Is a scientific career predictable?
Вставка
- Опубліковано 19 чер 2024
- Research on creativity has shown that people tend to produce their most important works when they are young. This is true for scientists too. But a new study finds that this is simply because scientists are more productive in their early years. When the study's authors discounted productivity, they found that a scientist has an equally good chance of producing a high impact paper towards the end of their career.
Read more in this news story: www.nature.com/news/1.20926
The study was published in Science and you can read it here: science.sciencemag.org/cgi/doi...
The data visualization was created by Mauro Martino (IBM Research). - Наука та технологія
I think that formula is ridiculous, if p has effectively no boundaries and cannot be determined (you know, the definition of chance), then there is no way of comparing the Qs of certain scientists in a sample, Q will always be entirely subjective. Although the conclusions seem cringeworthy to me, the data analysis made in the study seems very interesting.
Scientists actually have a lot of experience with finding characteristics that define random variables - although we can never put hard boundaries on chance, there is a sense in which we can see how it works. This is the foundation of the entire science of statistics, which finds and quantifies different kinds of chance. Although p at any one time cannot be exactly given, we know enough about how it varies to infer something about Q. For instance, if the peaks of C are in general smaller for one scientist than for another, we can infer that their Q is probably higher. Of course, there's a chance that we're wrong, but we always have a chance of being wrong whenever we reach a conclusion. The point of statistics is to reduce that chance. If you want to learn more about how statisticians and other mathematicians improve our understanding of different kinds of chance, Wikipedia is a good place to look, maybe starting here: simple.wikipedia.org/wiki/Probability_distribution
I'm also happy to answer any questions you might have about statistics or probability theory!
How do you nurture Q? Change the scientific system! the current system drains people and decreases motivation. Who wants to work to become a "good scientist" when publishing is the only thing that matters, regardless of quality?
Beautiful animation! What programs do you use to produce this?
If you find out, let me know please!
there is not a specific way to create visuals like this its combination of multiple programs you can create something similar with a standart video editing software and 3d software like blender, but i think if you have data this complex you need to write your own program or script
Makes me hopeful for the trajectory of my career. Thank you for this!
Amazing video, nature. You guys are doing what I'd love to do with my work in future.
0:24 woa, *Six Flags* guy invented the pacemaker!!
Absolutely well done and definitely keep it up!!! 👍👍👍👍👍
I am curious how they can say that Q is almost constant. Does this mean they were able to calculate the p values?
Btw:
Someone was listening to "remember the name" it seems.
Well a priori luck should be the same throughout your life.
So from the result of their analysis stating that your impact is independent of your age, you can deduce that Q should be a constant .
bgjhgjghkjuhgbg
Interesting. I would have assumed we need a new p value for each publication.
Chance to do not wash the dishes, chance to set up a perfect experiment and see higgs particels etc.
The value of p varies, but it follows a distribution that does not. A distribution tells us something about the possible values that p might take over a period of time. So, although we can never guess exactly where the peaks will be and how high they'll be, we can know something about what they'll look like - just like you don't know exactly each person will be on a crowded street, but you know it'll be more crowded on Sunday afternoon than on Tuesday morning. So we can see that, while the value of p changes for each publication, intuitively it's distribution should be the same - how old you are or how experienced you are does not usually effect how often you need to do the dishes, how often you find that perfect experiment, etc. IF you want to learn more about probability and statistics, Wikipedia is a good place to start: simple.wikipedia.org/wiki/Probability_distribution
Thanks for the anser. That was more than I had hoped for!
I'm glad I helped! I like telling people about statistics and probability, especially because it's very important in understanding a lot of the factoids we hear in the news, especially in politics and medicine.
Are you checking whether or not the particle went through one slit or the other?
Well you're certainly checking to see if the idea was one person or another and seeing there was one person that received it, or detected the idea, a detector screen, the screen of the prepared mind.
How is Q defined? As in... What exactly is it measuring and how is it measuring it? How do you get a number for Q?
Does these researchers achieved their most citationed paper after they completed their PhD or there are cases, in which this happened before their PhD? Can someone inform me on this?
and, ok, how do you measure LUCK?
That formula is not very scientific, right? since luck is a variable that multiplicates the other variable Q and Q itself "is constant throughout time" .. luck can't be really measured.
Who even came up with that
Scientists actually have a lot of experience with finding characteristics that define random variables like p - what you call 'luck'. Although there's no sense in which we can 'measure' luck, there is a sense in which we can see how it works. This is the foundation of the entire science of statistics, which finds and quantifies different kinds of chance. Although p at any one time cannot be exactly given, we know enough about how it varies to infer something about Q. For instance, if the peaks of C are in general smaller for one scientist than for another, we can infer that their Q is probably higher. Of course, there's a chance that we're wrong, but we always have a chance of being wrong whenever we reach a conclusion. The point of statistics is to reduce that chance. If you want to learn more about how statisticians and other mathematicians improve our understanding of different kinds of chance and luck, Wikipedia is a good place to look (try Simple English Wikipedia for articles without maths jargon), maybe starting here: simple.wikipedia.org/wiki/Probability_distribution
I'm also happy to answer any questions you might have about statistics or probability theory!
I agree with what you said. The concept of luck or chance can be either of the two, scientific and unscientific or, maybe (and this is something I feel more strongly about) it lies in the region where the two overlap.
It all depends on the context, and in the present one, I find the latter case (there being an overlap) to be true.
Without reading the paper: we have two variable p which does not depend on the sample (the scientist) and Q which depends on the scientist. If we say p is constant in every field and choosing a filed is itself a part of the Q, then we can make a nice model which will fit the data. The chance of a high impact paper will follow a Poisson distribution determined the factors of P, Q and time. If a researcher has more time in his career simply he/ she can roll the dice more often to publish a paper.
Chiara B when it comes to human beings, luck can be statistically measured (to a point of certainty) based off of common variations. Part of P may be, for instance, the number of "random"or serendipitous discoveries a scientist makes (ala Pasteur - he wasn't expecting the mold to have antibacterial secondary products. Someone else may have thrown that dish out upon seeing mold but Louis saw the ruin for the breakthrough it really was - P was mold doing its thang, Q was the skill in his correct interpretation.
is this theory are applicable for under developing countries were alots of problems behined beautiful minds?
Please don't tell me Q is static yet early age is a boon to P... I'm still merely on my way to becoming a real scientist yet already mid-twenties...
It hasn't deterred me but I can't say this is a promising thought. I suppose it's okay that not every scientist publishes a breakthrough.
What would one do if given a chance to refresh the way science works?
We are too scared to ask big questions which seems unrealistic for any funding! The Q is meaningless. We are training the current cohort to modify yellow pages.
this was a good one... Thanks Nature.. :)
bhau rai please don't take it otherwise... But before answering... Can you tell me the reason of interest ?
bhau rai OIST, Bhopal ....
I liked the voice of the narrator, so cute indeed! And thanks for the well-educated outreaching of cutting-edge sciences!
So a scientific career is more predictable than how memes become viral? Damn
Predicting career from no of citation is actually very unscientific thing to do.
wow this shit was good , more good stuff please.
How do we measure luck?
Diana Medina With "success" ;)
Sebastián López so with the number of citations in this case?
Diana Medina Although considering the scope the factors may vary, I'd say the most important factor for success is luck. So as the formula suggests success (I guess citations in this case) is directly proportional to luck. But luck is just bounded by probability.
So we have this equation: Citations=Impact, Impact=Random Potential(or Luck)* ability(Qi) . Then luck must be under a probability function which follows the same patterns over time (as they show by ignoring the timing of the publications) Where Q is the ability to take advantage of existing knowledge that does not increase with time. If the most important factor is luck according to your point of view, then what PF will luck fit?
Well, sorry for disappointing you but I'm not seeing this with maths, as I don't believe there's a sufficiently accurate maths model for success hiding in the data.
My Q Is soo big, I can barely walk
you have a very cute voice, speaker :)
the only true concerning meaning of that equation si that if your luck is equal to zero, your q can be as big as you want but the result will always be zero ahahhah
Some researchers are really involved in this kind of work? What a shame! This ideology of seeing research as made by a bunch of individuals is poisonous to every one. This negates science as a collective effort. This is indeed a dangerous political bias since it will have consequences on the funding and career of scientists. It seems pretty clear to me that, apart from a few unquestionable breakthroughs, the bibliometric indicators essentially are indicators of notoriety not of scientific quality. I expect the Q factor to be essentially a measure of the ability of selling its own work.
and now the study for this video gets a nobel prize for its great research on how scientists "are more effective when young". isnt that normal for any career?! people age, and in any case it would depend on the individual.
research on this, research on that. we do research on research thats being done on something else. why? "research" on why one becomes succesful and famous.
whats the point of this study in this video anyway?
.hm.