Intro to Business Statistics: Analysis of Variance ANOVA - ANOVA Table, BetweenGroups WithinGroups Total Sum of Squares Mean Squares FRatio

What This Is

The Analysis of Variance (ANOVA) table is a statistical tool used to compare the means of three or more groups to determine if there is a significant difference between them. A retail chain wants to know if average daily sales exceed $10,000 in different regions. They collect data on daily sales from three regions: North, South, and East. The ANOVA table helps them determine if the average daily sales are significantly different across these regions.

Key Formulas & Symbols

• Between-Groups Sum of Squares (SSB): SSB = k * (n * (x^2 - x?^2)) where k = number of groups, n = sample size, x = grand mean, x? = group mean.
• Within-Groups Sum of Squares (SSW): SSW =? (n_i * (x?_i - x?)^2) where n_i = sample size for group i, x?_i = group mean for group i, x? = overall mean.
• Total Sum of Squares (SST): SST = SSB + SSW.
• Mean Squares Between (MSB): MSB = SSB / (k - 1) where k = number of groups.
• Mean Squares Within (MSW): MSW = SSW / (N - k) where N = total sample size, k = number of groups.
• F-Ratio: F = MSB / MSW.
• Degrees of Freedom Between (dfB): dfB = k - 1 where k = number of groups.
• Degrees of Freedom Within (dfW): dfW = N - k where N = total sample size, k = number of groups.

Step-by-Step Procedure

1. State hypotheses: H?: = = ... = ?_k (no difference between groups) vs. H_a: at least one group mean is different.
2. Choose test: ANOVA is used for comparing means of three or more groups.
3. Compute test statistic: Calculate SSB, SSW, SST, MSB, MSW, and F.
4. Find p-value or critical value: Use the F-distribution table or calculator to find the p-value or critical value.
5. Compare to ?: Compare the p-value to? (default = 0.05) or compare the test statistic to the critical value.
6. Conclude: If p <-or test statistic > critical value, reject H? and conclude that at least one group mean is different.

Common Mistakes

• Mistake: Misinterpreting the F-statistic as a measure of effect size.
• Correction: The F-statistic is a ratio of variances, not a measure of effect size. Use other metrics like Cohen's d or ?² to measure effect size.
• Mistake: Failing to check assumptions (normality, equal variances).
• Correction: Check assumptions before conducting ANOVA. If assumptions are not met, consider alternative tests like non-parametric ANOVA or Kruskal-Wallis.
• Mistake: Misusing ANOVA for paired or repeated measures data.
• Correction: Use ANOVA for independent samples only. For paired or repeated measures data, use other tests like paired t-test or repeated measures ANOVA.

Quick Practice Problems

1. A company wants to compare the average salaries of three different departments. The data is as follows:

What is the F-statistic?

Answer: F = 10.5 (calculated using the formulas above) Explanation: The F-statistic is calculated by dividing the mean squares between (MSB) by the mean squares within (MSW).

1. A researcher wants to compare the average exam scores of three different teaching methods. The data is as follows:

What is the p-value?

Answer: p = 0.01 (calculated using the F-distribution table or calculator) Explanation: The p-value is calculated by comparing the F-statistic to the F-distribution table or using a calculator.

1. A company wants to compare the average sales of three different regions. The data is as follows:

What is the degrees of freedom between (dfB)?

Answer: dfB = 2 (calculated using the formula dfB = k - 1) Explanation: The degrees of freedom between is calculated by subtracting 1 from the number of groups.

Last-Minute Cram Sheet

1. ANOVA is used for comparing means of three or more groups.
2. F = MSB / MSW.
3. dfB = k - 1.
4. dfW = N - k.
5. p-value is NOT the probability that H? is true – it’s the probability of observing the data (or more extreme) if H? is true .
6. Assumptions: normality, equal variances.
7. F-statistic is a ratio of variances, not a measure of effect size .
8. Use ANOVA for independent samples only .
9. Check assumptions before conducting ANOVA.
10. Cohen's d or ?² are used to measure effect size.