When is enough, enough? A question of sample size

Michael Cramphorn, Add+Impact, suggests that we should not get over-obsessed with sample size in qualitative research

Michael Cramphorn

At one time people generally believed that unless you measured everything in a total census, you never could be really sure that what seemed to be, really was. Eventually some of them, (gourmets? cooks?) must have asked themselves questions such as ‘why is it that we can sample just a bit of food or wine and conclude that the rest is the same?’ The answer was, of course, that the wine and food had both been thoroughly mixed up and, unless you bit on a chili, any part of it tasted like any other.

The underlying principle is that the bigger the variation in what you are sampling the more tasting you require. When you have sampled enough to get the full experience, there is no need for further sampling unless something changes. In fact, further sampling adds nothing to the quality of understanding; it simply uses up resources and could make you drunk (or fat).