Determining Sample Size

Determining Sample Size

I am often asked, “how many people do we need to interview”?   This usually prompts me to ask three questions:

1)  How confident do you need to be in your results?   Our confidence in our survey increases with the sample size. Usually this is expressed in terms of the margin of error. With a sample of 100 our results may vary by as much as +/- 10%.  With a sample of 1,000 this margin of error  is reduced to +/- 3.2%. However, there is a law of diminishing returns and at some point it is simply not worth the cash outlay to gain a few points in confidence.  So it really boils down to how crucial is the decision that you have to make.

2)  What is your budget?  Question #1 leads to our second question. If we can afford to splash out on a big survey than I say “go for it”.  However, the reality is, most research is done on a a fairly tight budget and this will usually dictate how large of a sample we can recruit.

3)  How many ways do you want to break down your results?  The desire to look at the results regionally or by age or gender will have an impact on our sample size.  The more ways that we want to breakdown the results, the larger the total sample we will need.

I usually end my discussion by saying, your sample size is important but don’t ignore the response rate.  This is the topic of my next blog.

 

| Return
  

Post a Comment

${xf.ViewCommentTextBox("Email",200,true,"email")}
${xf.ViewCommentTextBox("WebSite",200,false,"url")}
  • B2B Market Research