well today i achieved my first set of results. and they look pretty good with great error bars and replicates pretty good. except they dont agree with the few papers in the literature, so my sup decided i did the experiment wrong. except i have no reason to doubt the experiment didnt work. The controls were exactly the same as the previous trials and nothing in the protocol was changed. how do i say i think ive done it right its just that it doesnt prove his theory thats the problem (this has been a running theme in my project where if its not as he expects he says ive done it wrong)

With all due respect this is your first set of results, so you're unlikely to be in a position to overturn an accepted theory in the literature. That's why your supervisor is playing it safe and questioning whether you've conducted the experiments correctly. I agree that it can be demoralising if your supervisor criticises your work, but all new PhD students are apt to make mistakes, it's part of the learning process, so his scepticism is understandable.

If after several sets of results you're still finding differences then you'll be in a much stronger position to challenge the accepted wisdom. That's when the science starts because you'll have to come up with a reason to explain the differences. Good luck

when i say first set i mean from the data that will make up my thesis as in also have done several other experiments along a similar line and found the same thing, but as running out of time these are the first set that will go into thesis.. if that makes sense! in terms of literature there is only one group who has done this and a handful of papers.. am i so naive to think that their results may not be conclusive?

Apologies for misunderstanding your first post. Are these results key to your whole thesis? If so, what does your supervisor suggest you do?

Ultimately though, if you're confident with your results and can defend them when it comes to the viva, it doesn't really matter whether your supervisor agrees with you or not.

Hi PhDgirly

in early renaissance it was thought that blood circulated in the heart going from the left to the right. Scolars in that time were convinced that there had to be a "whole" in the heart making this possible. For science sake at demonstrations they showed a hole in the septum (which they had previously made to strongen their case).

Harvey calculated that the keep a system like that going one needed to have 180 liters of blood!

There is always a first person who discovers that previous thoughts may be wrong.

Just food for thought. No garantee that your figures are right!

true but then i have no reason to think they are incorrect as such when the numbers match the published papers, its just they dont indicate the same things.
ill just have to come up with a good defence for my viva then!

When this happened in our lab, the supervisor decied to repeat the experiment, using the student's protocol. He got exactly the same results, which convinced him that the student hadn't done anything wrong.

There's often variation: I remember a newspaper article not long ago where a journalist sent a cannabis sample to be analysed (for THC content) to 5 different labs, and each one sent back a different result. There's variation between machines, time of day (bright light), temperature, and sample handling.

i suggested this to him but he is unwilling. think i might go over his head and speak with someone else.

When you say replicates do you mean at least three inderpendant experiments? Because if not and you havent repeated it at all yet then that is prob the first thing to do. This would also add weight to the fact that you didnt do it wrong as you are less likely to do it wrong three times in a row (as long as you dont go into autopiolt).

I would suggest looking at how your protocol compares with the published work. Maybe there is something subtly different which gives different results (eg temperature, timings, buffers etc etc). I have heard of cases where someone has published work, others have repeated, couldnt and subsequently found out there was a contaminant responsible for the observed result in the published work.

What does your supervisor want you to do about it then? Has he provided construtive advice?

i have three biological replicates then three technical replicates also. the results i have obtained have been from two different types of samples, but each experiment giving data going against the puplished dataset. superivisor has been less than helpful basically saying it has to be wrong!

Doing it wrong three times seems unlikely (though if you think you know the protocol inside out it is possible to unconsiously do the same thing wrong each time).

Has your supervisor sat down and gone through the protocol with you? Atempted to find out where you may have gone wrong? If not, I dont think it is really fair to make you keep on repeating it (as what will you change/do differently). Really he needs to do this, as just saying you did it wrong is not really very constructive.

he helped me to optimise the method and thought it was fine but he seems to think im im doing something wrong in the mastermix but i dont see how!

well, I guess that if he can point out exactly what you are doing wrong then you should probably try and correct it and try again. However, if he is just saying you did it wrong now go away and work out what it is yourself thats not really very useful (esp as you know that you have done nothing wrong). Just repeating it a million times wont help. There should be sound scientific reason why it isnt working. If he cant provide that then you have a pretty good argument that there is nothing wrong with the protocol.

Does he suggest that everything you do is wrong? Or just things that go against what he wants to happen? If he has no reason to doubt your ability normally then he should have no reason when something doesnt work the way you want it to work. Maybe you should talk to someone else in your department who knows something of this field. See if they can spot any problems with your protocol.

just when they go against his theories! i tried the method using samples of known status and it worked fine.

Do you have a positive and negative control?
At this point your supervisor does need to give you help on analysing the results if he expects you to interpret it his way, as you've no time to repeat if you're writing up(?)
Maybe you should talk to someone else senior, to see if they can find a potential problem or explain how it still fits your supervisor's theory.
If they couldn't find a problem, maybe they could help you defend it to your supervisor.
maybe the route is to say this evidence disputes the theory that....but more investigation is needed e.g. x and y and factors could have influenced the result or experiments a and b should be done in the future to test this further. The experiments could be right, but just don't support for theory for some obscure reason and may be just be an exception. There could be some secret ingredient in your chemicals that throws off the result and makes it an exeption.