Gary's survey has two additional outliers, 0 and 20, not in Susan's, so the question is asking about the effect of these outliers. going for easiest to hardest:
Range--Note that with a mean of 3.5 and range of 10 there could not have been a value greater than 10 in the Susan's survey. So the range in Gary's survey has to be increased to 20 because that is how far apart the two outliers are.
Median--This is unchanged because one outlier is less than the original median and the other is greater, leaving the middle value of 4 the same.
Mean--A quick way of answering this is to consider that we are adding two more numbers to the original total. So the new mean will be shifted in the direction of the average of those two numbers. Since the average of 0 and 20 is 10 and that is greater than the original 3.5, then the new average will be greater. A more difficult but accurate way to find this answer is by actually calculating the new average. If the mean of 10 samples is 3.5 then the total of their values would have to be 35, because working backwards we get 3.5 x 10 = 35. Now add the two new data points of 0 and 20 we get a new total of 55 for 12 samples which is a new mean of 55/12 = 4.58. that average is much greater than the original 3.5.
