ACT® · Science: Evaluating Scientific Arguments and Models with Evidence (8-13 items)
Identifying the Implications of a Model (EMI 301)
An implication of a model is a prediction the model forces to be true even though it never says it directly. You find it by taking the model's stated cause-and-effect one logical step further, into a case or measurement the model didn't mention.
Every model is built from a few stated claims, but those claims drag other things along with them whether the scientist says so or not. If a model says warmer water speeds up algae decay, and decay eats oxygen, it is also committed to a pattern in time: fish kills should show up more in the hottest weeks, because that's when the heat-driven decay runs longest. Nobody wrote that sentence in the model. It falls out of the model anyway. That's the implication, and it's the part you can actually go check with data the model never used to build itself.
To pull out an implication, restate the model's cause as a mechanism, then ask what else that mechanism would touch. A model about a crater's dust blocking sunlight and cooling the planet is a claim about reptiles dying off. But sunlight loss doesn't know or care about reptiles specifically. It should hit anything else that depends on sunlight too, like marine plankton. Predicting a plankton die-off in the same window is new ground, not a rephrasing. A restatement just repackages the model's own words; an implication reaches a case or consequence the model is silent about.
This matters most when a model throws off two implications that disagree. Warming both dries soil through evaporation and adds moisture to the air for rain. Both are implications of the same warming claim, but they push the forecast in opposite directions. When that happens, the model isn't broken, it's just unfinished: it needs a further measurement to tell which effect wins. This is exactly where the unit's evidence-testing skill picks up, since judging a model means following it to its implications and then seeing which one the data actually supports.
Key terms
- Implication
- A prediction that must follow from a model's claims even though the model never states it outright.
- Restatement
- A rephrasing of what the model already says, with no new predictive content.
- Mechanism
- The cause-and-effect process a model proposes, which you extend one step to find implications.
- Competing implications
- Two predictions from the same model that point in opposite directions, leaving the outcome undecided without more data.
Watch it explained
- How it worksHow do you follow a model one step past what it states to what it therefore has to predict?
- How it worksWalk me through turning a claim about a cause into a prediction somebody could go out and check.
- Common mistakeWhy isn't an implication of a model just a restatement of the model in different words?
- Going furtherWhat happens when two implications of one model end up pointing in opposite directions?
Check yourself
End of topic.