In statistics, how can I test if two ranges of data are significantly different?

Question

Please illustrate with an example and give an internet ref, if possible. Thanks

Answer 1

Here is a page showing which test is appropriate for what:

http://www.socr.ucla.edu/Applets.dir/ChoiceOfTest.html

Answer 2

You will have to get creative here....no standard test that I am aware of will answer this question. How about if you focus on comparing the minimum values and maximum values rather than the range per se? How about if you pretend one group's minumum is the true population value, and then test to see if the other group's minimum would be an outlier (maybe via Grubb's test http://www.itl.nist.gov/div898/handbook/eda/section3/eda35h.htm)

Or, rather than testing the ranges, what if you test for differences in dispersion via a test comparing standard deviations?

Answer 3

Box and whisker plots maybe? Then compare whether the ratio of Range to IQR is significantly different?

Answer 4

although i'm no longer a statistician, i'm an epidemiologist and function studied information. regrettably, taking the mid-selection because of the fact the mean is an particularly undesirable assumption. in reality, you haven't any concept what the distribution of values is between the bounds. it would desire to be exceedingly skewed. Assumptions relating to the SD are additionally deceptive. for occasion, on your n=10 occasion, the authentic mean might desire to be something from 2.2 to 3.8. And on your n=12 occasion, the authentic mean might desire to be 3.25 to 5.seventy 5. you purely don't be attentive to. i don't be attentive to of any attempt for such constrained information. I agree which you're able to no longer settle for conclusions drawn from inadequate information such as you have defined.

Answer 5

You would probably have to derive it from a particular distribution using the order statistics. I am not aware of any such hypothesis test.

Answer 6

One-way ANOVA