This post is written by Brian Flemming, a mathematician working as a Systems Engineer in Edinburgh. He completed an Engineering Doctorate (EngD) as a mature student at Heriot Watt University in 2014 and is now appreciating the freedom to continue studying and spend time away on the hills, without the associated “PhD-guilt” of neglecting the books. In this post Brian tells us about a situation we all dread: discovering a mistake in his thesis after it had been submitted….
It was a sickening moment. I was staring at the computer monitor in disbelief. My jaw sagged open at the realisation I’d made a mistake in my thesis, which I’d just handed in for examination only the day before. And not a trivial mistake either: anyone with a working knowledge of the subject would have spotted it straightaway.
It was a pure Adams-esque moment, of the “…accidently chang[ing] down from fourth to first instead of third thus making your engine leap out of your bonnet in a rather ugly mess…” variety.
I’d been preparing a short presentation of background material for the viva when I decided to check on a mathematical equation in a company report to which my thesis was partly related. I’d chosen not to use this report as a reference to avoid any awkwardness over access to material not in the public domain. The equation concerned was the realisation of a physical definition I’d quoted in my thesis. I checked a few other references for corroboration. The uncomfortable truth was there in stark black and white: I’d misinterpreted the definition.
My initial reaction was of mild panic. Not that those working around me noticed anything different, but inside I was quietly dying. After all that hard work making sure the thesis was as perfect as possible, and now this. It was if a large black hole had opened up in front of me, and there was nothing to stop me being sucked in.
My first thought was to try and recall the thesis. There was still a week to go before the deadline for submission. There was conceivably just enough time to correct the draft and resubmit. On the other hand, I was due to start a long-awaited holiday abroad in the next few days, which would effectively scupper any attempt at resubmission.
Missing the deadline would mean an extra six months delay in completing the degree, which I was unwilling to contemplate. I was already fed up with whole process of writing the thesis and wanted to get it out of the way as soon as possible.
Sheepishly I wrote an email explaining the situation to one of my academic supervisors. His reply was reassuring. “Don’t worry”, he said, “it is only a draft. Any mistakes will be dealt with at the viva, and corrections can be done afterwards.” (Editor’s note: Australian students note: you probably don’t have this option)The pressure was off temporarily.
Even so, I didn’t fancy the idea of undergoing a viva without having an answer to a thesis I knew was flawed.
I looked at the damage again with a fresh perspective. With great relief, it transpired that I had implicitly been using the correct definition after all, so that my results were still nominally correct. My thesis was concerned with an alternative approach to the standard method. To save time, I’d simply ploughed on ahead with the analysis: the introductory description had been added in later, which is when and where the mistake had occurred. It could just be a simple matter of correcting the error and moving on; on the other hand, this would also be a good opportunity to improve the argument by making a stronger connection between the standard definition and my alternative analysis.
Time was still of the essence though, and it would mean working through the holiday I’d planned with my long-suffering better-half as a celebration for my finishing the thesis. The books would, after all, be following me to France.
We agreed a compromise whence I’d work on our rest days between sightseeing trips. The total sum of the changes were two chapters swapped round, a new introduction drafted, conclusions revised, and follow-on effects traced throughout the remaining text.
The laptop glowed red hot crunching new data. It had been hard work but I’d done it, and the end result was a much stronger and more convincing argument. The question still to be answered though was why the mistake had happened in the first place.
Because my project involved a wide-ranging mixture of techniques spanning physics and statistics, my two academic supervisors were from the separate schools of mathematics and physical sciences respectively. I had spent considerable amounts of time writing up summaries of essential theory to aid inter-disciplinary understanding, in which differences in notation had also to be overcome.
Crucially, neither supervisor was particularly expert in the field in which the mistake had occurred, so that I was effectively policing my own understanding of the subject. I could remember distinctly the occasion when I’d checked the meaning of that particular definition. I trusted I knew what it meant. Unfortunately, I’d assumed the wrong interpretation and pushed on with the mountain of other work still to be done, unaware of the inadvertent discrepancy between the introduction and the rest of the analysis.
The viva took place approximately three months after the original draft had been submitted, during which time I’d continued to refine the replacement sections. It was a full discussion of the thesis, including my proposed changes. Having worked hard to overcome the flawed earlier submission I felt far more confident in discussing the material than I might otherwise have done. I passed: convincingly as it turned out. I’d demonstrated that I knew my stuff, and that is all I needed to do.
The moral of the story is that, no matter what the pressure, it pays to check and check and check yet again. Then check some more. In my case I’d let the pressure of work get to me instead of mentally setting aside the impending submission deadline, a tactic I’d found so effective in the past in maintaining a high standard of work. On the other hand, errors in execution are not necessarily fatal: the important point is your understanding of the subject matter, which includes knowing when you’ve made a mistake and what to do about it.
I should note, Brian had a viva to correct the record, Australian students will not have this opportunity. Have you had a near miss or PhD disaster? What did you do to fix it?
Four things you should know about choosing examiners for your PhD