error margin
Message posted by mitchy on September 30, 1999 at 12:00 AM (ET)

am i right:to start the sampling, i have to give a sample size then compute for margin error and the confidence level?the ans. are for the whole study, e.g., plus/minus 5, 95% confidence level for 400 sample. if i have 20 questions, all categorical/nominal data.can i specify for each data analyzed based on each question its own confidence level and margin error?Ex:1. are you aware of this certain ad? a.yes, b.no. Result:75% say yes, 25% say no, significant using chi-square for goodness of fit test, 95% confident (based on n=400). (is it right to use chisquare for frequency?)shld i use this for every question?

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: error
Message posted by marivic on December 4, 1999 at 12:00 AM (ET)