Garbage in, garbage out means that poor-quality input data can produce unreliable results, even when the analysis is sophisticated. “Unreliable” matters more than simply “inaccurate”: an apparently good result may still fail when repeated or used for a larger decision. section 2.1
A lucky result is not a reliable method
Imagine betting on a coin flip and winning once. That success does not prove the approach is dependable; the next flip may produce the opposite result. Likewise, an analysis based on poor data may happen to support a good business outcome without becoming trustworthy. section 2.1
Improve the reliability of decisions by:
- Examining the data closely before analysis. section 2.1
- Considering what each variable means in the specific business context. section 2.1
- Treating unavoidable data compromises as conscious limitations when evaluating results. section 2.1
- Checking for skewed inputs that could produce biased or unfair outputs. section 8.5
