By default, the y-scale of a histogram represents frequency (each bar represents the frequency of values within the specified bin), emphasizing the magnitude of each bin. If your audience doesn't have enough process knowledge to understand the frequency values, you can change the graph's y-scale type to recast these frequency values as percentages (each bar represents the percentage of all values within the bin), a format they may find more meaningful.
By default each bar represents the frequency of values within the bin. Change the y-scale type to Percent to make each bar represent the percentage of all values within the bin. Use Density when you want to compare distributions and the sample size differs. Density is also useful when you compare bars and the bin widths are unequal. Density is calculated as the proportion of observations divided by the bin width.
Accumulate values across bins: (Frequency and percent scales only) The bar heights accumulate from left to right. The height of each bar is equal the height of the bin plus all the previous bins.
Values on the y-axis represent estimated cumulative percentages. The estimated cumulative percentage is equal to the estimated cumulative probability multiplied by 100.
Values on the y-axis represent estimated cumulative probabilities. The cumulative probability for a value x is the probability that a random observation that is taken from the population will be less than or equal to x.
Minitab uses the median rank method (also called the Benard method) to estimate the cumulative probability (r) for each observation:
In this formula, i is the rank of the observation in the sample and n is the total number of observations in the sample. For the smallest value in the sample, i = 1 and for the largest value in the sample, i = n.
Values on the y-axis represent inverse cumulative probabilities.
The score values for the normal distribution and the lognormal distribution are the inverse cumulative probability of r, calculated using the standard normal distribution.
The score values for the exponential distribution and the Weibull distribution are calculated as LN(−LN(1−r)), where LN is the natural log function.