Question: How whould you measure whether you were doing a good job in an experiment?

Asked by awesomekid to Posty, Edward, Kate, Nathan on 22 Jun 2011.

0

  1. Photo: Kate PattersonKate Patterson
    answered on 21 Jun 2011:

    dear awesomekid,

    That’s a really great question. Every time we do an experiment we make sure we include ‘controls’. These we know should give certain answers because they have been tested many times before. So, if the controls don’t do what they should in an experiment, then we know the experiment has not worked very well.

    An example that I do in the lab is when I am growing cells. Let’s say I treat them with a chemical that I think should kill the cells. I will always grow an extra dish of cells as a control that I don’t treat with the chemical. I expect that dish of cells not to die. If they do… then I know something has gone wrong and I can’t trust the experiment.

    0

  2. Photo: Edward ShinEdward Shin
    answered on 22 Jun 2011:

    Hi awesomekid,

    This could be two questions.
    1. In a school science experiment, you know you’ve probably done a good job if you get the result that you’re supposed to get.

    2. In experiments done by “real scientists”, all you can do is to be very careful, as Kate says. This is because you don’t know, when you start the experiment, what the result will be.

    Note that in particle physics experiments, where you smash things together very hard and see what happens (see this quick explanation: http://simple.wikipedia.org/wiki/Large_Hadron_Collider), there are no ‘controls’ as such. This is very different to biology.

    In biology, you want to do the same experiment (e.g. does a drug kill off this disease?) 50 times, and if in 48 cases X happens and in 2 cases Y happens, what’s usually interesting is that ‘X happens most of the time’ (e.g. the drug is good at killing off this disease).

    In particle physics, you do the same experiment millions of times, and what’s interesting is the one time that something different happens. This is because particle physics is like throwing dice: if you do X, you can produce either result Y or result Z. It’s just chance. But you want to know if Z CAN happen, because if it CAN (even if it is rare), that can mean that a thirty-year-old theory (that says X causes Y) is not complete, and “new science” has been discovered. (It’s not so much the fact that X sometimes causes Z that is interesting: it is the fact that Z can happen at all.)

    0

Comments

  1. Photo: awesomekidawesomekid
    commented on 22 Jun 2011:

    thank you kate and edward you both have explained that very well 🙂

    0