Use the chain pattern or model → investigable question → variables → reasoned prediction or hypothesis → evidence. An investigable question names measurable variables and conditions. Questions such as “Should we…?” or “Is this good?” involve values and cannot be answered by experimental measurements alone.
A correlation means variables change together; causation means changing one produces a change in the other. Controlled investigations help test causation, while a confounding variable provides another possible explanation.
A prediction states an expected outcome. A hypothesis proposes a testable, directional relationship and a reason: If [independent variable changes], then [dependent variable changes], because [scientific reason].
A model may be a diagram, physical object, mathematical relationship or simulation. It can generate questions and predictions even though it simplifies reality; its predictions must still be compared with evidence.