What is Hypothesis Testing Procedure?

S3-SA3-0341

Grade Level:

Class 9

AI/ML, Data Science, Physics, Economics, Cryptography, Computer Science, Engineering

Definition

What is it?

Hypothesis testing is a statistical method used to make decisions about a population based on sample data. It's like checking if an idea or 'hypothesis' you have about something is likely true or false, using evidence from a smaller group.

Simple Example

Quick Example

Imagine you think the average price of a cup of chai in your city is Rs. 15. You ask 10 friends what they paid, and their average is Rs. 17. Hypothesis testing helps you decide if Rs. 17 is 'different enough' from Rs. 15 to say your initial idea (Rs. 15) is probably wrong, or if the difference is just due to random chance.

Worked Example

Step-by-Step

Let's say a mobile company claims their new phone battery lasts 10 hours on average. A tech reviewer thinks it's less. They test 50 phones and find an average battery life of 9.5 hours. They also know from past data that the battery life varies a bit, with a standard deviation of 1 hour.

Here's how they might think about it:
1. **Formulate Hypotheses:**
* Null Hypothesis (H0): The average battery life is 10 hours (μ = 10).
* Alternative Hypothesis (H1): The average battery life is less than 10 hours (μ < 10).
---2. **Choose a Significance Level:** The reviewer decides they want to be 95% confident, so their significance level (alpha) is 0.05. This means they're okay with a 5% chance of being wrong.
---3. **Calculate the Test Statistic:** They use a formula (like a z-score for large samples) to see how far 9.5 hours is from 10 hours, considering the sample size and variation.
* Z = (Sample Mean - Population Mean) / (Standard Deviation / sqrt(Sample Size))
* Z = (9.5 - 10) / (1 / sqrt(50))
* Z = -0.5 / (1 / 7.07)
* Z = -0.5 / 0.1414 = -3.53
---4. **Find the Critical Value:** For a 0.05 significance level in a 'less than' test, the critical z-value is approximately -1.645.
---5. **Compare Test Statistic to Critical Value:** The calculated Z-value (-3.53) is less than the critical Z-value (-1.645).
---6. **Make a Decision:** Since -3.53 is in the 'rejection region' (it's further into the negative side than -1.645), the reviewer rejects the null hypothesis.
---7. **Conclusion:** The reviewer concludes there is enough evidence to say that the average battery life of the new phone is indeed less than 10 hours. The company's claim is likely false.

Why It Matters

Hypothesis testing is crucial in fields like AI/ML to decide if a new algorithm performs better, in Physics to confirm experimental results, and in Economics to test theories. Data Scientists, Engineers, and Researchers use it daily to make informed decisions based on data, helping build better products and understand the world.

Common Mistakes

MISTAKE: Assuming that if you don't reject the null hypothesis, it means the null hypothesis is definitely true. | CORRECTION: Not rejecting the null hypothesis only means there isn't *enough evidence* to say it's false. It doesn't prove it's true, just like 'not guilty' isn't the same as 'innocent'.

MISTAKE: Confusing the significance level (alpha) with the probability of the null hypothesis being true. | CORRECTION: The significance level (alpha) is the maximum probability of making a Type I error (incorrectly rejecting a true null hypothesis). It's a threshold for decision-making, not the probability of the hypothesis itself.

MISTAKE: Only looking at the sample mean and deciding if it's 'different enough' without considering variation or sample size. | CORRECTION: Always use a formal test statistic that accounts for the variability in the data and the number of observations. A small difference might be significant with a large sample, but not with a small one.

Practice Questions

Try It Yourself

QUESTION: A school claims its students watch an average of 2 hours of TV per day. A survey of 30 students shows they watch 2.5 hours on average. If the standard deviation is 1 hour, what is the null hypothesis (H0) for testing this claim? | ANSWER: H0: The average TV watching time is 2 hours (μ = 2 hours).

QUESTION: A biscuit factory claims each packet has 100g of biscuits. A quality checker weighs 25 packets and finds the average weight is 98g. If they set a significance level (alpha) of 0.01, what does this alpha value represent in simple terms? | ANSWER: It means the checker is willing to accept a 1% chance of incorrectly concluding that the biscuit packets are underweight when they are actually 100g on average.

QUESTION: A new medicine is being tested to reduce fever. The current medicine reduces fever by an average of 2 degrees Celsius. Researchers test 100 patients with the new medicine and find an average fever reduction of 2.3 degrees, with a standard deviation of 0.5 degrees. If the critical Z-value for a 0.05 significance level (one-tailed test for improvement) is 1.645, what is the Z-score for the new medicine, and should the researchers conclude the new medicine is better? | ANSWER: Z = (2.3 - 2) / (0.5 / sqrt(100)) = 0.3 / (0.5 / 10) = 0.3 / 0.05 = 6. Since 6 > 1.645, the researchers should conclude the new medicine is better.

MCQ

Quick Quiz

Which of the following is the first step in a hypothesis testing procedure?

Collecting data

Formulating the null and alternative hypotheses

Calculating the test statistic

Making a decision about the hypothesis

The Correct Answer Is:

The very first step in hypothesis testing is always to clearly state what you are testing, which involves setting up the null and alternative hypotheses. All other steps follow this initial formulation.

Real World Connection

In the Real World

When you see an ad claiming a new shampoo makes hair 20% stronger, that claim likely went through hypothesis testing. Companies like Flipkart or Amazon use it to test if a new website layout increases sales. Doctors use it to determine if a new drug is more effective than an old one. Even ISRO scientists might use it to verify if a new rocket fuel mixture performs as expected.

Key Vocabulary

Key Terms

Null Hypothesis: The statement that there is no effect or no difference, assumed true until proven otherwise. | Alternative Hypothesis: The statement that there is an effect or a difference, which you are trying to prove. | Significance Level (Alpha): The probability of rejecting the null hypothesis when it is actually true (Type I error). | Test Statistic: A value calculated from sample data used to decide whether to reject the null hypothesis. | P-value: The probability of observing a test statistic as extreme as, or more extreme than, the one calculated, assuming the null hypothesis is true.

What's Next

What to Learn Next

Now that you understand the procedure, you should learn about different types of hypothesis tests, like Z-tests and T-tests, and when to use each. This will help you apply the procedure to various real-world problems and data!