What percentage of my conrol or test group must respond for the results to have statistical accuracy?

Question

I have a total of 60 program participants. Is there a formula that I can use? Please reference your source.

Answer 1

What you are really looking for is statistical power. Statistical accuracy is not really what statistics is about as much as it is relates to the likelihood that your results are from a real effect rather than random chance.

  

The percentage you need is driven by the effect size you expect and the "p" value you are trying to reach. There are a lot of calculators to help with this which are a little easier than calculating statistical power by hand.

You can use this link and calculator to help work out how to  address your effect size issue and figure the response rate you need.

Best of luck...