No! the way to earn trust in politics or science is to show that there is no place for fraudsters
We've recently published a series of papers in
@WellcomeOpenRes
reporting findings from a study of the impact of an extra X or Y chromosome (ie. trisomy) on development. Data collection took 6 yrs! I'm going to try a mega-thread summarising the main findings
Non-U.S. academics: If you had the opportunity to take the perfect academic job in the U.S., what (among these contrived options) would give you the most pause?
1) Restrictions on freedom of speech from the left
2) Right-wing politics
3) I wouldn't have pause
A massive 5 year, multi-school, $500M effort to improve teaching by measuring teaching effectiveness, helping teachers improve, and taking effectiveness into account in HR decisions resulted in... no improvements at all for students. Surprising outcome!
I've been asked about intros to data simulation.
Here's slides explaining the basic idea
With links to my scripts in R for beginners
See also refs toward end of this:
Please feel free to add other sources
7. It is not the action of a rational person to make an 80 min round trip with your wife and child in the car, in order to test whether you are fit to drive. Was this in fact a trip to celebrate your wife's birthday, as is widely suggested?
Academics like evidence, so maybe someone could do a study to find out whether the great timesink that is the mock REF makes any difference to the actual REF.
But perhaps first check out if the REF has improved or damaged research quality in the UK.
could I make a plea to drop headlines saying ‘how worried should we be?” and instead say “what action should we take?”
Make us active agents, not passive pawns
Sensible comments on pausing the REF, but maybe, just maybe, this hiatus will lead us to conclude that universities are a public good and we should stop spending so much time/money on proving their worth, so they can get on with their primary functions 1/2
This has a fantastically accurate evaluation of what is wrong with poster sessions - for presenters and audience - and has a neat solution : and Mike wants you to help him evaluate it.
Strongly recommended by this serious poster-session phobic.
Let's fix academic posters! Prepping a poster for
#SIOP19
and sick of the old "wall-of-text" poster design? Watch this cartoon to see a new, faster approach to designing research posters. Includes templates.
#betterposter