91Ó°ÊÓ

In probability and statistics, a probability distribution represents all the possible outcomes of a random experiment and the likelihood of each outcome. For instance, when we roll a fair six-sided die, each face—numbered from 1 to 6—has an equal probability of 1/6. Since all outcomes are equally likely with a fair die, the probability distribution is uniform.

However, real-world scenarios can deviate from this ideal distribution. When conducting experiments, like tossing a die multiple times, the observed frequencies might differ slightly due to random variation. To better understand how these observed differences occur, statisticians compare them to expected frequencies as a way to determine if any underlying shift from our assumed distribution might exist. This lays the foundation for methods like the Chi-squared test.

A contingency table is a type of data table that displays the frequency distribution of variables. In our die example, the table you saw organizes the outcomes of rolled faces and their respective frequencies. It's a compact way to show all this information in an accessible format.

Here's how it works: each row in the table represents a category of the variable tested—here, each face of the die—and each column could represent a different level of another variable (in this case, it's just counts of each face). For the Chi-squared test, the table allows us to easily compare observed frequencies (like the 18 times a 1 was rolled) against what would be expected if the assumptions about the die being fair were true.

This organization not only aids in clear data management but is crucial for proceeding with statistical tests to validate or reject hypotheses, such as fairness of a die.

Expected frequency is a concept that anticipates the number of times an event should happen based on probability. For our six-sided fair die tossed 120 times, the expected frequency for each number is derived from the probability distribution:

• Each number has a 1/6 chance of being rolled.
• Therefore, expected frequency for each face is \(\frac{1}{6}\times 120 = 20 \).

This means that, theoretically, each face should show up 20 times if the die is unbiased. In practice, actual results can differ slightly due to chance.

By comparing these ideal frequencies to the real observed ones, the Chi-squared test measures how much they differ to evaluate whether the die behaves fairly or not. This comparison tells us if there is any statistical significance in the difference between observed and expected values.

Hypothesis testing is a method used to make decisions based on statistical data. We start by proposing a null hypothesis, which is a general statement or default position that there is no effect or no difference. In our die example, the null hypothesis is that the die is fair.

Alternatively, the alternative hypothesis suggests that the die is not fair and that one or more sides are more likely to appear than others during a roll. Our goal is to use statistical tests to determine whether we can reject the null hypothesis in favor of the alternative.

In this process, the Chi-squared test is used to analyze the differences between observed frequencies and what we would expect if the null hypothesis were true. A significant Chi-squared statistic suggests that the difference is noteworthy enough to question the fairness of the die. If the Chi-squared value is smaller than a critical value for a chosen confidence level, we do not reject the null hypothesis. Otherwise, if it is large enough, we reject the null hypothesis, suggesting the die might not be fair.