Use the chain claim → assumption → method quality → source of error → conflicting evidence → justified conclusion → follow-up.
An assumption is treated as true without being checked. Examples include equal starting temperatures, moisture, composition or instrument calibration. Ask whether a false assumption could change the dependent variable.
Name the mechanism: parallax, timing inconsistency, uncontrolled stirring speed or calibration bias. A consistent offset is systematic error; repeating the biased method does not remove it.
Conflicting evidence should trigger comparison of samples, controls, equipment, range and conditions—not selective acceptance. A conclusion must stay within the tested species, temperature, cycle or operating range.