If i am not mistaken, the mean of all readings is your zero line, how much your highest and lowest readings differ from this are then used to calculate a second "mean"(You could consider the accuracy of the voltmeter at this point also, make a third "mean" and reduce that and the second to a fourth to replace the second and third) The percentage of readings that are different to the first mean by the magnitude of the fourth, or third had you not considered the accuracy factor,give a measure of the uncertainty.