…about chain graphs. You can get large, complicated graphs that are much more difficult to sort out! We’ll see that the problem is that, in general, you have to worry about bias when you’re trying to measure the total effect of one variable on another.
…tem?”. If you only observe how the system normally operates, you’ll generally get the wrong answer. For example, if you intervene to make sure your alarm never fails to go off (for example, by switching to a battery powered alarm clock), then you will underestimate the odds of being late to work. You’ll misattribute lateness due to traffic (which happens at the same time as your clock failing to go off!) as being due to your alarm clock, and so overestimate the effect of the alarm failing to go off.