sample size
After reading numerous articles in which IMPLAN is used, it is not clear if there are any guidelines about sample size. I see samples ranging from 10 to thousands. Is there a recommended minimum sample size? For example, I have expenditure data drawn from a survey of 209 people. Is this acceptable? Too small? Any guidance would be appreciated, thanks.

Hello Dick, Unfortunately, regarding sample size, there is no one ‘ideal’ percentage, but the target percentage will increase the smaller the total population of attendees. For example, if only 100 people attend, then 10% is only 10 people and would not likely be very representative – you’d probably want to aim for at least 20% coverage. But if 1,000 people attend, then 10% is 100 people and would be a good sample. If the sample size is small, it would be a good idea to try to get a variety of gender and age. Hopefully this helps. Please let us know if you have any additional questions. IMPLAN Support Team 
Thanks for the reply. So larger percentages would be ideal for purposes of external validity. My sample is a little more than 2%, which is a very common response rate in surveys of my population. Do I infer from your reply that there is no minimum (or ideal) sample size for the purposes of IMPLAN analysis or the software? [quote="IMPLAN Forum" post=17347]Hello Dick, Unfortunately, regarding sample size, there is no one ‘ideal’ percentage, but the target percentage will increase the smaller the total population of attendees. For example, if only 100 people attend, then 10% is only 10 people and would not likely be very representative – you’d probably want to aim for at least 20% coverage. But if 1,000 people attend, then 10% is 100 people and would be a good sample. If the sample size is small, it would be a good idea to try to get a variety of gender and age. Hopefully this helps. Please let us know if you have any additional questions. IMPLAN Support Team[/quote] 
Hello, Sample size required is based on the variation in the responses. If the expenses spent by visitors are similar, then not many samples are required to have confidence in the results. If expenditures vary widely than a much larger sample is required to have confidence in the results. Since you don't know ahead of time for your particular application you will need to find similar studies to use as proxies. We suggest searching the internet to find proxy studies. Some universities have departments that specialize in surveys. We hope this helps. Let us know if you have any further questions.
Please sign in to leave a comment.
Comments
3 comments