Determine what kind of error bar you need: standard deviation, standard error, confidence interval, or measurement uncertainty
Collect the data values for each group or measurement
Calculate the mean if you need error bars around an average
Calculate the standard deviation: square each value’s deviation from the mean, average those squares, then take the square root
Calculate the standard error: divide the standard deviation by the square root of the sample size
Calculate a confidence interval: multiply the standard error by the appropriate critical value and add/subtract from the mean
For measurement uncertainty, use the known instrument precision or propagate uncertainties from the measurements
Plot the mean or value as the center point
Add the error bars using the computed upper and lower limits
Check whether the error bars should be symmetric or asymmetric based on the data or uncertainty method used