The Inevitable Nature of Uncertainty
Uncertainty is an inherent human feeling, often focusing on the future and what will happen tomorrow. However, uncertainty is not confined to calendars; it also resides in understanding the reasons behind events. Questions like “why do I measure 1.65 meters?” or “what leads to cancer development?” illustrate this concept. It’s not just about wondering if a coin will land on heads; it’s about understanding the physical and contextual mechanisms that drive such outcomes.
Quantifying Uncertainty: The Role of Probability and Statistics
Since there’s no crystal ball to predict the future, we must rely on probability and statistics to quantify uncertainty. Probability provides a language to convert this lack of certainty into actionable insights, while statistics anchor this language to observable and measurable data.
Probability in Personal Concerns
Consider lung cancer as an example. Personal concern about this disease is often framed in binary terms (“will I get it or not?”), but the answer lies in probability. A 3%, 20%, or 80% probability of developing lung cancer are not guarantees, but they will influence decisions regarding screening frequency, habits, and priorities.
Determining this probability involves estimations based on data from individuals with similar age, environmental exposure, habits, and backgrounds. A serious study won’t assert “your risk is 3%” but rather “given your characteristics and available data, the risk is most likely between 1% and 5%.” This “most likely” can also be quantified, providing a range and coherence to our knowledge.
Factors Affecting Risk Estimation
The quality of these estimations depends on three pillars: data, design, and model. Data must accurately represent the population of interest. The data collection process should be well-designed with appropriate controls to avoid biases. Lastly, the mathematical simplification and statistical techniques used for estimation must be applicable to yield better results.
Understanding Causation vs Correlation
After determining the magnitude, it’s crucial to understand how much a factor changes the risk. Risks can be absolute or relative. Saying “smoking doubles the risk” is a relative statement: if a non-smoker has a 10% risk, doubling it results in 20%. While this sounds impactful when the risk goes from 0.5% to 1%, the practical difference is significant.
Interpreting Probability: A Guide
When informed about your lung cancer risk, probabilities are usually expressed as percentages (between 0% and 100%). They can also be presented as fractions (3% is 0.03; 20% is 0.2) or “x out of N” where, for example, one in a thousand equals 0.1% (0.001) or three out of ten, which is 30%.
Confusion arises when different formats are mixed in the same text. We tend to focus more on the total (N) than the part (x), potentially misinterpreting “one in a thousand” as a higher risk than “one in a hundred,” when in fact, it’s 0.1% versus 1%.
Once the number is understood, interpretation follows. There are two perspectives: frequentist and Bayesian. The frequentist view sees probability as the frequency of an event when repeating a process many times, working well in controlled and repeatable contexts. However, when discussing weather or illness—where repetition isn’t possible—the Bayesian perspective, treating probability as a degree of belief updateable with new evidence via the Bayes’ theorem, is more natural.
These perspectives also aid in understanding the uncertainty of estimations. Plausible value ranges for estimations are called confidence intervals under the frequentist paradigm or credibility intervals under Bayesianism. Both help measure estimation reliability. A 1% to 5% risk range is more precise than a 1% to 10% range, though both might yield similar levels of comfort regarding the disease. The more precise interval suggests better data or improved data usage.
Embracing Uncertainty
Uncertainty is inherent in daily processes and cannot be eliminated, but it can be measured. By doing so, we make uncertainty visible and honest. Probability and statistics don’t promise certainty but offer better questions and more informed decisions. It’s about navigating life’s complexities with improved judgment, not controlling every hidden corner but walking through them more wisely.
