I used to have a big concern for a little statistical mistake in my thesis (which wont be published), and now I have overcome that fear.

During that time of fear, my parents told me that all the thesis have mistakes. Even teachers, friends, everyone told me that thesis aren't perfect.

what do you think? How can a reader verify that a mistake was a honest mistake and not manipulation on purpose?

  • 4
    A perfect thesis is a done thesis. – StrongBad Jan 29 at 20:32
  • The question in the title is very different than the question in the last paragraph. What specific question do you want answered? – JiK Jan 29 at 20:33
  • I prefer the last one – Carlos Varas Tello Jan 29 at 20:34

Yes, many theses have mistakes. Some even have intentional errors. How a given work is judged depends on many factors, too numerous to list. If it was an honest answer then some rework will lead to a more valid result. If that valid result is essentially the same as what you wrote, it will almost certainly be dismissed as an oversight.

But you can't change the past. It is what it is. You could, perhaps, produce a publication based on your thesis work in which you cite the thesis and correct the error.

Ultimately a reader can accept your work as acceptable if not perfect or reject it. You have no control over that. Someone who tries to reproduce your results and comes up with very different results will probably write about it in their own publication. You should be the one to do this instead.

  • How can someone try to reproduce my results if the data is froma company and the thesis will not be published? – Carlos Varas Tello Jan 29 at 21:00
  • 1
    From the description of the problem and your results. This assumes it is "seen" by people. If it is never seen there is no issue whatever as long as your advisor and the university has accepted it. – Buffy Jan 29 at 21:02
  • My fear on those dark days was that maybe someday someone will have access to thesis and try to revoke my degree. – Carlos Varas Tello Jan 29 at 21:03
  • Not. Going. To. Happen. It is incredibly rare. But think about a follow up publication. – Buffy Jan 29 at 21:05
  • What is a follow up publication? – Carlos Varas Tello Jan 29 at 21:43

To add to what Buffy said, perhaps, if you can't publish the company specific data reanalysis, you could do some paper on the general flaw. (e.g. "regressing presmoothed data artificially inflates the apparent rsq") Or even a different dataset that you analyze properly and then put some warning about how a mistake could be made and then cite*/correct the other paper (even in a footnote...you don't even need to redisplay the data for this sort of "rowback" correction). But at least you are still getting SOME correction out there.

*And yes, you can cite unpublished work.

  • I dont want to publish my work, I just want to have a cool life. – Carlos Varas Tello Jan 30 at 0:11
  • It's probably fine. If it bugs you a lot, then figure out some method to put a correction out there (like Buffy suggests). If it bugs you less (than the work of doing a publication to sneak the rowback in) than leave it as is. You're probably fine. Nobody will read it. I bet even your company doesn't rely on it. And gawd knows there are a lot of company techreports with mistakes in them. – guest Jan 30 at 0:16

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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