91Ó°ÊÓ

In hypothesis testing, the null hypothesis, denoted as \(H_0\), serves as the default or baseline assumption. Essentially, it asserts that there is no effect or no difference; it stands as the statement we aim to test against. For instance, in our exercise with the mice, the null hypothesis is presented as \(H_0: p = 0.5\). This suggests that regardless of the noise stimulus, the proportion of mice that complete the maze in under 18 seconds remains at 50%. Why do we use a null hypothesis?

• It provides a statement that can be tested and potentially disproven.
• It helps researchers maintain objectivity by proposing no effect until evidence suggests otherwise.

After formulating \(H_0\), researchers gather data via experiments or studies. They then analyze this data to determine whether to reject the null hypothesis, generally aiming to do so only if faced with strong evidence suggesting an effect or difference exists.

The alternative hypothesis, denoted as \(H_a\), is the counterpart to the null hypothesis. It posits that there is indeed an effect or a difference, countering the assumption held by \(H_0\). In our maze exercise, the alternative hypothesis \(H_a: p > 0.5\) suggests that with noise as a stimulus, more than 50% of mice will complete the maze in less than 18 seconds. Roles of the alternative hypothesis in research:

• It allows researchers to express their expectations or predictions regarding the effect being tested.
• \(H_a\) guides the research in shedding light on whether a specific effect or change is real based on the collected data.

When analyzing results, rejecting \(H_0\) in favor of \(H_a\) implies that the evidence supports the existence of the effect or difference asserted by \(H_a\). It means that the researcher can reasonably conclude that the noise indeed influences the time mice take to complete the maze.

Statistical significance is crucial in hypothesis testing as it helps determine whether the observed effect or difference is likely due to chance alone or reflects a real underlying difference. A result is considered statistically significant if it is unlikely to have occurred by random chance, given the null hypothesis is true. Most commonly, a significance level \(\alpha\) is set, often at 0.05, indicating a 5% risk of concluding that a difference exists when there is none. How is statistical significance tested?

• Analyze the collected data to calculate a test statistic, which summarizes the observed evidence.
• Determine the p-value: the probability of obtaining the observed, or more extreme, results assuming \(H_0\) is true.

If the p-value is less than \(\alpha\), the result is deemed statistically significant, leading researchers to reject the null hypothesis. In our example, if the results show that significantly more than 50% of mice complete the maze under 18 seconds with noise, we would conclude the loud noise has a statistically significant effect on maze completion times. This affirmation relies on the evidence rather than random fluctuations or chance.