• Question: Do you have to try your experiments over and over again? If you do, how many times have you tried, when you've done your longest experiment?

    Asked by ambert2003 to Zach, Mark, Manolis, Elizabeth, Claire on 13 Mar 2014. This question was also asked by ratboy7.
    • Photo: Mark Wallace

      Mark Wallace answered on 13 Mar 2014:

      Yes! Science is all about trying new stuff. So you don’t ever really know if it’s going to work or not. Most of my ideas don’t work. Perhaps 1 in 20 work out.

      In terms of repeating the same experiment. It depends on how accurately you want to be able to say something about the result. If it’s not important, then only a few repeats will do. For a lot of my experiments you get a lot of ‘repeats’ for free, by measuring lots of individual molecules at a time. It’s not uncommon for us to collect data from 1000’s of molecules. So that’s 1000’s of repeats. But it only takes a few minutes!

    • Photo: Zach Dixon

      Zach Dixon answered on 13 Mar 2014:

      We always want to be sure that what we are seeing is a true result and so this means we have to see it in the same way as many times as possible!

      Sometimes getting these repeats is easy like Mark said, 1000s in just a minute! But other times getting the repeats is difficult, especially if doing one experiment takes a long time!

      Another way to see if our results are fair is by looking at the same thing using different experiments. If you get the same result using many different experiments it is more likely that it’s true!

    • Photo: Elizabeth Ratcliffe

      Elizabeth Ratcliffe answered on 13 Mar 2014:

      Another way we can look to see if our experiments are correct is to test lots of different combinations of the factors we are interested in, to build up a clearer picture of how these factors interact and see where the boundaries are i.e when does the experiment start to flop! This allows us to create whats called a design space – we can then do further smaller repeat experiments inside the design space. This helps if you have experiments where it takes a long time to get repeats or if you are trying to improve the experiment

    • Photo: Claire Vinten

      Claire Vinten answered on 17 Mar 2014:

      I’m lucky – because my experiments are trying to find out how people feel about certain topics, they can’t really go wrong! Anything that someone feels is important enough to say to me in an interview is data that I want to collect. The interviews can go off topic a bit, or finish really quickly which is not ideal. But generally I only have to do each one once.

Comments