Types of Error and Levels of Significance in Hypothesis Testing

Null Hypothesis ( $H_0$ ): A statement regarding a population parameter, such as $Mbewe ext{ is less than or equal to } 4$ .
Alternative Hypothesis ( $H_a$ ): The opposite statement that must be tested against the null hypothesis, such as Mbewe > 4.

There are two possible outcomes in hypothesis testing: - Reject the null hypothesis: This occurs when there is sufficient evidence to support the alternative hypothesis. - Fail to reject the null hypothesis: This happens when there is not enough evidence to support the alternative hypothesis.

Innocent Until Proven Guilty Concept: The null hypothesis is treated as the default position until evidence suggests otherwise. - Rejecting it signifies a shift in belief towards the alternative hypothesis. - Failing to reject means retaining the assumption of the null hypothesis.

In reality, one of these two statements must be true: - True Null Hypothesis: If $H_0$ is true and we fail to reject it - this is a good outcome. - False Null Hypothesis: If $H_0$ is false and we reject it - this, too, is a good outcome.
Problems arise when sampling errors occur due to non-representative samples, which leads to: - Type I Error: Rejecting the null hypothesis when it is in fact true. - Type II Error: Failing to reject the null hypothesis when it is indeed false.

Type I Error ( $ext{False Positive}$ ): - Defined as rejecting the null hypothesis when it is true. - Consistently involves, therefore, a claim that the alternative hypothesis is valid when it is not. - E.g., concluding that a blood culture contains Pathogen X when it truly does not.
Type II Error ( $ext{False Negative}$ ): - Defined as failing to reject the null hypothesis when it is false. - Involves not recognizing the validity of the alternative hypothesis in situations where it should be accepted. - E.g., concluding that a blood culture does not contain Pathogen X when it truly does.

Level of Significance ( $ext{Alpha}$ ): - Defined as the maximum allowable probability of wrongly rejecting a true null hypothesis (Type I error). - Notationally represented by the Greek letter $ext{Alpha}$ (α).
Probability of Type II Error ( $ext{Beta}$ ): - Represented as $\beta$ , it quantifies the chance of failing to reject a false null hypothesis.

Example 1: Blood Culture and Pathogen X: - Null hypothesis: The blood culture contains no trace of Pathogen X. - Type I Error (Rejecting $H_0$ ): - Concludes the cultures contain Pathogen X (incorrect conclusion).
- Type II Error (Failing to reject $H_0$ ): - Concludes the cultures do not contain Pathogen X (incorrect conclusion). - Discussion on which error is more serious is context-dependent. - In healthcare, concluding a patient is healthy when they are not (Type II) may be worse.
Example 2: Employment Rates: - Null hypothesis: The percentage of adults with jobs is at least 88%. - Type I Error (Rejecting $H_0$ ): Concludes that employment is less than 88% when it is, in fact, at least 88%. - Type II Error (Failing to reject $H_0$ ): Concludes the percentage of adults with jobs is at least 88% when it is actually less. - The severity of each type of error depends on context.

Both Type I and Type II errors are integral to hypothesis testing and need careful management.
The balance between alpha and beta is critical; choosing a low alpha increases the chance of beta error, and vice-versa.
The objective in hypothesis testing is to minimize the probability of both types of error while conducting statistically valid tests.