Since, d_i = x_i – A,

where d_i = deviation and A = Assumed mean

And u_i = d_i/h = (x_i – A)/h,

where h = class width