Hypothesis testing for a binomial probability p

Introduction to Hypothesis Testing

Hi there! Welcome to one of the most exciting and practical parts of your A Level Mathematics course. Have you ever wondered if a "lucky" coin is actually biased, or if a manufacturer's claim about a product is actually true? Hypothesis Testing is the mathematical toolkit that allows us to answer those questions using data.

In this chapter, we focus on testing a binomial probability \(p\). We are essentially acting as "statistical detectives," looking at a sample of results to see if there is enough evidence to suggest that the "true" probability of something happening is different from what we originally thought. Don't worry if it seems a bit abstract at first—once you learn the steps, it becomes very logical!

Section 1: The Language of Hypothesis Testing

Before we start calculating, we need to understand the "lingo." Statistical testing has its own specific vocabulary that you need to use correctly in your exams.

1. The Hypotheses

Every test starts with two competing statements:
• The Null Hypothesis (\(H_0\)): This is the "status quo." We assume nothing has changed and the probability \(p\) is exactly what it's supposed to be. In your exam, this always looks like \(H_0: p = \text{number}\).
• The Alternative Hypothesis (\(H_1\)): This is the "interesting" claim we are investigating. We think the probability has changed. This will look like \(H_1: p > \text{number}\), \(H_1: p < \text{number}\), or \(H_1: p \neq \text{number}\).

2. The Test Statistic

The Test Statistic is the actual result we observe in our sample. For a binomial test, this is simply the number of successes we count in our experiment.

3. The Significance Level (\(\alpha\))

Think of this as the "burden of proof." It is a percentage (usually 5% or 10%) that defines the threshold for "unlikely."
Analogy: If you tell me you can predict the future, and you get one coin flip right, I'm not impressed. If you get 20 right in a row, that's so unlikely to happen by chance that I might start believing you. The significance level is the mathematical line we draw to decide when we are "impressed" enough to reject the null hypothesis.

Did you know? The significance level is actually the probability of incorrectly rejecting the null hypothesis. Even if the null hypothesis is true, there is a small chance (equal to the significance level) that we get an extreme result just by luck!

Key Takeaway:

The Null Hypothesis (\(H_0\)) is the default assumption (\(p = \text{value}\)), and the Alternative Hypothesis (\(H_1\)) is what we suspect is actually happening.

Section 2: 1-Tail vs. 2-Tail Tests

How do we know which way to point the arrow in \(H_1\)? It all depends on what the question is asking.

1-Tail Tests

We use a 1-tail test when we are only interested if the probability has moved in one specific direction.
• Example: A gardener thinks a new fertilizer makes seeds germinate better (so \(H_1: p > \text{old value}\)).
• Example: A doctor thinks a new drug makes a disease less likely (so \(H_1: p < \text{old value}\)).

2-Tail Tests

We use a 2-tail test when we just want to know if the probability has changed at all, and we don't care if it's higher or lower.
• Example: A machine is checked to see if the proportion of faulty items it produces is different from the usual 2% (so \(H_1: p \neq 0.02\)).

Quick Trick: In a 2-tail test, we split our significance level in half. If the total significance level is 5%, we look for 2.5% at the very bottom and 2.5% at the very top.

Section 3: The Critical Region and Critical Values

Once we have our hypotheses and significance level, we need to find the "Danger Zone"—mathematically known as the Critical Region (or Rejection Region).

• Critical Value: This is the first value that falls into the critical region. It's the "tipping point."
• Critical Region: This is the set of values for the test statistic that are so unlikely to occur if \(H_0\) is true that we decide to reject \(H_0\).
• Acceptance Region: Any value not in the critical region. If our result falls here, we stay with the null hypothesis.

Analogy: Imagine a "keep off the grass" sign. The grass is the Critical Region. If you step on it (your test statistic falls in that region), you've broken the rules of the null hypothesis and we must reject it!

Quick Review Box:

• If Test Statistic is in the Critical Region \(\rightarrow\) Reject \(H_0\).
• If Test Statistic is not in the Critical Region \(\rightarrow\) Do not reject \(H_0\).

Section 4: The p-value Method

Another way to conduct the test is using a p-value. Most modern software and calculators use this method.

The p-value is the probability of getting a result at least as extreme as the one we observed, assuming \(H_0\) is true.

• If p-value \(\leq\) Significance Level: The result is "significant." Reject \(H_0\).
• If p-value \(>\) Significance Level: The result is not "significant." Do not reject \(H_0\).

Mnemonic: "If the p is low, the \(H_0\) must go!" (If p-value is less than or equal to the significance level, reject the null hypothesis).

Section 5: How to Conduct a Binomial Hypothesis Test (Step-by-Step)

Follow these steps every time to ensure you get full marks in your MEI H640 exam:

1. State your Hypotheses: Write down \(H_0: p = \dots\) and \(H_1: p \dots\).
2. Define your Distribution: State the model we are using, e.g., \(X \sim B(n, p)\) where \(n\) is the sample size and \(p\) is from \(H_0\).
3. State the Significance Level: Usually given in the question (e.g., 5%).
4. Calculate the Probability: Find the probability of getting your observed value \(x\) or more extreme.
Note: If \(H_1: p > k\), calculate \(P(X \geq x)\). If \(H_1: p < k\), calculate \(P(X \leq x)\).
5. Compare: Compare your calculated probability (p-value) to the significance level.
6. Conclusion (Two Parts):
Part A (Mathematical): State whether you reject or do not reject \(H_0\).
Part B (Contextual): Write a sentence in plain English explaining what this means for the specific situation (e.g., "There is sufficient evidence to suggest the new seeds are better").

Key Takeaway:

Never just say "Reject \(H_0\)". Always finish by explaining what that means in the real-world context of the question.

Section 6: Common Mistakes to Avoid

Don't worry if this seems tricky at first; many students make these same mistakes! Look out for these:

• Wrong arrow in \(H_1\): Read the question carefully. "Has it increased?" means \(>\). "Has it changed?" means \(\neq\).
• Using \(<\) instead of \(\leq\): In binomial distributions (which are discrete), \(P(X \leq 5)\) is very different from \(P(X < 5)\). Always include the observed value itself!
• Forgetting to split the % for 2-tail tests: If it's a 10% 2-tail test, you are looking for 5% at each end.
• Assertions: Avoid saying "This proves the probability has changed." We say "There is sufficient evidence to suggest it has changed." Statistics is about evidence, not absolute proof!

Final Chapter Summary

• We use Hypothesis Testing to see if a sample provides enough evidence to reject a default assumption (\(H_0\)).
• For binomial tests, we test the probability of success, \(p\).
• The Significance Level is the risk we are willing to take of being wrong.
• We compare the p-value (the probability of our result or worse) to the significance level to decide our verdict.
• Always conclude in the context of the problem!