I understand margin of error for a simple 'yes/no' answer.  But if we
are comparing mean dollars on a survey, and we are trying to compare
the results of two different means, how do we calculate a margin of
error that is plus/minus 7%?
One of the values is $16.00 and the other is $12.00.  The first value
is an independent answer.  The second is the sum of two subsets of the
first.  I want to know the margin of error and am told it is
plus/minus $7.00.  Is this correct?