What influence does breakfast cereal have on a baby’s gender, or one’s astrological sign on susceptibility to heart failure? Both showed statistically significant links in scientific studies—yet both may be completely random findings. This may make your life easier.
Using these studies as illustration, a recent Wall Street Journal article explores how the traditional mathematical standard for statistical significance is not infallible; and that much depends on study design and sometimes repeated implementation.
The point for CSR practitioners is not that controlled experiments or statistics lack value. In fact, they can be extremely useful tools. But it’s a cautionary tale that the “ideal” approach to measurement isn’t always perfect, and may be even more time-consuming and costly to get right than anticipated.
So, what do you do if you don’t have the time or resources for such an effort? First, recognize this is par for the course. Your colleagues in other functional areas, such as sales, marketing, and recruiting have similar limitations. So, use the techniques they use: sampling, pilots, or estimates based on similar implementations from within (or outside) of your organization.
The trick is to develop sufficient, if not perfect, information. Remember, very few new product launches, advertising initiatives, or other highly consequential business decisions would be made if everyone waited until they had bullet-proof statistics in hand for guidance. Business can’t afford to work that way. And neither can CSR activities.
