Complex problems can have a lot of moving parts, and it is easy to lose track of some of those. This why our working memory is a strong limiter in resolving complex problems. One way to address this limitation is to use a hypothesis map.
A hypothesis map is a graphical breakdown of a hypothesis. It uses argument mapping conventions (for a primer, read Austhink’s excellent tutorial on argument mapping) to analyze the thinking behind the hypothesis and to support this thinking by capturing the relevant evidence.
To illustrate the approach, let’s consider a concrete example. The 2013 Tour de France finished just a few days ago, and Christopher Froome won it with a commanding lead. So commanding, in fact, that many question whether he rode ‘clean’ or resorted to using performance-enhancing drugs. So let’s put this in a diagnostic issue tree.
Important note: the attached case is a work in progress. At this stage, I am not concluding whether Mr. Froome doped or not.
We break down the stem question in three branches and, eventually, map the problem with three hypotheses. The second hypothesis is the one addressing the potential doping situation: “Christopher Froome won the 2013 TdF because doping gave him the edge to beat his competition.” So let’s focus on that one.
To establish that Chris Froome doped, we need to establish that -1- our body of evidence supports this hypothesis and -2- our body of evidence doesn’t contradict this hypothesis.
Note that, to confirm this hypothesis, both arguments must be true: they are a set of necessary and sufficient conditions to support the hypothesis. A reasonable doubt on either of those should be enough for us to discard the overall hypothesis.
Let’s look at the first branch, about whether Chris Froome was willing to dope to win the Tour. There are three types of arguments for this claim: some in favor (reasons), some against (objections), and some against objections (rebuttals).
Reasons start with “because”. For instance, “he was willing to dope because [the rewards are high enough to offset the downsides]”. We capture reasons in green.
Objections start with “but”: “he was willing to dope but [doping is unethical and Froome doesn’t do unethical things].” This lead us to the implicit conclusion: therefore he didn’t dope. We capture objections in red.
The third type of element is the rebuttal (in orange), which are objections to objections.
Drilling down further, we meet another important element: the “and/or” connection. Whereas a “and” connects two or more co-premises, “and/or” fit between independent arguments. Let’s look at examples to explain the difference.
“We think that Chris Froome doped to win the 2013 TdF because: [he was willing to do it] and [he was able to do it].” Here, the two elements are co-premises: both are necessary conditions; prove that one is wrong and the whole argument collapses.
In contrast consider: “winning the TdF is highly rewarding because [winning the TdF earns a lot of money] and/or [winning the TdF earns a lot of prestige].” We identify two types of rewards associated with winning the Tour; even if one is not true, the argument has support from the other.
Next, use your hypothesis map to capture evidence. That the winner of the TdF earns €450k supports my claim that winning the TdF earns a lot of money. Generalizing, evidence is useful to support claims (reasons, objections, and rebuttals). All branches in your map should have some supporting evidence, even if some say ‘accepted as self evident’.
In the figure above, we consider that the claim ‘doping is unethical’ is self evident.
Finally, capture your synthesis for the branch.
Considering the evidence, my synthesis is that the reason presented is indeed valid, so I write it down in a box and put a check mark next to it.
You should go through this mechanic for all branches in your hypothesis map. I’ve linked below the latest version of my hypothesis map for this case. Again, I am not claiming that Mr. Froome either doped or didn’t because my analysis is not finished.
The take away: Hypotheses maps are a good way to capture a diagnostic analysis because of various reasons:
- They help ensure that the analysis is complete (CE)
- They help ensure that there are no redundancies in the analysis (ME)
- Likewise, in the heat of discussion, they help ensure that the various stakeholders don’t get stuck analyzing the same part of the problem over and over or forget some other parts
- They help ensure that people make up their mind based on the same material (consistency)
- They constitute a central repository where, at a glance, one can get all the information available on the case