Question of the Day

May 28, 2014

Are your papers reporting “discoveries”?

Or “demonstrating” something?


37 Responses to “Question of the Day”

  1. proflikesubstance Says:

    Why does it have to be one or the other? We commonly have to report discoveries that set the stage for demonstrating something. Sometimes those exist in the same paper, but often not.


  2. dr24hours Says:

    Demonstrating a methodology.


  3. AcademicLurker Says:

    Kind of depends on how you frame things. If you propose some novel mechanism or regulatory pathway or whatever, and then go on to experimentally demonstrate that it exists, did you “discover” the new mechanism or did you “demonstrate” that your proposal was correct?


  4. DrIgg Says:

    Demonstrating the feasibility of our discoveries.


  5. drugmonkey Says:

    AL: I am interested in any sub field or individual trends in the answer to that question. I was recently on a panel of faculty where the difference struck me.


  6. meshugena313 Says:

    Ugh. This gets straight to the bullshit about results being “too descriptive”. I still don’t know what that means. A paper was recently returned to us (by JBC, where I haven’t seen this before) with the comment that the “work was too descriptive and there were no experiments to probe mechanism”. We _reported_ abundant and clear new “discoveries” and our experiments to probe mechanism returned negative results. Our discoveries will clearly be useful for future mechanistic studies, but what the hell do you do when the straightforward experiments all return negative results? In the revision we added another “negative result” mechanistic probe (i.e. it doesn’t disprove the null hypothesis, for a formalistic version).

    Does every paper now have to be a CNS paper? Do they all have to discover _and_ describe?


  7. drugmonkey Says:

    I am intrigued by this notion that “demonstrating” an effect is the same thing as being descriptive. Which we all know is bad. Unless ChiP/SLIP/N/DiP is involved. or ‘omics. or something.

    It may explain why some people talk in “discovery” language all the damn time.


  8. old grantee Says:

    discovery (dis-cover) is a frequently used word in science and, particularly, when teaching natural sciences. It is a word found within the scientific method vocabulary.
    Discovery is uncovering something previously covered ( unknown). A discovery in science is usually accompanied by a demonstration of something that can be reproduced and validated. For example, a drug might be known for many years as having a beneficial effect in a specific disease condition but the mechanism might be totally unknown. Uncovering the mechanism of action of such a well-known drug in improving a specific condition is a discovery and, of course, comes with and after experimentation and demonstration.


  9. meshugena313 Says:

    Right, but does “dis-covery” have to include “uncovering the mechanism of action”? I’m not sure my implication above that “demonstration” = “descriptive” while “discovery” = “mechanism of action” is true.

    DM, not sure that there’s any real daylight between discovery and demonstration.


  10. eeke Says:

    And why is “descriptive” a bad thing???

    In fact, if you perform some sort of ‘omics experiment, doesn’t that qualify as descriptive work?


  11. Dave Says:

    The descriptive thing is becoming ridiculous to be honest. You see it all the time, especially in human studies (even when interventions are performed).

    I am intrigued by this notion that “demonstrating” an effect is the same thing as being descriptive. Which we all know is bad. Unless ChiP/SLIP/N/DiP is involved. or ‘omics. or something

    Blimey. Change the record. Descriptive critiques actually became a lot more popular during the qRT-PCR and (especially) microarray era, and is applied frequently to studies using NGS techniques as well.


  12. drugmonkey Says:

    “Descriptive” is the “too many notes” of the modern bioscience critique.


  13. meshugena313 Says:

    “Descriptive” as a critique drives me up the fucking wall. Isn’t our job to describe how nature works? It’s now a freakin pejorative, and an easy way to dismiss something. That makes no sense.

    I have never once used the term in any review I have done, and will never use it.


  14. These terms are meaningless per se, and only gain meaning in a particular specific experimental context. So your question is underdetermined. Please reformulate.


  15. drugmonkey Says:

    It’s a simple question PP. Do you use either, neither or both terms in reference to your own work?


  16. I recommend everybody interested in the question to stop whatever they are doing and go out and read the late, great David Hull’s book “Science as a Process”. The whole point of the book is to point out that some scientists make a big deal about how revolutionary their work is and others don’t. And that there is no real objective distinction between the real novelty of the work in question.


  17. gingerest Says:

    Over here in the nebulous world of observational studies of human health and behavior, only arrogant people say they discover things. We mere mortals find things, and if we find something we think we understand, it demonstrates or shows that. Oddly, whether a finding is novel or not has nothing to do with whether we understand it, just whether anyone else has thought to ask about it. Novel findings are good, because they mean there’s a knowledge gap, which you can now fill with a paper, which will be high-profile if you can figure out a way it will translate, and otherwise you’ll shop it around until you find someone willing to publish it. Exceptionally novel findings aren’t publishable because everyone thinks you probably did something wrong.

    Non-novel findings only get published in high-profile journals by senior people who are owed favors.

    “Descriptive” is a term of art for epidemiologists, and means a paper that isn’t hypothesis-driven, much less a trial, but that needs to be published because so little is understood about the group of people or the disease or the behavior you’re studying that the findings are guaranteed to be novel. Or it’s the first chapter of someone’s thesis and has to be published separately because otherwise they won’t get three papers out of their study and also this way they can cite their first paper in the next two.

    All bets are off for RCTs, but that comes back to that arrogance thing and also the fact that an awful lot of the time someone running an RCT is trying to sell you something.


  18. MTomasson Says:

    I can’t help but think this is an example of what Wittgenstein meant…most philosphical “problems” are simply the use of vague language.


  19. MTomasson Says:



  20. Geologist Says:

    Most of my work is finding new things in the environment, that we didn’t know existed before. So we think of them as discoveries. But when I cross over and do interdisciplinary work then often what we are doing is demonstrating how some process works.


  21. Emaderton3 Says:

    In my experience in basic science research (cell biology), descriptive gets you scorn and criticism. Mechanism gets you favorable grant reviews and funding.


  22. dsks Says:

    Papers I’m on have usually been about discovering which of our brilliant and profound hypotheses in re protein structure and function live to fight another day and which ones are demonstrably horse shit… and then finishing it all off with some vague platitudes about the potential for alleviating cancer/pain/heart disease/hemorrhoids/sectarian strife/civil war &c.


  23. Ola Says:

    Why all the hate on “descriptive”? When used appropriately it’s a perfectly good criticism. As a grant reviewer for NIH, the delineation occurs at the level of “what the fuck does this mean for a patient suffering from the disease this study section is focused on?” (because study sections are almost always disease-centric).

    Put differently: suppose Aim 1 is to characterize some new pathway, then in Aim 2 you’ll determine its role in process X, then in Aim 3 you play with a drug that hits the pathway in a relevant disease model. Without Aim 3, “descriptive” is right on the money. Aims 1 and 2 do very little for anyone suffering from the disease.

    Given the above, I think a better word that should perhaps be used instead of descriptive is “passive”. To what extent will the research simply examine from afar, versus getting in there and messing with the system (i.e. active). Passive vs. active is a huge determinant of the “significance” score on a grant review, and to be frank the NIH much prefers active.


  24. dsks Says:

    Ola, that’s not what people usually mean when using the term”descriptive” . It’s generally used – slightly pejoratively – to differentiate between pure serendipitous observational science and hypothesis-driven science. However, as has been hashed previously on this blog and others, that distinction is not always clearly defined.


  25. E-rock Says:

    Ola, then you get into the problem that Aim 3 presupposes the success of Aim 2. What if the pathway is not described in the way you thought it would … What is the point of Aim 3 in that case? It helps people because now we know that pathway is not involved process X and research can go on to something. You’re putting too much emphasis on predetermined positive results.


  26. meshugena313 Says:

    Ola – doesn’t something need to be “described” (and not in a passive sense, as at least in my particular bioworld, much intervention and technology is necessary to even get to the description) before it can be “actively” messed with? And what do you do with the gobs of observations, not publish them or stack them along with a full paper’s worth of “active” experiments?


  27. E-rock Says:

    To answer the question, I generally demonstrate. Occasionally I have described discoveries (unexpected, actually), and in grants said that in the process of demonstrating X, if Y results from it, it would be a significant discovery, Z. Even so, demonstrating X is still important. No idea how this plays in the minds of reviewers.


  28. rxnm Says:

    To me, “discovered” means you observed a thing or process that had not been observed before.

    “Demonstrate” (or it’s low-IF cousin “show”) implies you are proving a hypothesis, but is mostly meaningless.

    Elucidate, find, reveal, test, explore, investigate…. they are all empty place holders with respect to actual data. You report the results of experiments, period. Everything else is window dressing and gamesmanship. I think most people use these terms interchangeably in rotation.


  29. Dr Becca Says:

    “Descriptive” is an utter bullshit critique used only by people privileged enough in their niche to have come in with all the basics worked out for them already.


  30. drugmonkey Says:

    So you are saying “mechanism” is the uninteresting, self-referential and masturbatory endgame when all the actual interesting stuff has been “discovered”?

    You may have a point.


  31. rxnm Says:

    “Mechanism” is 5 parts prestige, 1 part data, and 3 parts bullshit in the abstract and discussion.


  32. RunningViolet Says:

    I use “discover” judiciously and in reference to previously unreported biological phenomena. But that doesn’t have anything to do with mechanism, which is really what I think most reviewers care about. For mechanism, I tend to “probe” and “interrogate” cells and molecules in order to “demonstrate” how the phenomenon occurs and how it can be manipulated. But going back to “discovery” I think that’s equally important because it is the novel part of the stuff I do.


  33. drugmonkey Says:

    How does one interrogate something that cannot talk?


  34. Chris Says:

    We tend to use “develop” and “design” more commonly than “discover” or “demonstrate”. Probably field-specific – I’m in chemistry.


  35. qaz Says:

    In my experience, the key difference is that “demonstrate” is something engineers do (*) while “discover” is something scientists do. In science, we are setting out to test something – to see if it occurs or not – to see if it is true or not. Both hypothesis testing and descriptive exploration are “discoveries” not “demonstrations”, because in both cases, we don’t know what we’re going to find. In engineering, we are trying to make something work from principles we think we understand.

    Science journals don’t like demonstrations because they are not discoveries. So I make all my students change “we show that X occurs” to “we found that X occurs”.

    * I think it is important to recognize that building a successful demonstration is neither trivial nor simple.


  36. geo Says:

    Interesting Discussion. Creativity is the Driving Force, and with this Approach, you ought to do all-right –

    Insight > Hypothesis > Experiments > Initial Validation> More Experiments with Rigorous Testing of Hypothesis > Initial Confirmation of Hypothesis > Mechanistic Testing of Hypothesis > Initial Validation of Hypothesis > Alternative Testing of Hypothesis for Independent Confirmation of Hypothesis > Confirmation of Discovery.

    No Descriptive Approach Involved. Mechanism Trumps Everything Else except Initial Hypothesis.

    If Discovery is of Importance, than Publish at the Highest Level Possible.

    Fame and Fortune. Credit for Discovery.


  37. lionel Says:

    sad, sad, sad


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: