Inference and hypothesis testing - Mathematics: Applications & Interpretation IB Study Notes

Imagine you're trying to figure out if a new type of fertilizer makes plants grow taller. You can't test it on every single plant in the world, right? That would be impossible!

So, what do you do? You pick a small group of plants, say 50 of them, and give them the new fertilizer. This small group is called your sample. Then you compare their growth to another small group of plants that got the old fertilizer.

Inference is like looking at how your 50 plants grew and then making a smart guess (or an inference) about whether the fertilizer would make all plants grow taller. You're using a small piece of information to understand a bigger picture.

Hypothesis Testing is the official way we check if our guess is probably true or just a lucky accident. It's like setting up a scientific experiment with rules to see if your fertilizer really made a difference, or if the plants just grew taller by chance. We start with an idea (a hypothesis) and then use our sample data to see if there's enough evidence to support that idea.

Let's say a famous chocolate company thinks their new recipe for chocolate chip cookies makes them taste even better. They can't ask everyone in the world to try them, so they decide to do a taste test.

1. The Idea (Hypothesis): The company's main idea (their alternative hypothesis) is that people will prefer the new recipe. But for testing, we usually start with the opposite: the null hypothesis, which says there's no difference – people will like both recipes equally.
2. The Sample: They gather 100 people from a local shopping mall. This group of 100 is their sample (a smaller group representing all potential cookie eaters).
3. The Test: Each person tries both the old and new cookie recipes (without knowing which is which, to be fair!) and says which one they prefer.
4. The Data: Let's say 70 out of 100 people preferred the new recipe.
5. The Inference: Now, the company needs to decide: Is 70 out of 100 a strong enough sign that everyone would prefer the new recipe? Or could it just be a random happenstance that 70 people in this specific sample liked it more, and if they picked another 100 people, it might be 50/50?
6. The Conclusion: Using hypothesis testing, they'll calculate if getting 70 preferences for the new recipe is so unusual (if there truly was no difference) that they can confidently say, "Yes! The new recipe is better!" If it's not unusual enough, they might say, "We don't have enough evidence to say the new recipe is better, it could just be random." This helps them decide whether to launch the new cookie recipe or stick with the old one.