ACT Science: Interpret Confidence Intervals and Margin of Error - Skip the Statistics Jargon

Confidence interval (CI) is a range. If a study shows "95% CI: 10-20," it means: based on the data, there's a 95% chance the true value falls between 10 and 20. Margin of error is half the confidence interval width. If the CI is 10-20, the margin of error is (20-10)/2=5. This tells you how far off the measured value might be. On the ACT, confidence intervals test whether you understand that measurements have uncertainty, not exact values.

Practical example: A poll shows "60% of voters support the candidate (95% CI: 55-65%)." This means: if you repeated the poll many times, about 95% of those polls would show support between 55-65%. The single measurement (60%) is just one estimate; the true value might be anywhere in the range. Wider intervals (55-70%) show more uncertainty. Narrower intervals (58-62%) show more certainty.

Trap 1: Thinking the confidence interval shows the range of individual data points. Wrong: "95% CI: 10-20" does NOT mean 95% of the data falls between 10-20. It means the true population average (not individual data points) likely falls in this range. Trap 2: Confusing 95% CI with a range where 95% of values lie. 95% CI means 95% chance the true parameter is in the range, not that 95% of data is in it. These sound similar but mean different things. Trap 3: Assuming a narrower CI is always better. Narrow CIs suggest higher precision, but they might result from overconfidence or a biased sample. Check the study design before trusting a narrow interval. Traps 1 and 2 stem from confusing the CI (about the population parameter) with data distribution (about individual data points).

Fix: remember that confidence intervals answer the question: "Where is the true value?" Not: "Where are the data points?" This distinction prevents all three traps.

Interval 1: "Average temperature: 25°C (95% CI: 20-30°C)." (Interpretation: There's a 95% chance the true average temperature is between 20-30°C. Margin of error: (30-20)/2=5°C.) Interval 2: "Drug efficacy: 75% (95% CI: 70-80%)." (Interpretation: There's a 95% chance the true efficacy is between 70-80%. Margin of error: 5%.) Interval 3: "Reaction time: 500 ms (95% CI: 490-510 ms)." (Interpretation: There's a 95% chance the true average reaction time is between 490-510 ms. Narrow interval suggests high precision.) Interval 4: "Population size: 10,000 (95% CI: 5,000-15,000)." (Interpretation: There's a 95% chance the true population is between 5,000-15,000. Wide interval suggests high uncertainty.) All four examples show that confidence intervals represent uncertainty around the measured value, not the distribution of individual data points.

For each interval, ask: "How wide is the range?" and "What does that tell me about the precision of the measurement?" Wider ranges suggest less certainty; narrower ranges suggest more certainty about where the true value lies.

Confidence intervals appear in 1-2 ACT Science questions per section, usually asking you to interpret what the CI means or compare uncertainty between two measurements. These questions are accessible if you understand the basic concept: CI is a range around the measured value, showing where the true value likely lies. You don't need statistics knowledge; you just need to understand that measurements have uncertainty, and the CI quantifies it.

Spend 20 minutes understanding confidence intervals using the explanations above. By test day, you'll interpret CIs correctly and answer these questions with confidence, boosting your Science score by 1-2 points per section.