Subscribe Bookmark RSS Feed

quantile box plots

outlier

Community Trekker

Joined:

Jun 23, 2011

Anyone know why the quantiles box plot is seemingly different depending on the platform used.? E.g. for the distrbution platform the end of the whiskers are 0% and 100% (i.e. min and max values). For Fit Y by X the quantile box plot ends are 10 ad 90%. Then of course there is the graph builder platform. Going back to the distribution platform the 2nd lines are 2.5 and 97.5%, however I have sometimes seen these as 10% and 90%.

Can anyone shed any light onto the underlying logic behind these differences..?

Another question: is there an easy way to add confidence interval (like mean diamonds) to the median values on quantile box plot..? Normally for my data I expect the median to be the most meaningful number (with obvious outliers filtered out). Therefore, for comparing groups I would like to focus on median values with some confidence interval added.