I have several data sets, each of which includes several tests. Each test contains several temperature samples at different locations. For each data set, I have an average of the different tests for each location and some measurement of error (standard deviation, confidence interval, etc.).
I want to produce a bar graph where bars are grouped by sample location, so there will be as many bars per group as there are data sets, and as many groups as there are sample locations. The number of tests in each data set doesn't matter, as I am not using test data, I am using the averages of test data in each data set. This is, so far, fairly straightforward.
However, I also want to graph the error measurement as error bars (whiskers extending up and down from each vertical bar, or whiskers extending right and left from each horizontal bar, depending on whether I choose a column or bar graph type in Excel).
These error bars should not be calculated by Excel from the data series it is given. Instead, they should be taken from the error measurement series I have prepared along with the average from the test data in each data set. I'm not sure how to accomplish this.