ABSTRACT
The core notion of reliability is simple: The measurement instruments applied to observations must be consistent over time, place, coder, and circumstance. As with all measurement, one must be certain that one’s measuring stick does not develop distortions. If, for example, one had to measure day-to-day changes in someone’s height, would a metal yardstick or one made of rubber be better? Clearly the rubber yardstick’s own length would be more likely to vary with the temperature and humidity of the day the measure was taken and with the measurer’s pull on the yardstick. Indeed, a biased measurer might stretch the rubber yardstick. Similarly, if one wanted to measure minority presence in television commercials, as described in chapter 2 (this volume), one would find different results by using an untrained coder’s assessment or by using trained coders with explicit coding instructions.
