Imagine you want to measure soft drink preferences unobtrusively. How would you do this? Think about this before you go on reading. You probably thought about sitting in a restaurant or school...

Imagine you want to measure soft drink preferences unobtrusively. How would you do this? Think about this before you go on reading.

You probably thought about sitting in a restaurant or school cafeteria and counting the number of people who bought certain soft drinks. But did you consider that not all soft drinks may be available in this particular location? Did you consider the time of day (since some people may prefer caffeinated soft drinks but will avoid them in the evening, or some people who don’t normally choose caffeinated drinks might choose them in the late afternoon when their energy level is low)? What about other factors that could affect your measurements? Thinking about this now, do you trust your original method to give accurate results?