Doing an online, mail-in or telephone survey?
Thanks to the convenience and low cost of the many online tools now available to radio programmers, the task that once was the exclusive territory of people who specialized in research is now being performed week in and week out as we send out surveys to samples which may or may not be representative of the group of people we want to listen more.
Statisticians say that one response in 2,000 "WILL" be wacko, no matter how perfect the sample is!
That's where a very arcane calculation - standard deviation from the mean - comes in. It can't tell you if your entire sample was not the one you wanted to target, but if you know that the vast majority of your data comes from folks in the right age group, the correct geographical locations, a good balance of gender and qualitative profiles, it can help you find the outliers who can greatly skew your results.
Hand check each of your survey responses before tabulating them by computer. All it takes is one such form to throw otherwise very good data WAY OFF.
You actually may get more actionable and reliable research from a smaller, more consistent, cleaner sample.
Tale of Two Bar Graphs: Country Shares and Music Scores - If you want to see just how long the current ‘up’ ratings cycle has been going on for country, this just-released Nielsen bar chart and This Week In Music...
4 hours ago