darthtire Posted February 20, 2011 Report Share Posted February 20, 2011 Hi guys,I was just wondering when to know when the data you have collected experimentally is normally distributed. Is there a method that exists, to show whether it is in Gaussian Distribution?EDIT: I have 4 groups , each with 5 trials -- if that helps at allIn my IA i want to use the t-test, however the condition is that the data has to be normally distributed, soo I don't know where to goThanks! Reply Link to post Share on other sites More sharing options...
dessskris Posted February 21, 2011 Report Share Posted February 21, 2011 what t-test? what kind of data do you have? I think you are thinking too much Let's say you investigate the change in temperature over time..then just plot a graph of temp against time.Same with statistical data. What do you mean by normally distributed data? Reply Link to post Share on other sites More sharing options...
Keel Posted February 21, 2011 Report Share Posted February 21, 2011 (edited) Yes I would have to agree with Desy about you thinking too much. You seem to be overlapping your HL maths too much with your science (unless of course your task was indeed to prove that the data obtained would be normally distributed, but I don’t see why your teacher would give such an assignment). Even if you were to use the t-distribution you would need to convert the variance of the sample Sn2 into the unbiased estimator of the variance [(n/n-1) Sn2], the mean of the sample would be the same as the estimated mean and you would go from there to find the t value. (so long as you know the actual mean as well)To answer your question of how you can tell whether the data follows a normal distribution, I don’t think you can as your data is discrete whereas the distribution is continuous, so I think it’s an issue of trial and error. Like finding a graph which fits all the points. So it would only be an assumption. Edited February 21, 2011 by Keel Reply Link to post Share on other sites More sharing options...
kiwi.at.heart Posted February 28, 2011 Report Share Posted February 28, 2011 Unfortunately you cant do a t-test because it only works when there are two groups of trials involved, and even then you need at least 10 trials of each group. Don't try to over complicate your IA. Generally the mathematics needed for IAs should be capable by someone studying Maths studies so don't over-think it and ask your teacher how much mathematical analysis you really need. Reply Link to post Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.