91Ó°ÊÓ

A one-sided hypothesis test is used when we are interested in determining if there is an effect in only one direction. In this scenario, you expect a change but want to test if that change is specifically higher or lower than expected. This is essentially a directional test, where you're trying to see if the data heads towards a particular direction.

For instance, consider when PepsiCo wants to find out if their new Diet Pepsi appeals to more teenagers than their standard formula. Here, PepsiCo is only interested in seeing if there's more appeal, making it a one-sided test.

• The null hypothesis ( H_0 ) would claim that "The new formula does not appeal to more teenagers." This implies either no change or a lesser degree of appeal.
• The alternative hypothesis ( H_a ) is more specific, stating that "The new formula does appeal to more teenagers."

Such scenarios are common in cases where proving an increase or improvement over a baseline is required.

A two-sided hypothesis test is utilized when we are checking for differences without specifying a direction. Here, you want to see if there is any change, not favoring an increase or decrease. This involves checking for any significant difference, regardless of direction.

For example, when a business student wants to see if there's a preference between Diet Coke and Diet Pepsi, they're simply interested in any preference. The test considers either brand being more preferred than the other.

• The null hypothesis ( H_0 ) would state: "There is no preference between Diet Coke and Diet Pepsi."
• The alternative hypothesis ( H_a ) considers both possibilities: "There is a preference between Diet Coke and Diet Pepsi."

Two-sided tests are useful in situations where deviation from the status quo in either direction is significant.

The null hypothesis ( H_0 ) is a fundamental concept in hypothesis testing. It represents the idea that there is no effect or no difference, and any observed effect is due to random chance. Essentially, it's your starting assumption before collecting data.

When testing hypotheses, the null hypothesis serves as the default position that you seek to test against. For example, in the budget override case, the null hypothesis could be "The proportion supporting the budget override is less than or equal to two-thirds." This means that the current support is not sufficiently high for the override to pass.
The strength of evidence is weighed to decide whether to reject the null hypothesis. If data strongly contradicts H_0 , it might be rejected, favoring the alternative hypothesis ( H_a ). Conversely, if there's insufficient evidence, H_0 is not rejected.

The alternative hypothesis ( H_a ) is the statement you aim to find support for in a hypothesis test. It represents the existence of an effect or difference in the population that the null hypothesis dismisses.

For example, when testing if the stock market goes up or down with equal probability, the alternative hypothesis might be "The probability of the stock market going up does not equal the probability of it going down." This suggests that there's a difference in probabilities, counter to what the null hypothesis states.

• In the case of the reformulated Diet Pepsi, H_a indicates a stronger appeal to teenagers.
• For the budget override poll, H_a implies support greater than two-thirds for approval.

The goal in testing is often to gather enough evidence to support H_a , showing that the observed effect is likely genuine and not due to random chance.