ACT® · Science: Evaluating Scientific Arguments and Models with Evidence (8-13 items)
Complex Hypotheses, Predictions and Conclusions (EMI 601)
A complex claim with words like only or always, or with two parts joined by and, is only as strong as its weakest piece. Check each piece against the data separately before deciding if the claim survives.
Words like only and always are absolute. They claim something happens in every case, or in no other case, which means a single counterexample in the data breaks them. A claim that a drug works only above 50 mg dies the moment the table shows an effect at 30 mg. You don't need to study the whole table, just find the one number that falls outside the boundary the claim sets.
Two-part conclusions, usually joined by and, need splitting into two separate claims before you touch the data. Test each half on its own evidence. One half can hold while the other fails, and the honest answer reports that split rather than forcing a single verdict. A claim that an additive raises yield and shortens reaction time might survive on yield, rising from say 40 to 55 g, while reaction time actually rises instead of falling. That's a half-true conclusion, and the correct answer says so directly instead of rounding it up to fully true or fully false.
This connects to the model-versus-measurement question that sits right next to it. A model can predict a number, say 310 MPa for alloy strength at the fastest cooling rate, but if the measured value at that rate reads 260 MPa, the measurement wins. Models are built on assumptions that can be wrong; measured data is what actually happened. The same discipline applies throughout this category: read the claim carefully, isolate its parts, and let the data arbitrate each one.
Key terms
- Absolute qualifier
- A word like only or always that makes a claim apply to every case, so one exception disproves it.
- Two-part conclusion
- A claim joined by and that makes two separate assertions, each needing its own check against the data.
- Model
- A predicted relationship or calculated value derived from a theory, which must yield to contrary measurements.
- Measurement
- An actual recorded data value from an experiment, which overrides a model's prediction when the two disagree.
Watch it explained
- How it worksHow do you split a two-part conclusion into halves and test each half against the data separately?
- Worked exampleFind whether a claim that a drug works only above 50 mg survives results showing an effect at 30 mg.
- Going furtherWhat changes when a complex conclusion is backed by a model but contradicted by the measurements?
Check yourself
End of topic.