Here you should use Chebyshev's inequality

Compute the z-score for $14.80: z=(14.80-25)/(10.20)= ??b. Assuming we can verify that the data set is approximately normally distributed, what percentage of time will the cost of a full-course meal be greater $14.80?

Now look the z-score up in a table of the cumulative normal.

+/-1 sigmac. Assuming we can verify that the data set is approximately normally distributed, 68% of the dinner prices will be within what range?

RonL