Welcome to the World of Sampling!

In this chapter, we are going to explore why we don't always look at every single piece of data available to us. Whether you are checking if a batch of lightbulbs works or predicting the result of a national election, you are using sampling.

Don't worry if Statistics feels a bit abstract at first. Think of sampling like being a chef: you don't need to eat the entire pot of soup to know if it needs more salt; you just need one well-mixed spoonful! By the end of these notes, you’ll understand how to pick that "spoonful" correctly so your conclusions are accurate.

1. Why Sample? (The Census vs. Sample Debate)

In a perfect world, we would look at the population (the whole group). This is called a census. However, in the real world, a census is often impossible. Here is why we use a sample (a smaller selection) instead:

  • Population Size: Sometimes the population is just too big! Imagine trying to weigh every grain of sand on a beach.
  • Cost and Time: It is much cheaper and faster to ask 1,000 people their opinion than to ask 60 million.
  • Destructive Testing: This is a big one for Further Maths! If you want to test how much pressure a glass bottle can take before it breaks, you have to break it. If you tested the entire population, you would have zero bottles left to sell!
  • Act of Sampling: Sometimes, the very act of measuring something changes it. We want to ensure our data remains relevant and the population remains unchanged by our study.

Did you know? A sample can be mathematically viewed as \( n \) observations taken from a random variable. This allows us to use all those cool probability formulas we learn later in the Statistics Major!

Key Takeaway:

We sample because it is practical, cost-effective, and preserves the population.


2. Features of a "Good" Sample

Not all samples are created equal. If you want to know how much the average person spends on shoes, but you only interview people outside a luxury boutique, your data will be biased. To make a sample useful, it must have these features:

  • Unbiased: It shouldn't systematically favor one outcome over another.
  • Representative: It should "look like" the population. If the population is 50% women, your sample should ideally be around 50% women too.
  • Relevant: The data collected must actually answer the question you are asking.

Quick Review Box:
Bias is the "villain" of statistics. It’s a systematic error that makes your results unreliable. Always ask: "Is there something about how I picked this group that might skew the results?"


3. The Importance of Sample Size

In experimental design, the size of your sample (\( n \)) is vital. You might hear people talk about Effect Size—this is just a way of measuring how "strong" a result is.

Analogy: Imagine tossing a coin. If you toss it twice and get two Heads, you wouldn't be surprised. But if you toss it 1,000 times and get 1,000 Heads, you'd be certain the coin is rigged!

A larger sample size helps us:

  1. Reduce the impact of "flukes" or random chance.
  2. Increase our confidence in the effect size we observe.
  3. Provide a clearer "picture" of the population.

Common Mistake to Avoid: Don't assume a huge sample always fixes everything. A massive sample that is biased is still a bad sample! Quality and quantity both matter.


4. The Advantage of Random Sampling

The syllabus highlights that Random Samples are the gold standard for Inference (making a "best guess" about the population). Why? Because the probability basis of the selection is known.

When every member of a population has an equal chance of being chosen:

  • We can use mathematical models to calculate exactly how likely our results are.
  • It removes human choice, which is often the source of hidden bias.
  • It allows us to "scale up" our findings to the whole population with a calculated level of certainty.

Memory Aid: "RUB" Your Data!
A good sample should be:
Representative
Unbiased
Big enough (Appropriate Size)


5. Chapter Summary Checklist

Before you move on to Discrete Random Variables, make sure you can explain:

  • Why we can't always take a census (e.g., destructive testing).
  • What features make a sample "good" (unbiased, representative).
  • How sample size affects our interpretation of the results.
  • Why random sampling is better for making mathematical predictions.

Don't worry if this seems a bit "wordy" for a math subject! These concepts are the foundation for the heavy-duty calculations you'll do later in the course. Get the concepts right now, and the numbers will make much more sense later!