Close Home Forum Sign up / Log in

Negative results...

T

Hello,
Has anyone on here successfully written their thesis and defended their viva with lots of "negative" results? It feels very disheartening and would just be encouraging to hear that it can be done.
Tudor

P

What do you mean by negative? As in - not as expected? Or you've done it wrong?

S

It was the case of the project of the PhD student I was hired to help. Basically the results did not support the hypothesis of the mutated protein not binding the RNA, because it obviously bound it, just not as much. Instead of being bent on the original hypothesis, we looked at what else might cause the differences that resulted in the biological effect, and we found that it has different affinity to motors (we already had a clue that this is either something in binding or in transport). Sometimes when you get stuck on the original hypothesis you miss real findings. It does not have to be as we wanted it to be. I'd look at what the "negative results" tell you, because every experiments tells you something, and imagine in your head how the process might go. Also remember that you set out to disprove hypotheses, not to prove them.

T

People in my department call it "negative results" or "negative findings". I don't like the term myself - as it doesn't quite make sense. Basically - yes, you didn't find what you were expecting - or you didn't find anything "significant" - either statistically or clinically. It happens a fair bit in my field... interventions etc are not effective when you expected them to be. Has anyone here had a thesis that basically didn't report any significant results? It would just be encouraging to hear some accounts of the research still being appreciated, and successfully defended, in the viva.

S

When I was a bit younger, PLOS one tended to accept those as far as they were scientifically sound. This is a pain in the community, because we're not supposed to find exactly what we think, otherwise it wouldn't be called scientific research. Punishing it would be like actively punishing fairness. I've seen theses without papers, too, but I'm based in Poland, so I don't know about elsewhere.
That said, you can take those negative findings and as I said, build on it. Just like you'd follow the positive findings. Because they always tell you something, especially if they stand in contrary with previous research, and I don't believe that every single of your experiments didn't tell you anything.

T

Yes, this is great, and I agree with all you've said. What I would like to hear though is some experiences of where people have been successful defending their work in the viva when it was a bit disappointing / uninspiring in the results department. Thankfully my first study has some "significant" findings. But the latter two are so boring (although I do appreciate that there are things to be learnt from them - especially for myself as a developing researcher). I'm just freaking out at it a bit.

S

If this is the case, it's probably defendable, even if you don't get accolades for it. Don't worry. Ask your advisor about it. If you're worried, it might be good to sprinkle the research with a well-thought fancy method or two, and perhaps something even shows up on the radar then. They understand it's science and it doesn't always result in breakthrough. Write it like a proper researcher, perhaps a beautiful thesis on negative results will be enough. Not all my experiments worked as well - the medical part was mostly negative, and people appreciate that I don't point in wrong directions.

T

Cool! Yeh, I think so too. Think I just wanted to hear it said to me. And thankfully, I don't have any problem writing things well, so that part should go OK. I'll certainly see what my supervisor says when we next meet. She is encouraging but realistic, which is good. What do you mean "people appreciate that I don't point in wrong directions"?

T

There is no need to chase the elusive 'p<0.05' (particularly as p-values are meaningless anyway and such chasing typically ends up in the dubious practice of outcome-switching).

A thesis of 'negative' results is perfectly fine! It also makes for a more interesting discussion.

T

Thanks Thesisfun! It just feels like everything is worthless without that darn p value! I completely disagree with that, but it seems the culture. Your words are extremely encouraging; thank you! : ) : )

C

The quants side of my work was a bit of a mixed bag - couple of significant things and some non-significant. The non-significant stuff was quite useful in terms of generating suggested tweaks to the models I was testing and further research ideas, and to be honest, it was hardly mentioned in the viva.

Hi Tudor, what methodologies are you using (generally speaking), quantitative and spss or are you also using qualitative? And are you in social or physical sciences, or a professional discipline? I ask this as these variations on qualitative methods provide room for different interpretations of results.

Social sciences and disciplines such as education, seem to allow for no significant results, provided these lead to a position or clarification on something. My first study (I am in education but I used social science methodologies) had a disappointingly small sample size-so though there were results, from the sample of 34, I could not really claim that results were indicative of a broader population or amounted to much of anything that could be considered statistically significant. I used this as an exploratory data chapter, which set the scene for the two other studies. The second study had a bigger base of 178 participants-was both qual and quant, and the third study was purely qualitative and the sample size was appropriate for the analysis and methods used.

In my thesis, I used the first study, as outlined, as a scene setter or initial point of reference and I referred back to this chapter in the same way in the analysis chapter, which followed the 3 data chapters. Having said that, both externals noted that this first data chapter was the least interesting in the thesis, but did not really comment on it other than that. None of my required corrections were anything to do with data, data analysis, methods or results.

T

Thanks for sharing Pjlu. Mine is purely quantitative and would fall under social sciences. I am doing an "alternative format" thesis, which will comprise an intro, three "papers" (or at least written up manuscripts in paper format), and a discussion. I've finished collecting data, so I can't really go back and do more based on what I've done already. There are limitations such as sample size (this is often a limitation in my field so not overly worried - I recruited as much as I could) and participant characteristics not being as I had hoped, but those are not really things I cannot do anything about.

T

Ps. Your sample sizes are impressive!

Thanks Tudor. The thing I always remember from my Masters supervisor was her saying "your thesis tells the story of your data". So is it possible to think of it in this way? That even though results in some parts were negative, that these negative data indicated " .... " in the overall story or thesis?

54060