Made a huge mistake in my bachelor's thesis

Question

I recently submitted my thesis and also defended it successfully. But then I found mistakes in the statistical analyses of my research.

When I was doing my bachelor's there was no biostatistics in my curriculum. So I didn't have a grasp of statistics at that time. That is why during my research my co-supervisor did the statistical analysis of the data. I started doing an online course on biostatistics, and decided to check the data of my thesis. And I found that my supervisor had made mistakes while doing statistical analyses.

After this discovery I contacted my supervisors and they told me that, "you shouldn't worry because nobody is going to read your thesis." But the real problem is that I am going to apply for masters at different universities, and the admission committee of universities will definitely ask for my thesis. What if they read my thesis? Should I make corrections in the soft copy or not?

Answer 1

First of all, are you sure that the analysis is wrong? Was it confirmed by your supervisors when you reached out? In general, it is more likely that you cosupervisor knows more than you do after a single online course.

Second, are you sure that the admission committees will ask for a copy of your thesis? Personally, I was asked for a copy of my thesis when I applied for a PhD, but not for master programs.

If the thesis contains a significant error and you are asked to provide it, you have two choices:

1. In the thesis I have seen, the student is usually expected to have done all the work (under supervision). If that is not the case, as in your situation, it may be good to include a statement of contributions, where you clearly explain who has done what. By stating that somebody else has done the statistical analysis, you can hope/expect that the errors do not reflect poorly on you (again, assuming that the thesis is indeed flawed, that you are asked to provide it, and that somebody reads in detail enough to spot the error). Including such a statement, if not present already, might be a good idea even if you go for the second option.
2. You fix the error, and when you are asked for your thesis, you provide the fixed version mentioning that you are sending them a revised copy that fixes the mistakes you have found after submission and incorporating the feedback you have received.

Answer 2

I'd be very surprised if anyone asks to read your undergrad thesis for PhD submission. But if they did I would add an erratum, or a postscript spelling out what happened and what the effect was.

To me following things up like this, admitting the mistake, and discussing and outlining the consequences would be a massive plus in the column of any PhD applicant.

Answer 3

First, relax. Take a deep breath. Take another. And relax again.

Second, as others have said, it is pretty unlikely that anyone on an admissions committee will ask for your thesis or hunt it down. It's even more unlikely that it will be read with anything more than a cursory skim, and vanishingly unlikely that someone will try to replicate the statistical analysis from the data you provide. People just don't have that kind of time on their hands.

So unless it is a truly howling error, visible from a brief skim, you're probably okay on that account. (There are types of statistical error that obvious-- using a blatantly wrong type of test, for instance, so it is not impossible. But someone does have to be really engaged as a reader to spot these things.)

Third, remember, you are talking about an undergraduate thesis, not a published journal article. You've done your best, which was good enough to satisfy whoever assigns grades. I will not defend bad statistical analysis, but I will absolutely defend a person who makes a good faith error at the unpublished undergraduate level.

Fourth, even if this were a published article, you would unfortunately be a statistic yourself:

Of the 157 articles published in radiology journals, 10 had no statistical errors, while 147 had at least one statistical error. The most frequently encountered error was “errors in summarizing data” with a rate of 66%. This was followed by “incorrect representation of P values” with a rate of 42%. The least frequently encountered error was “statistical symbol errors” with a rate of 3%. There was no statistically significant difference according to impact factors.

Radiology journals! Rife with statistical errors!

Finally, what I would advise you to do is to redo the analysis on your own, and hang on to the results. It's the right thing to do. It builds character. It builds an ethos of learning from your mistakes, after which you can resolve at least never to make that particular mistake again. If necessary, it gives you something to talk about if an admissions committee does ask you about it. (And some interviewers like to ask questions such as, "Tell me about a learning experience you had," or "Tell me about a difficult experience or challenge you had." This is a great anecdote for that situation.)

And not coincidentally, it will make you feel better.