• Question: How do you react when your work does not prove your hypothesis?

    Asked by Gremblin to Angela, Claire, Ian, Robert, Sarah on 10 Nov 2014.
    • Photo: Angela Stokes

      Angela Stokes answered on 10 Nov 2014:


      Hi Gremlin

      Another great question from you. In my case if a hypothesis has not been proved it probably means that the clinical trial I am waiting for the results from has not been successful, that is that the new medicine has not been as effective as another treatment or as successful as we had hoped.

      It is quite sad when this happens because it means that we haven’t been able to find a new medicine to help people overcome their illness, but we can’t allow ourselves to give up when this happens as patients rely on us to help them, so we generally have a look at the data to see whether there is a small section of the people who have taken the medicine who have had some benefit.

      We also use the results to look again at how the medicine was supposed to work and try and find out why it didn’t so we gather more information to help us move forward with our knowledge of how the medicine works.

      So even when we have a negative result we can get some positive information from it.

    • Photo: Ian Cade

      Ian Cade answered on 10 Nov 2014:


      As long as the ‘negative result’ is fairly clear cut, such as:

      A + B does not give C it instead gives D and a little E

      then no problem at all. A result even negative one is still a result that tells you something about the system you are working with.

      Alternatively if the result is a little more ambiguous, such as:

      A + B looks as if it gives C as expected… but not all the data supports the assignment “product = C” and sometimes the procedure doesn’t work at all.

      then that is a bit frustrating! And you have to carefully try to trace the problem. Is it bad reagents, bad solvent, dirty glassware, a change that you haven’t noticed (maybe the first time you did the reaction you weren’t being very careful and let some air in and the reaction actually needs a little oxygen?). Or are you misinterpreting the data and putting too much weight on the data that supports your preferred outcome and ignoring the contrary evidence…

      In my most recent brush with this type of thing, the problem was:

      A + B does not give C
      instead A + B gives D and E
      … but a 1 to 1 mixture of D and E spectroscopically look very much like what C was predicted to look like… damn and blast it all!

      This misinterpreted result wasn’t detected for months and inspired us to pursue a new line of research… Fortunately, although the initial result was bogus, it turned out that the related systems that it had inspired did work in the way we thought the initial system did.

Comments