Since there's only one standard deviation for a normal curve, we can use z score to solve this problem.
From the question, we know two probabilities for 56 and 92.
For scores
For scores > 92, the probability = 80/500 = 0.16
The z score for the upper tail with an area of 0.16 is 0.9945
The z score for the lower tail with an area of 0.02 is -2.0537
The formula for z = (score - mean score) / standard deviation
Let's consider two cases here, mean is 87 and 80
Case1: Mean=80 (92-80) / standard deviation = 0.9945
(92-80) / standard deviation = 0.9945, standard deviation = 12.06,
(56-80) / standard deviation = -2.0537, standard deviation = 11.68,
Round to the integer. Both are 12 so the standard deviation is 12.
Case2: Mean=87
(92-87) / standard deviation = 0.9945, standard deviation = 5.027652
(56-87) / standard deviation = -2.0537, standard deviation = 15.09471
Those two deviations are way different. From those two numbers we also know that mean has to be smaller than 87 in order to make standard deviation to be
very close.
The answer is A.
This is my explanation and hope it makes sense to you.