Welcome to the World of Estimation!
Ever wondered how scientists know how many fish are in a giant lake without catching every single one? Or how news channels predict election results before all the votes are counted? They use Estimation! In this chapter, you will learn how to take a small "snapshot" of data (a sample) and use it to make a very smart guess about the "whole picture" (the population). Don’t worry if it sounds like magic at first—it’s actually just clever use of numbers!
1. Estimating Population Characteristics
The main goal of estimation is to use what we know about a small group to describe a much larger group.
Using the Sample Mean
If you take a representative sample from a population, the sample mean (average) is often a great estimate for the population mean.
Example: If you measure the heights of 30 students in Year 11 and find their average height is 165cm, you can estimate that the average height of all Year 11 students in the country is also roughly 165cm.
Predicting Proportions
We can also use samples to predict proportions (how often something happens).
Quick Review: A proportion is just a fraction or percentage of a group.
If 10% of your sample has blue eyes, you can estimate that approximately 10% of the whole population has blue eyes.
The Median Rule of Thumb
A simple way to estimate characteristics is using the median. We can predict that approximately half of the population will be above the sample median, and half will be below it. This is a common exam "estimate" question!
Quick Review Box:
• Sample: The small group you actually measure.
• Population: The entire group you want to know about.
• Estimate: An educated guess based on data.
Key Takeaway: What happens in a fair sample is likely to be happening in the whole population.
2. The "Capture-Recapture" Method (Higher Tier Only)
If you need to estimate the total size of a population (like animals in the wild), we use a special technique called Petersen Capture-Recapture.
How it works: Step-by-Step
1. Capture: Catch a group of individuals from the population.
2. Tag: Mark them safely (so they are "tagged") and count them. Let's call this number \( n_1 \).
3. Release: Let them go back into the wild to mix thoroughly with the rest.
4. Recapture: Later, catch a second group. Let's call the total number in this second group \( n_2 \).
5. Count Tags: Count how many in the second group are already marked. Let's call this \( m \).
6. Calculate: Use the formula to find the total population (\( N \)).
The Formula
The formula to estimate the total population \( N \) is:
\( N = \frac{n_1 \times n_2}{m} \)
Memory Aid: "The CTRR Mnemonic"
Catch, Tag, Release, Recapture!
Important Assumptions
For this estimate to be accurate, we have to assume ideal conditions. In exams, you are often asked what might make the estimate wrong. Common assumptions are:
• The population stayed the same (no births, deaths, or migrations).
• The marks/tags did not fall off or disappear.
• The tagged animals mixed perfectly back into the whole group.
• Being tagged didn't make the animals easier or harder to catch the second time.
Common Mistake to Avoid: Don't forget that if the tag makes an animal "trap-shy" (scared to be caught again), your value for \( m \) will be too small, making your total population estimate \( N \) way too big!
Key Takeaway: Capture-Recapture is a ratio. We assume the proportion of tagged items in our second catch is the same as the proportion of tagged items in the whole population.
3. Reliability and Sample Size
Is an estimate always right? Not exactly! It’s just an estimate. However, we can make it more reliable.
The Power of "Big"
The most important rule in Statistics is: The larger the sample, the more reliable the result.
Analogy: If you want to know if a giant pot of soup is salty, tasting one tiny drop might be misleading. Tasting a whole spoonful gives you a much better "estimate" of the flavor!
Replication
Replication means doing the study again. If you repeat your sampling and get similar results every time, your estimate is much more dependable. If you only do it once, it could be a fluke!
Did you know?
Political pollsters usually talk to about 1,000 people to estimate the views of 67 million people in the UK. Because they use a large, random sample, they are often very close to the truth!
Quick Review Box:
• Small Sample: Likely to be biased or affected by "freak" results.
• Large Sample: More likely to represent the true population mean.
• Reliability: How much we can trust our results to be consistent.
Key Takeaway: If a question asks how to improve an investigation, the answer is almost always: "Use a larger sample size."
Summary Checklist
Before you move on, make sure you can:
• Identify the population and the sample in a story.
• Use a sample mean or median to guess a population characteristic.
• (Higher Tier) Use the Capture-Recapture formula \( N = \frac{n_1 \times n_2}{m} \).
• (Higher Tier) List the assumptions needed for Capture-Recapture to work.
• Explain why a larger sample makes an estimate better.
Don't worry if this seems tricky at first! Just remember that estimation is all about using the small bits of info we have to make the best possible guess about the big stuff we can't see. You've got this!