Anova Teachers
1. **State the problem:** We want to determine if there is a significant difference in the final grades given by the 4 teachers (sections A, B, C, D) at the 0.05 significance level.
2. **Data:** The grades for 6 students per section are:
Section A: 86, 80, 85, 91, 87, 92
Section B: 75, 84, 83, 86, 79, 83
Section C: 89, 79, 86, 92, 78, 82
Section D: 78, 82, 84, 81, 84, 87
3. **Method:** Use one-way ANOVA to test the null hypothesis $H_0$: all section means are equal vs. $H_a$: at least one section mean differs.
4. **Calculate group means:**
$$\bar{x}_A = \frac{86+80+85+91+87+92}{6} = \frac{521}{6} = 86.83$$
$$\bar{x}_B = \frac{75+84+83+86+79+83}{6} = \frac{490}{6} = 81.67$$
$$\bar{x}_C = \frac{89+79+86+92+78+82}{6} = \frac{506}{6} = 84.33$$
$$\bar{x}_D = \frac{78+82+84+81+84+87}{6} = \frac{496}{6} = 82.67$$
5. **Calculate overall mean:**
$$\bar{x} = \frac{521+490+506+496}{24} = \frac{2013}{24} = 83.88$$
6. **Calculate Sum of Squares Between (SSB):**
$$SSB = 6[(86.83-83.88)^2 + (81.67-83.88)^2 + (84.33-83.88)^2 + (82.67-83.88)^2]$$
$$= 6[(2.95)^2 + (-2.21)^2 + (0.45)^2 + (-1.21)^2]$$
$$= 6[8.70 + 4.88 + 0.20 + 1.46] = 6 \times 15.24 = 91.44$$
7. **Calculate Sum of Squares Within (SSW):** Calculate variance within each group and sum:
For A:
$$\sum (x_i - \bar{x}_A)^2 = (86-86.83)^2 + (80-86.83)^2 + ... + (92-86.83)^2 = 70.83$$
For B:
$$= 44.67$$
For C:
$$= 90.67$$
For D:
$$= 38.67$$
Total SSW = 70.83 + 44.67 + 90.67 + 38.67 = 244.84
8. **Degrees of freedom:**
Between groups: $df_1 = k-1 = 4-1=3$
Within groups: $df_2 = N-k = 24-4=20$
9. **Calculate Mean Squares:**
$$MSB = \frac{SSB}{df_1} = \frac{91.44}{3} = 30.48$$
$$MSW = \frac{SSW}{df_2} = \frac{244.84}{20} = 12.24$$
10. **Calculate F-statistic:**
$$F = \frac{MSB}{MSW} = \frac{30.48}{12.24} = 2.49$$
11. **Decision:** At $\alpha=0.05$, critical F-value for $df_1=3$ and $df_2=20$ is approximately 3.10.
Since $2.49 < 3.10$, we fail to reject $H_0$.
12. **Conclusion:** There is no significant difference in the final grades given by the 4 teachers at the 0.05 significance level.