91Ó°ÊÓ

To understand the concept of expected value, let's consider an example involving jury selection. The idea is to predict the average outcome if we were to repeat a process several times. In the context of our exercise, we are asked to calculate the expected number of African Americans in a jury pool of 200 people, provided that the population is 24% African American. This is crucial because the expected value serves as a benchmark to decide whether an observed outcome is unusual under the null hypothesis.

Using the formula for expected value, which in probability theory is given by the formula \( E(X) = n*p \), where \( n \) is the sample size and \( p \) is the probability of the event occurring, we can carry out our own calculations. So if the null hypothesis assumes a perfectly proportional representation, we multiply the number of jury members (200) by the proportion (0.24) to predict an average of 48 African Americans in the jury pool. Whenever you come across such problems, remember that the expected value is not a guarantee of what will happen in a single trial, but a long-term average across many trials.

When dealing with categorical data, like the number of African Americans in a jury pool, we often use the binomial test to evaluate the observed outcomes against the expected outcomes under the null hypothesis. A binomial test compares the number of successes (in this case, selected African Americans) to what we would expect to see by chance. It’s based on the binomial distribution, which is a type of probability distribution with two possible outcomes: success or failure.

In our example, 'success' is selecting an African American for the jury pool. If the null hypothesis is true, with a jury pool of 200, about 48 members should be African American. Any significant deviation from this expected number might suggest a non-random process at play. To complete a binomial test, we would calculate the probability of observing the actually counted African Americans in the jury pool (e.g., 40 in Pool A and 26 in Pool B), which would then generate a p-value to interpret.

The p-value is a pivotal concept in statistical hypothesis testing. It measures the probability of an observed result (or one more extreme) occurring by random chance if the null hypothesis is true. A smaller p-value indicates that the observed data is less likely to occur under the null hypothesis, hence suggesting that the null hypothesis may not adequately explain the observed data.

In the context of our jury pool, the p-value for Pool A and Pool B would indicate how likely we are to see such a distribution of African American jurors purely by chance. A lower p-value in Pool B compared to Pool A suggests that a pool with 26 African American members (well below the expected 48) is less likely to occur due to random variability alone. This may prompt further investigation into possible reasons beyond chance that explain this discrepancy. The commonly accepted threshold to determine 'statistical significance' is a p-value lower than 0.05, which suggests that the result is unlikely enough under the null hypothesis that it prompts considering alternative explanations.