2.19 Detecting a Change in Level with Variability

Just as when you were estimating the level of one phase, variability makes it a little more difficult to compare two phases. And, just like when you were looking at one phase with more variability, it can be helpful to imagine a band that covers most of the data points within each phase. Look at this graph. Let’s say that it depicts the number of minutes a girl rides her bike every day. In baseline, most of the data points are around 15, even though the highest is around 17 and the lowest is around 10.