Welcome to the World of Statistical Enquiry!

Ever wondered how scientists decide if a new medicine works, or how pollsters predict election results? They don't just guess; they follow a rigorous process called the Statistical Enquiry Cycle (SEC). In this chapter, we focus on the most important part of any investigation: Planning and Data Collection. Think of this as the "foundation" of a house—if the foundation is shaky, the whole building will fall down, no matter how pretty the curtains are!

The Statistical Enquiry Cycle (SEC) Overview

The SEC is a five-stage loop that statisticians use to solve real-world problems. While you will be tested on the whole cycle across your three papers, today we are focusing on the first two stages:

  1. Initial Planning: Asking the right questions and deciding the "rules" of the game.
  2. Data Collection: Gathering the raw ingredients for your analysis.

Note: Later chapters will cover Stage 3 (Processing), Stage 4 (Interpretation), and Stage 5 (Evaluation).

Stage 1: Initial Planning

Planning is about more than just picking a topic. It involves being specific and anticipating problems before they happen. Here is what you need to do:

1. Define the Question or Hypothesis

A good investigation starts with a clear hypothesis (a statement you can test) or a specific question. Instead of saying "I want to study height," you should say "I want to test if students in Year 13 are significantly taller than students in Year 7."

2. Identify Related Factors

In the real world, things are rarely simple. If you are measuring test scores, you must identify other factors that might affect the result, such as hours of sleep, time spent revising, or even the room temperature. In statistics, we try to control these factors or record them so they don't bias our results.

3. Exploratory Data Analysis (EDA)

Before doing a massive study, statisticians often do EDA. This is like a "mini-investigation" or a pilot study. It helps you see the general shape of the data, spot obvious outliers, and decide which statistical tests might be appropriate later on. It’s the "look before you leap" stage.

4. Justifying the Plan

You must be able to explain why you chose a specific method. If you choose to sample 50 people, why not 500? If you choose to use a survey, why not an experiment? A key part of your justification must be how you plan to avoid bias.

Quick Review: Planning involves defining a testable hypothesis, identifying confounding factors, and performing EDA to ensure your plan is robust and unbiased.

Stage 2: Data Collection – The "How To"

Once you have a plan, you need data. There are two main ways to get it:

  • Primary Data: Data you collect yourself (e.g., your own experiment or survey).
  • Secondary Data: Data collected by someone else (e.g., from the internet, newspapers, or government publications). You must always acknowledge your sources when using secondary data.

Sampling Methods

You usually cannot ask every single person in a population (the whole group). Instead, you take a sample. Here are the methods you need to know for your exam:

  • Simple Random Sampling: Every member of the population has an equal chance of being chosen, and every possible sample of size \( n \) has an equal chance of being selected. This is usually done using random number tables or a calculator. It is "without replacement."
  • Unrestricted Random Sampling: This is similar to simple random, but it is "with replacement," meaning the same individual could be picked twice.
  • Systematic Sampling: Choosing every \( k \)-th element from a list (e.g., every 10th person on a register).
  • Cluster Sampling: Dividing the population into groups (clusters), then choosing a few clusters at random and surveying everyone inside those clusters.
  • Stratified Sampling: Dividing the population into layers (strata), like age groups or genders, and then taking a random sample from each layer.
    • Proportional: The number of people from each layer matches the population's proportions.
    • Disproportional: You intentionally over-sample a small group to ensure you have enough data to analyze them properly.
  • Judgmental Sampling: The researcher uses their own "expert" judgment to choose who should be in the sample. This is prone to bias!
  • Snowball Sampling: Starting with one person and asking them to suggest others. This is useful for hard-to-reach groups (like people with a rare hobby).

Avoiding Bias in Data Collection

Bias is the enemy of statistics. It makes your results unreliable. Common sources of bias include:

  • Leading Questions: "Don't you agree that stats is the best subject?" (This pushes the person to say yes).
  • Non-response Bias: When a certain type of person refuses to answer your survey.
  • Sampling Bias: Choosing a sample that doesn't represent the whole population (e.g., surveying people at a gym to ask about the nation's average fitness).

Key Takeaway: Your choice of sampling method depends on your budget, time, and the type of population. Always justify your choice by weighing the advantages against the limitations.

Experimental Design

If your enquiry is an experiment (where you change something to see what happens), you need to know these concepts:

  • Experimental Error: The difference between your results and the "true" value, often caused by things you couldn't control.
  • Replication: Doing the experiment many times. If you only do it once, it could be a fluke!
  • Randomisation: Assigning people to groups by chance to ensure the groups are as similar as possible.
  • Control vs. Experimental Groups: The experimental group gets the treatment (e.g., a new energy drink), while the control group gets nothing or a "placebo" (e.g., water).
  • Blinding:
    • Blind: The participant doesn't know if they are in the control or experimental group.
    • Double-blind: Neither the participant nor the researcher knows who is in which group. This prevents the researcher from accidentally influencing the results!

Blocking and Paired Comparisons

Sometimes we know a factor will affect our results (like "age" in a memory test). We use Blocking to group similar people together before we start, ensuring that each block is represented in both the control and experimental groups. Paired comparisons are even better—this is where you compare two very similar people (or the same person at two different times) to cancel out individual differences.

Common Mistakes to Avoid

Don't forget the context! Edexcel examiners love context. Never just say "the sample was biased." Say "the sample was biased because only people who like football were asked, so it doesn't represent the whole school's opinion on sports."

Sample Size (\( n \)): Remember that a larger sample size generally reduces the standard error and makes your results more reliable, but it costs more time and money. It's always a balance!

Quick Review Box:
1. Primary = You collected it; Secondary = Someone else did.
2. Random = Equal chance for all.
3. Stratified = Groups (strata) are represented fairly.
4. Double-blind = No one knows who got the "real" treatment.
5. Hypothesis = Must be specific and testable.

Ready for the next step? Once your data is collected, you'll head into "Data Processing and Presentation," where we turn those raw numbers into meaningful graphs and calculations!