• Question: How do u know that everything you find out isnt fake and how do you know it is reliable ??

    Asked by 07kearneyl to Vicki, Keith, Hywel, Emma, Alastair on 18 Jun 2010 in Categories: .
    • Photo: Alastair Sloan

      Alastair Sloan answered on 18 Jun 2010:

      Good question. You repeat your experiments 3 4 or 5 times and always have 2 or more controls

    • Photo: Keith Brain

      Keith Brain answered on 18 Jun 2010:

      Good question: we have to trust people most of the time, check them for ourselves some of the time, and look for internal consistency (see below). If you know a particular scientific field very well, then you get to know all the major research groups – over the years, you can build up an idea of how liable they are, and that will affect how much you trust them.

      Internal consistency is an important clue – in many scientific paper the same information can be presented in different ways, and so you can often cross-check some of the results by looking at other results and doing some calculations. Sometimes, you can also check for consistency of the results with well-establish theories or ideas … if there is conflict, then perhaps the theories/ideas are wrong … but then again, perhaps there is something “wrong” with the results …

    • Photo: Vicki Stevenson

      Vicki Stevenson answered on 18 Jun 2010:

      There’s two parts to the information I find out. The first is what I read from other scientists in journals. These papers are always assessed by other scientists before they are published, so they are pretty reliable.
      The other part is the information that I generate from experiments. It’s important to make sure that everything is calibrated first – so if I’m looking at temperatures I’ll test the sensors in ice water and in boiling water to make sure that the results are right. It also takes a bit of common sense – if you’ve made a calculation and the results don’t look right, I go back over my calculations and split them up and check each individual part – that usually works! If I was still concerned, I would go to a friend in my department and ask them to check over my work and see if they could find a problem.
      It’s always worth staying alert – equipment can breakdown and you wouldn’t want that to corrupt your data.

    • Photo: Emma Carter

      Emma Carter answered on 18 Jun 2010:

      That is a really good question. It’s just something you always have to bear in mind and make sure you’ve thought of everything you should have thought of (if you haven’t the peer reviewers will soon pick it up). It means we need to keep very good records of experiments and results and think about why we get the results we get. For example, I have to anaylse my results to find out if my device moved because of the Casimir force or because of some sneaky electrostatic forces that shouldn’t have been there but may have. For some experiments, scientists will have a ‘control’ so they have something to compare their study with. We also have to bring statistics into the mix sometimes so that we know if we have enough results for the study to be statistically significant. Hope that helps!

Comments