Decision tree analysis is a structured, systematic method used in complex decision making problems. It consists of a diagram of the decisions, external events, and likely outcomes involved in a decision.
The article provides a schematic of the basic parts of a decision tree diagram:
In this diagram, the squares represent the possible decisions, with the extending lines representing the opposing options at the point of decision. Circles represent the external events, or points of uncertainty. The lines extending from circles indicate external, uncontrollable events and/or circumstances. The analyst can choose to list a probability above these lines to indicate the likelihood of the outcomes.
The article recommends assigning a quantitative measurement of the perceived benefit of each outcome (desirability). Once all of the outcomes are listed and contain a measure of value, it is time to evaluate the decision. Add the benefit measures of the end outcomes that trace back to a particular choice via a specific path. The preferred pathway is that which results in the highest level of desirability.
In the event where a tree has more than one decision point, it is best to calculate the latest stages first. "Identify the choice that gives the highest overall desirability and leave only that branch (removing the decision point). Do the same with the remaining squares, working your way to the left (to the first decision point in the sequence)."