It is extremely difficult to get a truly
representative sample without doing a very time- and cost-intensive
random selection of phone numbers or addresses. Survey
companies can do this, but it is really not necessary.
Your group can be confident in your data and stand by
your claims if you sample as many different types of people
within the parameters of the group you are looking at
and by having enough people fill out the survey. To capture
the attention of media and policymakers, a survey of at
least 500 people is considered solid; having 500 to 1,000
completed surveys is even better.