5

Is it just to check that the paper's results are correct? Should I expect that the authors describe their methods for arriving at the right answers / solving a problem? I've often found that on a key part of the paper, e.g., the main result, the authors just give you the data without explaining. I then have to verify that the data is correct and the results are correct but this takes a bit of work on my part, which I am happy to do, but I wonder why these steps aren't shown on the paper itself.

Is what I described the usual expectation when reading a research paper? That I shouldn't expect researchers to divulge how they arrived at the right answer and perhaps that is a way for them to protect their work, e.g., whatever method they came up with that worked ... could have the potential to be used again and again to publish more papers?

I also tried contacting researchers directly to discuss their work. All of them are happy to talk - but mostly on a superficial level, e.g., "I've done X,Y, and Z, my co-author has done A,B and C."

  • Generally, a failure to clearly demonstrate the correctness of results is a weakness in a paper, and shouldn't be tolerated, but sometimes is. And of course, what is clear to one reader may not be clear to another. But if you could say what your field is, it might be possible to give more specific answers. – Nate Eldredge Jan 19 '17 at 21:07
  • 2
    "That I shouldn't expect researchers to divulge how they arrived at the right answer": many people -- I'm one of them -- arrive at the right answer through ways that are really insignificant and are not worth publishing. In addition, some of us can't even remember how they arrived to the right idea. – Massimo Ortolano Jan 19 '17 at 21:19
  • 1
    I left that part out of my question but that is certainly a possibility which I have discussed with my advisor, too, @massimoortolano. Considering this possibility too, these papers don't seem useful; there's no method / explanations that can be extracted from the paper, studied, and generalized / tweaked to solve the bigger problem, which is my (naive?) expectation when reading a paper. – Michelle Jan 19 '17 at 21:30
  • 3
    What area are you working in? I think the answers to your questions could differ quite a lot depending on that. – Dan Romik Jan 20 '17 at 2:52
  • 1
    Why are you reading the paper? Are you a reviewer? Are you interested primarily in the statements of the results, or in the high-level techniques, the low-level technical details, or the definitions, or the pictures? Are you interested in the paper because you want to apply it, because you want to teach it, because you're going to be tested on it, or because you want to extend it? – JeffE Jan 21 '17 at 16:00
3

I think there are two uses of the word "expectation" relevant to your question:

1) Expectation as in "standards": this would be the type of "expectations" a teacher might have of students: that they show up to class, complete assignments, read course materials, etc. This is the aspirational sense of the word.

2) Expectation as in "predictions of what is going to happen": a more realistic, perhaps Bayesian understanding of what publications are like in your field as you read more papers.

I don't think you are wrong to have high expectations in the first sense, especially from work published in high tier journals in your field. You should expect results to be presented clearly, empirical and statistical methodology to be detailed, and conclusions drawn to be clearly connected to the actual results and not just the predictions and world view of the authors.

In reality, your expectations in the second sense should not be as lofty. You should expect abstracts to not really tell you what limited scope the paper actually addresses You should expect to need to dig into not just the paper in hand but previous papers by the same authors and others they reference obliquely to figure out their methodology. You should expect to find conclusions that are suggested by but not shown convincingly by the data at hand. You should expect to find statistical approaches that are naive, inappropriate, or misleading.

In my personal experience, you will find authors in your field whose work you really cherish because they meet your Expectations. These won't necessarily be the most famous papers, but you will enjoy and learn from most of the publications they write. The other papers that instead conform to your expectations might be just as important, you will be held just as responsible to understand them and incorporate them into your work, but they won't be as fun, you will wonder how they ever got past peer review, and you might even start to doubt your field in general.

Don't fret! All you can really do is try to meet your own expectations.

0

Let's say that you decide to do some work on the same or a related problem, and that you write a paper about your work. Near the beginning, you get to say:

Previous work by Terse and Succinct addressed such-and-so aspect of my-topic, but their results are not reproducible, because they did not clearly outline their approach.

And then make sure you clearly outline your approach.

I understand your frustration... but do you see that in a way, their terseness makes it easier for you to get your work published?

  • I'm not sure whether saying something like this at the beginning of a paper makes the paper easier to publish in practice. I agree that it should, wherever it's correct. – darij grinberg Jan 22 '17 at 7:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.