ABSTRACT
How can a researcher determine reliability? The classic model is to measure twice and then check to see if the two sets of measurements are consistent with each other.
First, consider interobserver reliability. Suppose a researcher wants to test the hypothesis that tall individuals are waited on more quickly in retail stores than short individuals. To do this, the researcher could observe unobtrusively1 to (1) classify each customer as tall, medium, or short; and then (2) count the number of seconds from the time each customer enters the store to the time a salesperson greets the customer.
