SAMPLING DISTRIBUTIONS - T, F, χ²

In statistics, a sampling distribution is a probability distribution of a statistic that is derived from all possible samples of a given size from a population. Understanding these distributions is crucial for making inferences about a population based on a sample. When the population standard deviation is unknown, or the sample size is small, we often rely on specific probability distributions like the t-distribution, F-distribution, and Chi-square (χ²) distribution. These distributions are fundamental tools for hypothesis testing and confidence interval estimation.

The t-Distribution (Student's t-distribution)

The t-distribution is used when we are estimating the mean of a normally distributed population and the population standard deviation is unknown. In such cases, we use the sample standard deviation as an estimate. The t-distribution is similar to the normal distribution (Z-distribution) in that it is bell-shaped and symmetric around zero. However, it has heavier tails, meaning extreme values are more likely than in a normal distribution. This is because the sample standard deviation is a less precise estimate than the population standard deviation, leading to more uncertainty.

The shape of the t-distribution depends on a parameter called the 'degrees of freedom' (df). The degrees of freedom are related to the sample size. For a one-sample t-test, df = n - 1, where n is the sample size. As the degrees of freedom increase, the t-distribution approaches the standard normal distribution.

Key Characteristics of the t-Distribution:

• It is bell-shaped and symmetric about the mean (0).
• The mean, median, and mode are all 0.
• It is defined for all values of t from -∞ to +∞.
• The variance is greater than 1 (specifically, df / (df - 2) for df > 2).
• The tails of the distribution are heavier than the normal distribution, indicating a greater probability of extreme values.
• The shape depends on the degrees of freedom (df).

Uses of the t-Distribution:

• One-Sample t-Test: To test if the mean of a single population is equal to a specific value when the population standard deviation is unknown.
• Independent Samples t-Test: To compare the means of two independent groups.
• Paired Samples t-Test: To compare the means of the same group at two different times or under two different conditions.
• Confidence Intervals for the Mean: To construct confidence intervals for the population mean when the population standard deviation is unknown.

Example: One-Sample t-Test

Suppose a researcher wants to test if the average height of a certain breed of dog is 25 inches. They take a sample of 10 dogs and find the sample mean height to be 26.5 inches with a sample standard deviation of 2 inches. Here, the population standard deviation is unknown. The degrees of freedom would be df = 10 - 1 = 9. The researcher would use a t-distribution with 9 degrees of freedom to determine if the observed sample mean is significantly different from 25 inches.

The F-Distribution

The F-distribution is used primarily for comparing the variances of two or more populations. It is also used in the context of Analysis of Variance (ANOVA) to test if the means of multiple groups are equal. The F-distribution is derived from the ratio of two independent chi-square distributions, each divided by its respective degrees of freedom. It is always non-negative and is skewed to the right.

The F-distribution has two sets of degrees of freedom: the numerator degrees of freedom (df1) and the denominator degrees of freedom (df2). These degrees of freedom are associated with the variances being compared or the groups being analyzed in ANOVA.

Key Characteristics of the F-Distribution:

• It is non-negative (values range from 0 to infinity).
• It is skewed to the right.
• The shape depends on two parameters: the numerator degrees of freedom (df1) and the denominator degrees of freedom (df2).
• As df1 and df2 increase, the F-distribution approaches a normal distribution.
• The mean of the F-distribution is approximately df2 / (df2 - 2) for df2 > 2.

Uses of the F-Distribution:

• F-Test for Equality of Variances: To test if the variances of two populations are equal.
• Analysis of Variance (ANOVA): To test if there are any statistically significant differences between the means of three or more independent groups.
• Regression Analysis: To test the overall significance of a regression model.

Example: F-Test for Equality of Variances

A quality control manager wants to know if the variability in the weight of product A is the same as the variability in the weight of product B. They collect samples from both products and calculate their sample variances. The F-statistic is calculated as the ratio of the larger sample variance to the smaller sample variance. This F-statistic is then compared to a critical value from the F-distribution with appropriate degrees of freedom (related to the sample sizes of A and B) to decide if the variances are significantly different.

The Chi-Square (χ²) Distribution

The Chi-square distribution is a probability distribution that arises from the sum of squared standard normal random variables. It is used in several types of statistical tests, particularly those involving categorical data and tests of variance.

Like the t-distribution, the Chi-square distribution is characterized by its degrees of freedom (df). It is a non-negative distribution and is skewed to the right. The skewness decreases as the degrees of freedom increase, and the distribution becomes more symmetric.

Key Characteristics of the χ² Distribution:

• It is non-negative (values range from 0 to infinity).
• It is skewed to the right.
• The shape depends on the degrees of freedom (df).
• The mean of the χ² distribution is equal to its degrees of freedom (df).
• The variance of the χ² distribution is equal to twice the degrees of freedom (2 * df).

Uses of the χ² Distribution:

• Chi-Square Test for Independence: To test if there is a significant association between two categorical variables in a contingency table.
• Chi-Square Test for Goodness-of-Fit: To test if an observed frequency distribution fits an expected distribution.
• Chi-Square Test for Variance: To test if the variance of a single population is equal to a specific value.
• Used in conjunction with the F-distribution (as F is a ratio of chi-squares).

Example: Chi-Square Test for Independence

A market researcher wants to know if there is a relationship between a person's preference for a certain brand of coffee and their age group (e.g., young, middle-aged, senior). They collect data and create a contingency table showing the observed frequencies of preferences across age groups. A Chi-square test for independence is used to determine if the observed counts are significantly different from what would be expected if there were no association between brand preference and age group. The degrees of freedom for this test are calculated as (number of rows - 1) * (number of columns - 1).

SAMPLING DISTRIBUTIONS - USES IN SIGNIFICANCE TESTING

Significance testing, also known as hypothesis testing, is a core concept in inferential statistics. It's a formal procedure used to determine whether the results of a study are statistically significant or likely due to chance. The sampling distributions of the t, F, and χ² are the foundation upon which these tests are built. They provide the probability framework needed to decide whether to reject or fail to reject a null hypothesis.

The General Framework of Significance Testing

Regardless of the specific test (t-test, F-test, χ² test), the process of significance testing follows a similar structure:

1. State the Hypotheses:
   • Null Hypothesis (H₀): A statement of no effect or no difference. It represents the status quo or a baseline assumption.
   • Alternative Hypothesis (H₁ or Hₐ): A statement that contradicts the null hypothesis, suggesting an effect or difference exists.
2. Set the Significance Level (α): This is the probability of rejecting the null hypothesis when it is actually true (Type I error). Common values are 0.05 (5%), 0.01 (1%), or 0.10 (10%).
3. Choose the Appropriate Test Statistic: Based on the type of data, the research question, and assumptions about the population (e.g., known/unknown standard deviation, sample size, type of variables), select the correct test (t-test, F-test, χ² test).
4. Determine the Critical Region/Calculate the p-value:
   • Critical Region Approach: Find the critical value(s) from the relevant sampling distribution (t, F, or χ²) based on α and the degrees of freedom. The critical region consists of values of the test statistic that would lead to rejecting H₀.
   • p-value Approach: Calculate the p-value associated with the observed test statistic. The p-value is the probability of observing a test statistic as extreme as, or more extreme than, the one calculated from the sample, assuming H₀ is true.
5. Make a Decision:
   • If the calculated test statistic falls within the critical region (or if the p-value ≤ α), reject H₀.
   • If the calculated test statistic does not fall within the critical region (or if the p-value > α), fail to reject H₀.
6. Interpret the Results: State the conclusion in the context of the original research question, considering the significance level and the decision made about H₀.

Using Sampling Distributions in Significance Testing

1. t-Distribution in Significance Testing

The t-distribution is used when testing hypotheses about population means when the population standard deviation is unknown and the sample size is relatively small (often n < 30), or when the population is assumed to be normally distributed.

• One-Sample t-Test:
  • H₀: μ = μ₀ (Population mean equals a specific value)
  • H₁: μ ≠ μ₀ (two-tailed), μ > μ₀ (right-tailed), or μ < μ₀ (left-tailed)
  • Test Statistic: t = (x̄ - μ₀) / (s / √n), where x̄ is the sample mean, s is the sample standard deviation, n is the sample size, and μ₀ is the hypothesized population mean.
  • Decision: Compare the calculated t-value to the critical t-value from the t-distribution with df = n - 1, or compare the p-value to α.
• Independent Samples t-Test:
  • H₀: μ₁ = μ₂ (Means of two independent populations are equal)
  • H₁: μ₁ ≠ μ₂, μ₁ > μ₂, or μ₁ < μ₂
  • Test Statistic: Depends on whether equal variances are assumed. If equal variances are assumed, a pooled variance is calculated. If not, separate variances are used, resulting in a more complex df calculation (Welch's t-test).
  • Decision: Compare the calculated t-value to the critical t-value with appropriate df, or compare the p-value to α.
• Paired Samples t-Test:
  • H₀: μ<0xE1><0xB5><0xA0> = 0 (The mean difference between paired observations is zero)
  • H₁: μ<0xE1><0xB5><0xA0> ≠ 0, μ<0xE1><0xB5><0xA0> > 0, or μ<0xE1><0xB5><0xA0> < 0
  • Test Statistic: t = d̄ / (s<0xE1><0xB5><0xA0> / √n), where d̄ is the mean of the differences, s<0xE1><0xB5><0xA0> is the standard deviation of the differences, and n is the number of pairs.
  • Decision: Compare the calculated t-value to the critical t-value with df = n - 1, or compare the p-value to α.

2. F-Distribution in Significance Testing

The F-distribution is used when testing hypotheses about the equality of variances or comparing means of three or more groups.

• F-Test for Equality of Variances:
  • H₀: σ₁² = σ₂² (Variances of two populations are equal)
  • H₁: σ₁² ≠ σ₂²
  • Test Statistic: F = s₁² / s₂² (where s₁² is the larger sample variance)
  • Decision: Compare the calculated F-value to the critical F-value from the F-distribution with df1 = n₁ - 1 and df2 = n₂ - 1 (for a two-tailed test, use α/2 and adjust df if needed, or use the upper tail critical value for α and consider the ratio of variances).
• One-Way ANOVA:
  • H₀: μ₁ = μ₂ = ... = μ<0xE2><0x82><0x96> (All population means are equal)
  • H₁: At least one population mean is different.
  • Test Statistic: F = (Variance between groups) / (Variance within groups) = MSB / MSW. MSB (Mean Square Between) and MSW (Mean Square Within) are calculated from sums of squares and degrees of freedom.
  • Decision: Compare the calculated F-value to the critical F-value from the F-distribution with df1 = k - 1 (k = number of groups) and df2 = N - k (N = total number of observations).

3. Chi-Square (χ²) Distribution in Significance Testing

The χ² distribution is used for tests involving categorical data or for testing hypotheses about a single population variance.

• Chi-Square Test for Independence:
  • H₀: The two categorical variables are independent.
  • H₁: The two categorical variables are dependent (associated).
  • Test Statistic: χ² = Σ [ (O - E)² / E ], where O is the observed frequency and E is the expected frequency in each cell of the contingency table. Expected frequencies are calculated based on the assumption of independence.
  • Decision: Compare the calculated χ² value to the critical χ² value from the χ² distribution with df = (rows - 1) * (columns - 1).
• Chi-Square Test for Goodness-of-Fit:
  • H₀: The observed distribution fits the expected distribution.
  • H₁: The observed distribution does not fit the expected distribution.
  • Test Statistic: Same formula as the test for independence: χ² = Σ [ (O - E)² / E ].
  • Decision: Compare the calculated χ² value to the critical χ² value from the χ² distribution with df = (number of categories - 1).
• Chi-Square Test for Variance:
  • H₀: σ² = σ₀² (Population variance equals a specific value)
  • H₁: σ² ≠ σ₀², σ² > σ₀², or σ² < σ₀²
  • Test Statistic: χ² = (n - 1) * s² / σ₀², where s² is the sample variance and σ₀² is the hypothesized population variance.
  • Decision: Compare the calculated χ² value to the critical χ² value from the χ² distribution with df = n - 1.

LARGE-SAMPLE TESTS FOR MEANS AND PROPORTIONS

When dealing with large sample sizes, the Central Limit Theorem (CLT) becomes very powerful. The CLT states that the sampling distribution of the sample mean (and sample proportion) will be approximately normally distributed, regardless of the shape of the population distribution, provided the sample size is sufficiently large (typically n ≥ 30 is considered large enough). This allows us to use the standard normal distribution (Z-distribution) for hypothesis testing and constructing confidence intervals, simplifying calculations compared to using t-distributions for means.

Large-Sample Tests for Means (Z-tests)

These tests are used when the population standard deviation (σ) is known, or when the sample size is large (n ≥ 30) and the sample standard deviation (s) is used as a reliable estimate of σ.

1. One-Sample Z-Test for the Mean

• Hypotheses:
  • H₀: μ = μ₀
  • H₁: μ ≠ μ₀ (two-tailed), μ > μ₀ (right-tailed), or μ < μ₀ (left-tailed)
• Assumptions: Population is normally distributed OR sample size is large (n ≥ 30). Population standard deviation (σ) is known OR sample size is large enough for s to be a good estimate of σ.
• Test Statistic: Z = (x̄ - μ₀) / (σ / √n)
• Decision: Compare the calculated Z-value to the critical Z-value from the standard normal distribution (Z-table) corresponding to the significance level α, or calculate the p-value and compare it to α.

Example: A manufacturer claims their light bulbs last an average of 1000 hours. A consumer group takes a sample of 100 bulbs and finds the sample mean lifetime to be 980 hours, with a sample standard deviation of 50 hours. Assuming the sample size is large enough to use s as an estimate for σ, they can perform a Z-test to see if the mean lifetime is significantly less than 1000 hours.

2. Two-Sample Z-Test for the Difference Between Means

Used to compare the means of two independent populations when sample sizes are large.

• Hypotheses:
  • H₀: μ₁ - μ₂ = D₀ (Often D₀ = 0, meaning μ₁ = μ₂)
  • H₁: μ₁ - μ₂ ≠ D₀, μ₁ - μ₂ > D₀, or μ₁ - μ₂ < D₀
• Assumptions: Both samples are independent and randomly selected. Both sample sizes are large (n₁ ≥ 30 and n₂ ≥ 30). Population standard deviations (σ₁ and σ₂) are known OR sample standard deviations (s₁ and s₂) are used as estimates.
• Test Statistic: Z = [(x̄₁ - x̄₂) - D₀] / √[(σ₁²/n₁) + (σ₂²/n₂)]. If σ₁ and σ₂ are unknown, use s₁ and s₂ in their place.
• Decision: Compare the calculated Z-value to the critical Z-value or compare the p-value to α.

Example: A company wants to compare the average sales performance of two different marketing strategies. They implement strategy A in region 1 and strategy B in region 2. After a period, they find the average sales for strategy A (sample size 50) is $15,000 with a standard deviation of $2,000, and for strategy B (sample size 60) is $16,500 with a standard deviation of $2,500. Using the large-sample Z-test, they can determine if there's a significant difference in average sales.

Large-Sample Tests for Proportions (Z-tests)

These tests are used to make inferences about population proportions when sample sizes are large.

1. One-Sample Z-Test for a Proportion

• Hypotheses:
  • H₀: p = p₀
  • H₁: p ≠ p₀ (two-tailed), p > p₀ (right-tailed), or p < p₀ (left-tailed)
• Assumptions: Sample is random. Sample size is large enough such that np₀ ≥ 10 and n(1-p₀) ≥ 10.
• Test Statistic: Z = (p̂ - p₀) / √[p₀(1-p₀)/n], where p̂ is the sample proportion (x/n), p₀ is the hypothesized population proportion, and n is the sample size.
• Decision: Compare the calculated Z-value to the critical Z-value or compare the p-value to α.

Example: A political pollster wants to test if the proportion of voters supporting a candidate is greater than 50%. They conduct a survey of 400 voters and find that 220 support the candidate (p̂ = 220/400 = 0.55). Using a Z-test, they can determine if this sample proportion is significantly greater than 0.50.

2. Two-Sample Z-Test for the Difference Between Proportions

Used to compare the proportions of two independent populations.

• Hypotheses:
  • H₀: p₁ - p₂ = D₀ (Often D₀ = 0, meaning p₁ = p₂)
  • H₁: p₁ - p₂ ≠ D₀, p₁ - p₂ > D₀, or p₁ - p₂ < D₀
• Assumptions: Both samples are independent and randomly selected. Sample sizes are large enough such that n₁p̂ ≥ 10, n₁(1-p̂) ≥ 10, n₂p̂ ≥ 10, and n₂(1-p̂) ≥ 10, where p̂ is the pooled proportion calculated under the null hypothesis (p̂ = (x₁ + x₂)/(n₁ + n₂)).
• Test Statistic: Z = (p̂₁ - p̂₂) / √[p̂(1-p̂)(1/n₁ + 1/n₂)], where p̂₁ and p̂₂ are the sample proportions for group 1 and group 2, respectively.
• Decision: Compare the calculated Z-value to the critical Z-value or compare the p-value to α.

Example: A pharmaceutical company is testing a new drug designed to reduce cholesterol. They conduct a study with two groups: one receiving the new drug and a control group receiving a placebo. After the study, they find that 60 out of 200 patients in the drug group showed improvement (p̂₁ = 0.30), while 40 out of 200 patients in the placebo group showed improvement (p̂₂ = 0.20). A two-sample Z-test can determine if the proportion of improvement is significantly higher in the drug group.