You can assume that the data is normally distributed and then plot the data with this assumption.
With probit scale, this is just a straight line - without any structure.
But in real life, what you want to do with a cumulative distribution is actually the opposite of this:
check for deviations. Without having the data at hand, this is not possible.
