91Ó°ÊÓ

A Type I error is like a false alarm. It occurs in hypothesis testing when we reject a true null hypothesis. Imagine it as sounding the fire alarm when there's no fire. This type of error is also known as a "false positive."
In statistics, decisions are rarely perfect. Sometimes, tests can mistakenly identify an effect when there is none, leading to a Type I error. To control this risk, we set a significance level, denoted by \(\alpha\).
For example, if \(\alpha = 0.05\), it means we accept a 5% risk of wrongly rejecting the null hypothesis. It's crucial to choose this level wisely because it balances the risk of error with the precision of the test. In our given exercise, this level is indeed set at \(0.05\), meaning a 5% probability of committing a Type I error.
Key points about Type I Error:

• Occurs when a true null hypothesis is rejected.
• Significance level \(\alpha\) represents its probability.
• Common values are 0.05 or 0.01.
• It represents a "false positive" result in testing.

Type II error happens when we fail to reject a false null hypothesis. This is similar to ignoring a fire alarm when there's actually a fire. Such an error is also labeled as a "false negative."
The probability of making a Type II error is denoted by \(\beta\). Unlike \(\alpha\), \(\beta\) is not typically prefixed; it's calculated based on the power of the test.
In our exercise, the power of the test is 0.78. This means there's a 78% chance that the test correctly identifies a false null hypothesis. From the formula \(1 - \beta = \text{power}\), we calculate \(\beta\) to be 0.22.
That translates to a 22% chance of not detecting an effect when there is indeed one. Type II errors are generally more common when a test lacks sensitivity.
Important facts about Type II Error:

• Occurs when a false null hypothesis is not rejected.
• Represented by \(\beta\), its calculation depends on the test's power.
• It is a "false negative" result.

Statistical power measures the likelihood that a test will detect an effect when there is one. It's the ability of a test to correctly reject a false null hypothesis, akin to hearing the fire alarm when there's indeed a fire.
In our scenario, the power is 0.78, which indicates that there's a 78% probability of identifying a false null hypothesis correctly. Higher power implies a lower risk of making a Type II error. It’s calculated as \(1 - \beta\).
Power is influenced by several factors:

• Sample size – larger samples usually lead to higher power.
• Effect size – larger effects are easier to detect, increasing power.
• Significance level \(\alpha\) – altering \(\alpha\) can impact power.
• Variability in data – less variability generally increases power.

Evaluating a test’s power helps researchers optimize their experiments for meaningful results, balancing the potential for Type I and Type II errors.