The Causal Gap: Which Alternative Must Disappear?
When an argument says that X caused Y, ask: What else could have produced Y? If a rival explanation remains, the conclusion is not secure. The best strengthener makes that rival unavailable.
Run this procedure:
- Mark the proposed cause and effect. Write the argument’s causal claim as “X → Y.”
- Name the strongest alternative. Ask what could have produced Y even if X had no effect. Look for a preexisting difference, a simultaneous change, selection into groups, or coincidence.
- Predict the needed evidence. The best strengthener will show that the alternative was absent, equal across groups, or unable to account for the result.
- Test every choice against that alternative. Do not ask merely whether a choice makes X sound beneficial. Ask whether it removes a specific competing cause of Y.
The goal is to reduce credible routes to Y, not collect favorable facts about X. If people who received X also differed from everyone else in another relevant way, evidence about X’s general reputation will not eliminate that difference. Evidence that groups were alike before X directly attacks the rival.
Define “eliminating” precisely. The answer need not prove that no other explanation is imaginable. It must remove the serious alternative that the argument leaves open. A choice can therefore suffice without establishing the causal conclusion with certainty.
The Rival-Cause Test: Evidence That Closes the Escape Route
Consider a common pattern:
- Group 1 received X and later had more Y.
- Group 2 did not receive X and had less Y.
- Therefore, X caused Y.
The obvious alternative is that Group 1 already had some feature that made Y more likely. For example, motivated people may have been more likely both to choose X and to achieve Y. It strengthens the conclusion to show motivation was comparable before the groups diverged.
Other alternatives require different repairs:
- Different starting conditions: Show the groups were equivalent before X.
- Self-selection: Show assignment to X did not depend on traits related to Y.
- A simultaneous intervention: Show no other relevant change occurred only for the X group.
- A time-based coincidence: Show Y increased after X, not before it or independently of it.
Do not overreach. If the argument concerns whether X caused Y in one population, evidence that X causes a different outcome elsewhere may not remove the local rival. Likewise, a choice showing that X was implemented correctly may confirm that X occurred without showing that nothing else caused Y.
A useful internal sentence is:
“If this choice is true, the strongest remaining explanation for Y would be ______.”
If you cannot fill that blank with the original rival, the choice probably does not perform the needed causal work.
One Case, One Rival: Find the Choice That Removes It
A city compared two neighborhoods. Residents of Northside received free air-purifying plants, while residents of Southside did not. Six months later, Northside reported fewer respiratory complaints than Southside. The city concluded that the plants reduced respiratory complaints.
Practice
Which one of the following, if true, most strengthens the city’s conclusion?
The forced-choice discipline matters here. (C) is not merely a favorable fact about plants. It changes the comparison: if the neighborhoods began alike on the outcome and its obvious environmental cause, the strongest escape route for the causal conclusion is closed.
High-Yield Points - ⚡ Biggest Takeaways
Identify the strongest rival cause of the observed effect. Prefer evidence showing the groups began alike or were assigned independently of the outcome. Reject facts that praise the cause without removing the rival explanation.
Unlock the full lesson and continue reading
Signup to continue reading this lesson and unlimited access questions, flashcards, AI notes, and more