[Marketing Research]Determining Sample Size for Online Surveys

How to determine sample size for online surveys varies hugely.

Factors on this decision include budget constraints, sample availability, time available, etc. Obviously a sample of 1000 individuals will produce more accurate survey results versus a sample of only 100 people. This is simply because generally, a larger group of people will more accurately reflect the attitudes of an even larger group, simply because there is more individual variety, and this therefore will yield a larger variance in answers.

Now, admittedly this is a general and over-simplified answer. Complicated statistics on sample sizes, confidence intervals, and statistical validity on how all of these relate would better explain how this all works.

A similar question was posted on Infosurv's website, and the following response was provided:
"To achieve statistically valid results, we recommend that a statistically significant sample of n=400 completed surveys be obtained for online market research surveys. This will provide statistical validity at the 95% confidence level with +/- 5% confidence interval. For clients on a tighter budget, we sometimes recommend a sample of n=200 completed surveys which provides statistical validity at the 95% confidence level with +/- 7% confidence interval. In any event, a reliable sample source is very important, as is an unbiased random sampling methodology." http://www.infosurv.com/blog/2005/01/market-research-survey-questions.html#comments

"n" in this case simply represents the number of people comprising the sample.

As you can see then, there needs to be a decent sized sample in order to eliminate as much bias as possible. This does not mean that it is necessary to conduct a survey with 100,000 participants in order to get accurate answers.

What does this mean for you? When there is a chance to win cash or product as a survey participant, keep in mind that your chances of winning are decent. The chances of winning the jackpot in a national lottery are usually more than 1 in 10,000,000, and you must purchase a ticket, which involves a cost. You could very well be offered a chance to win a plasma tv worth $5000 for 15 minutes of your time, when there are only a few hundred other participants. What are you waiting for?