 
 For relative frequencies, if variable f is the relative frequency and the total relative frequency is 1, you can estimate the mean using sum(f * m)
 The estimate we obtain from these data differs somewhat from the mean of the original set of 30 values  it would be closer if the original set was larger.
This formula uses the midpoints as the best estimate of each class interval's location  which, aside from the problem of sample size, depends upon how those class intervals were originally decided upon.
