Develop a hypothesis testing

Get Started. It's Free
or sign up with your email address
Develop a hypothesis testing by Mind Map: Develop a hypothesis testing

1. Step2. Select a level of significance

1.1. It is the probability of rejecting the null hypothesis when its true. It is designated as α (1-Confidence level)

1.2. 0.05 is often used for consumer research project 0.01 for quality control 0.1 for political polling

1.2.1. Type 1 error

1.2.1.1. Rejects Ho while Ho is true (We caught the wrong suspect) This is more dangerous error. The is the α, aka the significance level of a test

1.2.2. Type 2 error

1.2.2.1. Fails to reject Ho while Ho is false. (we let the suspect go) Denoted as β

2. Step5. Make a conclusion and decision

3. Step1. State the null and alternative hypothesis

3.1. The null hypothesis (Ho) Alternative hypothesis (H1), its also called reseach hypothesis They are mutually exclusive and collectively exhaustive

3.1.1. Types of hypothesis Ho always carries the equal sign, we develop the H1 first and then Ho

3.1.1.1. H1: > (right side) H1: < (Left side) H1: = (Equal)

3.2. *In this technique, we only state the following and not “accept” When we reject: statement is not good Fail to reject: Statement is good H1 is what we want. The null hypotheis not being rejected does not prove that Ho is true; it means that we have failed to disprove Ho Hypothesis should be concise and clear

3.2.1. Indicating signs

4. Step3. Identify a test statistics

4.1. If we have population SD

4.1.1. Z statistics

4.2. Sample SD (population SD is unknown)

4.2.1. T statistics

4.3. For population proportion

4.3.1. z proportion statistics

4.4. *There are many other test statistics

5. Step4. Reject or fail to reject

5.1. we will find: Z (for alpha) Z (critical value from test statistics)

5.1.1. Then compare them under the normal distribution curve. If Z (critical value) > Z (Alpha) => reject

5.2. P value

5.2.1. The p-value is a probability that measures the evidence against the null hypothesis. The p-value tells you how likely it is to get the observed result in your sample if the null hypothesis is true.

5.2.2. **How to calculate?** 1. Find the probability from the test statistics (0.45 for example) 2. P value = 0.5 - 0.45 (for One tail) p value = 2*(0.5-0.45) (for two tail) 3. Compare with the significance level (10% for example). If less than α then reject. The smaller the P, the stronger evidence to reject Ho