A message posted from Jay Ratnayake ([log in to unmask]):
Dear All,
Would you/your colleagues be able to enlighten me regarding the
following? I received a mix of comments and suggestions.
A mail survey was conducted among people who contracted a vector-borne
disease between Feb and July 2002 (total cases for the period = 350;
total
cases for the entire year = 877). Although target population was 350
only
224 gave the consent to take part in the survey. Of 224 only 201
returned
the completed survey form.
Q1. Can I treat this data (201 cases) as population data?
Q2. If it is a sample, what should I do to minimise self-selection bias?
Q3. Can I use past data to stratify this information and select a random
sample from the existing database?
Thanking in advance,
Jay
|