# of survey respondents to represent general population?

[Smath]

I made an online survey regarding the use and preference of digital music, and I was wondering how many participants I would need to represent the general population. What I'm seeking with the results of the survey is to figure out what people want out of digital music formats (like MP3s) on their computers and portable media players (like iPods). For my original hypothesis, I used the sample size chart found in the Science Buddies article found here: https://www.sciencebuddies.org/science- ... ants.shtml . So far, I've gathered 113 respondents who have taken the survey, which, according to the sample size article, is enough for a confidence level of at least 90%.

You can find the online survey I made here at: http://www.surveymonkey.com/s/digitalaudiosurvey

[deleted-42343]

I'm glad you checked out the article.

The number you survey is going to depend on a few things:

1) Amount of time/money you have available

2) How accurate do YOU want the results to be?

You have to keep in mind that after you get all the results, you'll have to analyze them, which also takes time. You have a respectable number of participants now. If you have more time and want to make your results even more accurate you could send out more surveys. More data is always better, but at the same time you do have quite a lot now. It also says your project is due in 2 days, so it sounds like you don't have enough time. I would focus on writing up a good report and doing data analysis instead.

When companies send out surveys, they probably set a goal for their error margin. If they wanted a really low error (1% or less) they'd have to survey more than 10,000 people! That costs time and money. In research, most people aim for an error of +/- 5% or less (which would be 500 people in this case). The acceptable error margin probably depends on the industry/area of science, too.

I hope that answers your question!