PhD issues

Post Reply New Topic

So I have competed a PhD in science at a major research university and am doing a post doc that is going much better than it was 3 weeks ago.

However, I am wondering what to do if it turns out that something P produced for my thesis is not 100 % reproducible when other scientists try and run the same code and/or if it turns out I made computing mistakes that may change my thesis and how valid it was.

Basically, the project involved taking a highly complex code written by a computer scientist, without comments or clear organization, comprehending ti and restructuring it to make various computations for analysis. It is possible that a fit value I calculated was not being done 100 % correctly because the code needed to be rewritten in a way that neither I nor the computer scientist who wrote it were aware of at the time and so I would need to make the necessary adjustments. There's a good chance this is not the case but it is possible. Also, for one condition, I got a fit value of 1.45 after a number of runs and someone else may get repeated runs of something like 1.75, 1.69 or 1.58, in part because the procedure uses semi random processes.

I am worried about the possibility to either discovering something I did not realize as a graduate student and needed to fix it in later publications, or having someone else rerun the procedure and get somewhat different values as I mentioned above and this being used as grounds to revoke my PhD. It is also possible that since the procedure had about 50 or so steps that I may have accidentally overlooked 1-2 of the 50 steps being done a certain way that may result in slightly different values when someone tries to replicate what I did.

There was no plagiarism anywhere in my thesis and no made up data. All the runs in the thesis were actually done and all the fit values and functions used in the thesis were authentically generated and there was no doctoring the results by artificially making them lower or replacing a given data set with a lower value to make it look better. In fact the PI even expressed some concern at one point about the values i was getting but I maintained that these are the true values and did not doctor it at all.

In light of this how plausible would it be to have a PhD revoked on grounds of the university being angry about the fact that I did not realize something was missing in the code the computer scientist sent or being angry that my results are not 100 % duplicable? Should I really be thinking about how to address this?

If you just made a mistake and didn't commit fraud or plagiarism, then no, you certainly don't have anything to worry about at all. Your committee accepted it, that's good enough. If you plan to do additional publishing, then you can just publish a correct result, that's all.

Yep.....I agree with Darmok

If there were serious concerns, I would wager that they would have been bought up before and during your defense.

I would reckon that if the thesis passed the examination and there was no fraud or plagiarism and your advisors okayed it and so on, it should be okay

Have you discussed the concerns with your doctoral supervisor?

The fear of failure of 100% replication can be set aside. Science is not about being 100% right all the time in every detail, (if it were, then replication would not even be required!). It is about honest and careful inquiry carried out in a scrupulous fashion in good faith. Human error manifests in science in many ways - wrong theory, wrong method, wrong analysis, wrong conclusions - the history of science is absolutely littered with these. Then there are (a lot more these days it seems) errors of omission and commission. Cheating in science is a big issue now, even journals such as Nature have been sucked in. If you know that you have completed your doctorate with integrity to the best of your ability at that time and to the satisfaction of the doctoral committee, then let matters take their course. If someone produces findings that are in contradiction to yours, treat that as new information to be considered, rather than evidence of some personal failure. Science as you know is accumulative and a collective endeavour; each work forms part of a greater whole in any discipline, and even studies that are completely wrong in their conclusions may nevertheless contribute something that is useful.

Suspicions are raised in science sometimes when the 100% comes up just too often to be credible - it is rarely possible to replicate experiments without any degree of difference and different samples introduce more variance however scrupulous the design and method. It may be that you are perfectionistic by nature as many HFAs in academia are, and maybe accepting that will help you see this in a wider perspective rather than as something to attach anxiety to.

Celebrate your achievement and give yourself credit for it.

Out of interest what was the original code?

_________________
( If I ignore a reply it's not intentional I get distracted, send me a PM to prompt me )