Appreciating A/B testing with a real-world example

In the last section of this chapter, let's talk about A/B testing. Unlike previous topics, A/B testing is a very general concept. A/B testing is something of a geeky engineer's word for statistical hypothesis testing. At the most basic level, it simply means a way of finding out which setting or treatment performs better in a single-variable experiment. Most A/B testing can be classified as a simple Randomized Controlled Trial (RCT). What randomized control means will be clear soon.

Let's take a real-world example: a consulting company proposes a new working-hours schedule for a factory, claiming that the new schedule will improve the workers' efficiency as well as their satisfaction. The cost of abruptly shifting the working-hours schedule may be big and the factory does not want the risk involved. Therefore, the consulting company proposes an A/B test. Consultants propose selecting two groups of workers, group...