Inference
An inference is a logical explanation or conclusion drawn from observations and prior knowledge, not something seen directly.
Definition
An inference is an idea you form to explain what you observed: you did not see it happen, but the clues point to it. It is a reasoned interpretation built from evidence and prior knowledge, and because it can be wrong it must be tested before it becomes a conclusion. In lab science, inferences connect measured data to mechanisms nobody can watch directly (molecular motion, friction, genetic factors), and they are constrained by controls, by alternative explanations, and by the limits of the data. The strongest inferences are the ones that survive deliberate attempts to rule them out.
Example
Bean plants in a closet are pale and bent toward the door crack; a student infers they are reaching for light. Salt dissolves in $30\ \text{s}$ at $20^\circ\text{C}$ but $12\ \text{s}$ at $50^\circ\text{C}$; she infers that faster-moving water molecules break the crystals apart more quickly, where the timing is the observation and the molecular story is the inference. A cart's measured speed runs about $8\%$ below the energy-conservation prediction, so a group infers that friction and wheel rotation absorb the difference, then strengthens the inference by repeating on a smoother surface and watching the gap shrink to $3\%$.
Key Insight
Observations come from your senses and inferences come from your brain, so keep them in separate columns of your notebook. Two scientists can draw different inferences from the same observation, and that disagreement is where new experiments come from. Inference is unavoidable, since nobody can watch an atom, and the discipline lies in stating what the inference depends on and what observation would prove it wrong.