I think it means that you take the difference between the two people in each category, square it (to get rid of the positive or negative sign so that youre comparing apples to apples) and then compare the results across the population.
I think thats how you would do it for one parameter, but i'm not sure how you correlate the 4 aspects, unless you took the difference in all 4 areas, summed those and then squared the results.