front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |review

Sample Size for Surveys - Comparisons 

If the objective is a comparison, i.e. a statistical test, then you must specify the level of errors you can tolerate and the size in absolute units of the smallest difference worth detecting.  This is the smallest value that can be credibly attributed to an effect and not to noise, or just random variation. i.e. a 5% in a Prevalence, or a 20% in an Odds Ratio.