We don't have standard deviation, so we need to calculate it from the simple statistic given: p=0.68 and 1-p=0.32. The sample size, n, is 34099. Standard error replaces standard deviation and is equal to sqrt(p(1-p)/n)=sqrt(0.68*0.32/34099)=0.002526. From the confidence level of 90%, we can calculate a=1-0.9=0.1. The critical probability is 1-a/2=0.95. The z-score associated with this is 1.645. So margin of error=1.645*0.002526=0.0041556.
So this gives p=0.68+0.0041556 or 67.584%<68%<68.416%.