Short answers, pulled from the story.
A scientific control is an element of an experiment designed to minimize the influence of variables other than the independent variable being studied, reducing the risk of confounding. It provides a baseline for comparison between experimental and control measurements, increasing the reliability and validity of results.
A negative control is a non-treatment condition expected to produce no effect, used to detect confounding or verify that a result is not a false positive. A positive control uses a known-effective treatment or substance to confirm that the experiment is capable of detecting real effects at all.
Confounding occurs when an extraneous variable is related to both the independent variable and the outcome, distorting the apparent relationship between them. Uncontrolled confounding can lead to incorrect policy recommendations, ineffective interventions, and flawed scientific understanding.
A Negative Control Exposure is a variable that should have no causal effect on the outcome but shares the same confounding mechanism as the exposure being studied. If a statistical association is found between the NCE and the outcome, it signals that unmeasured confounding is likely distorting the primary analysis.
Unblinding occurs when a trial participant deduces or obtains information that was meant to be concealed, such as whether they received an active treatment or a placebo. It reintroduces the bias that blinding was designed to eliminate, and meta-research has found high rates of unblinding especially in antidepressant trials.
Randomization assigns subjects or plots to treatment groups by chance, ensuring that differences between groups are distributed equally rather than systematically favoring one condition. In crop yield experiments, for example, randomly assigning treatments to plots of land prevents variation in soil fertility from skewing results.