1 - due 10/4/99

Article :http://old.seattletimes.com/text/2018321244.html

The authors defined getting worse as a change (in blood pressure, fasting insulin, cholesterol or triglycerides) that was more than 2 standard deviations above zero (or below, whichever was bad). For all of the measures (except fasting insulin) the mean change was about zero. Assuming that the changes are normally distributed, what percent of individuals would you expect to get worse on any given measure just due to random variability alone?

For this question, I am particularly confused on how to use the "2 Standard Deviations" to find the percentage of individuals that would "get worse" simply from random variability. Could you clarify on how to approach this problem? Thanks