-
Asked by jordanclough96 to Anne, Carolyn, Joe, Mariana, Nick on 15 Mar 2010 in Categories: Test methods.
Question: how many times would you do a test to make sure you dont get any anomolies
-
Nick Bradshaw answered on 15 Mar 2010:
It depends on how reliable the test is, but as a rule I don’t believe a result until I have seen it three times.
-
-
Joseph Devlin answered on 15 Mar 2010:
When we run experiments, they typically involve between 12-24 participants and the results are based on the group as a whole. When we find something entirely new, we always run the experiment again with new participants to make sure it wasn’t just a weird chance. If we still are worried about the results, we’d typically run it again. But it’s best when other labs run a version of the experiment and find similar results — those are more independent and the replication is more meaningful in some ways.
-
Carolyn McGettigan answered on 16 Mar 2010:
There are some limits on how many times we can repeat an MRI study, because of the time and expense needed to run experiments. Also, the research journals don’t tend to want to publish studies that are just replications of previous work. These are definite disadvantages in the way the field operates, in my opinion.
However, usually each study in a programme of research will be linked to the previous one, in terms of the methods and the materials we use. So, we can look for consistencies across our experiments. We also do a lot of reading of published research from other laboratories to see how our results compare with similar studies they have done.
-
Mariana Vargas answered on 16 Mar 2010:
The number of repetitions you do for an experiment depends a lot on how variable your measurements are.
For the recordings I do, 10 repetitions of the same experiment under the same conditions is a good number.
Comments