Видео

I disagree that you should use the observed effect size, I think you should use the MCID or the effect size that was selected during the a priori power calculations

You should make it explicit that you’re defining the false positive risk as the posterior probability of H0! Also great lecture

Amazing! Thanks for sharing

So lowering a given fields p-value threshold to 0.005 in terms of publication standards would achieve a FPR of <0.05?

Hello Professor, off-topic here, but I learned a lot from your explanations regarding the false positive risk and the misunderstanding of P(D|H) and P(H|D). It was one of the few extremely intuitive descriptions of the problems I’ve heard. Can I ask you about your take on Magnitude Based Inference? It’s an inference technique that grew out of sports science. Superficially I don’t see problems with it but it has been criticised as being statistically unsound. Could you give your input, if you’re aware of it?

Death ruclips.net/video/FfpBchzBUHs/видео.htmlsi=qQSYaWLUEWTtg7Gb in small doses

Thanks for the talk! Writing “P values < 0.05 were considered significant” is pretty much universal in biological science papers. Would there be value in this context of quoting a smaller P value equivalent to the 5% false positive risk? Of course, one consequence would be less gets published, but maybe that is a good thing?! That suggestion aside, I do like the suggestion to quote FPRs _in addition to_ P values. It’s a fairly gentle but significant(!) move in the right direction.

Wonderful. I would like (with permission and attribution) to use some of this in my talk "You can't just pull facts off the internet" designed to get students to read actual papers & books in depth and not just do Google searches when looking up information.

wow! It such a pleasure to watch single channels in action

26:26 all the positives gives true risk; what's the denominator?

17:15 what if 99% are not that impacted by a diagnosis, but you roll out large screening tests.

25:01

Lovely to watch this. David was my piano teacher back in the 1970’s when he lived in Grange Hill. I’m not sure I was his best pupil but I enjoyed the lessons and David’s patience! At the time I was not so interested in classical music but wanted to learn to read music so I could learn how to play the songs of my heroes at the time (Elton John, Billy Joel, Simon & Garfunkel) but now in my 50’s I am returning to classical. Please remember me to David if you are still in touch. Best regards.

No nonsense from this guy.

Does she speak any other languages?

Thank you for the kind comments. There is one change that I'd like to make in this video. What I called "false discovery rate (FDR)" in this video, I would now call "false positive risk (FPR)". The reasons for this change in terminology are explained in a 2017 paper: royalsocietypublishing.org/doi/pdf/10.1098/rsos.171085 A 2019 paper goes into more detail about the false positive risk: www.tandfonline.com/doi/full/10.1080/00031305.2018.1529622 and I made a video about it too: ruclips.net/video/jZWgijUnIxI/видео.html

Amazingly helpful. Tried to find resources online to help me understand what a permutation test was and this video was the only one that helped. Thanks a lot.

🎉🎂🎈

Thank you very much for your very helpful contributions to EBM, I have very much enjoyed your papers. So the FPR is essentially 1-PPV ? How does this relate , conceptually, to positive likelihood value? ( which I understand as the risk a positive test is a true positive divided by the risk it is a false positive)

Okay. I am very new to this and still a rather confused on what is being discussed here--I get the general theme, but the specifics are eluding me. For example, at min. 17:00 (approx.), where you are unpacking the screen tests. I do not understand the prevalence number: how can our screening tell us that 100 people have a condition when whatever testing we're doing, the sensitivity factor discussed, can tell us that only 80 people have the condition??? To me, it seems our prevalence is 8/10 of 1% (0.008) rather than 1% How can we possibly know the prior prevalence when we can only empirically confirm a lesser number? Apologies for the glaring ignorance here for those more seasoned with this, but I'm not getting this. I'll keep trying.

An amazing introductory slide; thank you for your work, Professor!

I love the modulation near the end, just after 20:35

The introduction is already out of date. In May 2019 Rosi, and her fiancé, Stephen Caudwell, were both elected as Councillors. And in October 2019 it was announced that Rosi had been selected as the Green Party candidate for the next parliamentary election.

Thank you for this interview. I strive to be as well rounded and accomplished as her.

You can find collected links to all the stuff I've written about his at www.onemol.org.uk/?page_id=456

I made a ratio version of FPR calculater. www.tutorialspoint.com/execute_r_online.php?NGsTgJ It needs package 'pwr' to execute, but graph is shown without it. I am afraid there are many bug in it, due to my poor understanding. Any improvement or debugging is welcome.

I made a ratio version of FPR calculater. www.tutorialspoint.com/execute_r_online.php?NGsTgJ it needs package 'pwr' to execute, but graph is shown without it. I am afraid there are many bug in it, due to my poor understanding.

Rosi's father told me a couple of fascinating things after seeing this interview. "When we moved house in 1985 - Rosi at the age of 8 - her then primary-school Head said to me as a sort of parting shot, something to the effect of "you do know you are holding a tiger by the tail, don't you?"". and "when I asked Professor Harold Simmons, her PhD supervisor at the time of that graduation, what he thought she would be doing in twenty years time, he replied "running the country, I should think" " The first prediction was certainly right. The second one -we'll see.

Thank you Professor for your lifetime of work, it's been a staple of my diet as a pharmacology student at UCL. And of course, thank you for fighting pseudoscience wherever and whenever you see it.Truly awesome!

There is now a version of this video with better volume for most of the soundtrack, thanks to editing with Lightworks. See ruclips.net/video/jZWgijUnIxI/видео.html

Postcript. Rosi now runs the Combat Sports Clinic, at www.combatsportsclinic.net/. They have recently produced a video neck strength training, designed to help people who do contact sports -things like rugby, boxing, muay thai and MMA. I've seen only the preview, but there is certainly nothing quackish about it. It's about strength training. You can buy it at courses.combatsportsclinic.net/collections

This was very helpful, thank you very much!

This talk id now superseded by the 2018 version, atruclips.net/video/iFaIpe9rFR0/видео.html

Thank you so much Prof. Colquhoun. I am studying on analysis methods of biological screens and came across this video. This is really helpful to having a more precise understanding of what p value is and what kind of mistakes we will make if using it inappropriately.

Your sound is set very low, for the bulk of this video. Only the intro and the excerpt of Kristin Lennox were what i'd call 'reasonable' so i turned the volume on my speakers up. If the Liberty Bell March hadn't been set low too, i might have jumped out of my skin :-D P.S. RUclips recommends i watch a video called "Fakes in the art world - The mystery conman" after this one. I wonder how on Earth they come up with values for the prior probability that a sculpture, or painting, were genuine in progeny, in the world of art. What a statistically-scary thought :-/

Thanks professor

Thanks for the video, very well explained. And also, on an unrelated note, I couldn't help but notice what a great voice you have, sir; it felt like watching a Discovery channel documentary! Good for your students! :)

Dear Professor, thank you very much for the helpful video! I still struggle with estimating or even understanding what p(real) really means. Let’s assume I investigate the effect of tobacco on an endurance parameter by recruiting smokers and nonsmokers and compare their endurance. What would be a good estimate of the probability of a real effect? I mean we already know that tobacco use correlates neg. with endurance, so should we assume that p(real) is 1? Thanks a lot!

Is this a Bayesian approach? Prevalence should be the a-priori probability of the condition being diagnosed in the population subject. Am I right?

I found this presentation generally on the spot and really not that provocative from a ’metric point of view. What I found most interesting was the focus on individual cases and the fact that researchers that are recognised as successful (in terms of prizes or general recognition) would not have met simple (and stupid) baseline counts for performance as demanded by both internal and external forces in academia. But most (if not all) bibliometricians are aware and of the shortcomings of the H-index, JIF, the skewness of citation distributions and other simple measures and would never use them in evaluative practice (or when they do, in very controlled and modest forms ((really!)). Instead it is more likely that it is lazy administrators or cheap leaders in academia that would carry out simple DIY-bibliometrics for the detriment of scholarly practice in the local setting in the short run and as a potential negative consequence in research in general.

thanks for talking about this important topic. as junior faculty it can sometimes feel like I am a slave to the impact factors... and it really hurts curiosity-driven science. but since metrics are not going to go away, is(are) there metrics that is (are) meaningful?

This is wonderful. Thank you.

Excellent overview. One significant problem: several times in the video, the screen goes black for quite awhile. for example 2:04-2:43.

Thank you!

Thank you very much professor Colquhoun!

Finally, a common sense approach. Excellent, level-headed analysis.

Phil Hammond..Look at the last 3 decades...the leaders of the royal colleges ,the patient organisations and nursing over the last few decades have been part of the cause of this mess as much as successive governments.I emigrated to Australia as a Dr because the leaders of medicine in the UK were either spineless or expedient..The only thing to improve the situation is a new zeitgeist of junior drs with more principles than their elders..I am talking about GPs giving up the 24 hr commitment to patients.I am talking about the catastrophic sudden unmonitored reduction in Junior dr hours and the loss student nurses being able to nurse patients on the ward.I am talking about the use of untrained nurses to see patients in walk in centres without proper supervision.I am talking about poor care that was by no means only seen in Stafford Trust.It was seen in all hospitals over those years.The junior drs have had enough of the BS

What a hero. 7 day full emergency and non-emergency services on a 5 day budget, which can barely provide safe 7 day emergency services and 5 day non-emergency services as it is with cuts. Why do our European counterparts invest more money than us? Why do they have more doctors? More scanners? More beds? I thought we were the 6th largest economy in the world? The health of the British public is worth more than that. Our family and children are worth more than that.