In past, the percent of smokers capable to quit smoking "cold turkey" was reported to be no more than 0.10 . If we wished to make new estimate of that percent that now were capable to quit "cold turkey" with percision (margin of error) of plus and minus 0.02 and with 99% confidence, how large sample would be required? What sample size would be needed to make that estimate in plus and minus 0.03 and with 95% confidence?