Why a Single Number Isn't Enough
Imagine a news report says: "The average American spends $3,200 per year on eating out." That sounds precise, but it came from a survey of just 500 people. How close is that number to the truth for all 330 million Americans?
A single number (called a "point estimate") gives you a best guess, but it tells you nothing about how reliable that guess is. That's where confidence intervals come in. They give you a range that's likely to contain the true answer.
What Is a Confidence Interval?
A confidence interval is a range of values that we believe contains the true population value, based on our sample data. Instead of saying "the average is $3,200," you'd say "we're 95% confident the true average is between $2,950 and $3,450."
That range - $2,950 to $3,450 - is the confidence interval. The "$250 above and below" is the margin of error.
The Election Poll Example
Before an election, a polling company surveys 1,200 likely voters. They find that 52% support Candidate A.
But they don't just report "52%." They report: "52% ± 3%, with 95% confidence."
This means: based on this sample, we're 95% confident that the true level of support for Candidate A, among ALL voters, is somewhere between 49% and 55%.
Notice that 49% is below 50%. So even though the poll shows Candidate A "ahead," the confidence interval tells us the race could still go either way. That's incredibly useful information that the single number "52%" doesn't reveal on its own.
What "95% Confident" Really Means
This is one of the most commonly misunderstood phrases in statistics. Here's what it does and doesn't mean:
It does NOT mean: "There's a 95% chance the true value is in this specific interval." Once the interval is calculated, the true value either is in it or it isn't - there's no probability about it.
It DOES mean: "If we repeated this entire study many times - taking a new sample each time and computing a new interval - about 95% of those intervals would contain the true value."
Think of it like a fishing net. If you cast the same type of net 100 times, you'd expect to catch the fish about 95 times. Any single cast might miss, but the method is reliable over the long run.
Margin of Error
The margin of error is the "plus or minus" part of a confidence interval. It tells you how far off your estimate might be. A smaller margin of error means a more precise estimate.
Three things affect the margin of error:
- Sample size: Bigger samples produce smaller margins of error. Surveying 2,000 people gives a tighter range than surveying 200.
- Variability in the data: If everyone in the population is very similar, you need fewer data points to estimate the average. If people vary widely, you need more.
- Confidence level: A 99% confidence interval is wider than a 95% one. Demanding more certainty means accepting a wider range.
A school wants to know the average time students spend on homework each night.
They survey 50 students and find an average of 1.8 hours, with a 95% confidence interval of 1.5 to 2.1 hours.
If they'd surveyed 200 students instead, the interval might narrow to 1.65 to 1.95 hours - a more precise estimate from more data.
If they only surveyed 15 students, the interval might widen to 1.2 to 2.4 hours - much less useful because the range is so broad.
How Sample Size Affects Width
This is one of the most practical insights in statistics. The relationship between sample size and precision isn't linear - it follows a "square root" rule.
To cut your margin of error in half, you need to quadruple your sample size. Going from 100 to 400 people cuts the margin of error in half. Going from 400 to 1,600 cuts it in half again.
This explains why most national polls survey around 1,000-1,500 people. Beyond that, the improvement in precision isn't worth the extra cost. Surveying 10,000 people instead of 1,000 only makes the interval about three times narrower - often not worth the tenfold increase in effort.
Different Confidence Levels
You can choose different confidence levels depending on your needs:
- 90% confidence: Narrower interval, but more chance of missing the true value.
- 95% confidence: The most common choice. A good balance between precision and reliability.
- 99% confidence: Wider interval, but you're almost certain the true value is included.
Higher confidence means a wider net. You're more likely to catch the truth, but your estimate is less precise. It's a trade-off you choose based on how much risk you're comfortable with.
Reading Confidence Intervals in the News
When you see confidence intervals reported, here are some practical tips:
- If two confidence intervals don't overlap, the groups are likely genuinely different.
- If a confidence interval for a difference includes zero, the difference might not be real.
- A very wide interval means the estimate is unreliable - probably from a small sample or highly variable data.
- Always check the confidence level. A 90% interval looks tighter than a 99% interval from the same data, but it's less reliable.
A confidence interval gives you a range of plausible values for a population measure, rather than a single guess. The margin of error tells you how wide that range is. Larger samples produce narrower, more precise intervals. When you see "95% confidence," it means the method works 95% of the time over many repetitions - not that there's a 95% chance this particular interval is correct. Confidence intervals are essential for understanding how much trust to place in any estimate.