ANOVA F-Test – AP Statistics

Learning Objectives

Comparing $k$ Means: A t-test compares 2 groups. ANOVA compares $3+$ groups at once. Are all the group means the same ($H_0$)?

Between-Group Variance: How far apart are the specific group averages from the "Grand Average"?

Within-Group Variance: How spread out or noisy is the data inside each individual group?

The F-Ratio Signal/Noise: $F = \frac{Variance\_Between}{Variance\_Within}$. If the groups are far apart (huge Between) but very tight internally (tiny Within), the F-Ratio explodes. We Reject $H_0$!

Tags

ANOVAF-TestVarianceSignal-to-Noise

       ONE-WAY ANOVA F-STATISTIC

       $F = \frac{MS_{Between}}{MS_{Within}} = \frac{\text{Signal}}{\text{Noise}}$

Grand Mean $\bar{\bar{x}}$

Group Averages

Individual Noise (Within)

Signal (Between-Group Distance)

Group Means Spread = 0

Noise (Within-Group Scatter)

Internal Variance ($s^2$) = 100

ANOVA TABLE OUTPUT ($k=3, n_T=30$)

$MS_{bet}$ (Signal): 0.00

$MS_{wit}$ (Noise): 100.0

F-Ratio: 0.00

VERDICT ($P \leq 0.05$ ?)

FAIL TO REJECT H₀

$H_0: \mu_1 = \mu_2 = \mu_3$. When the spread is 0, the groups completely overlap. Any visual gap is just random noise ($F < 1$). To prove they are fundamentally different, you need to pull them far apart (Increase Spread) OR tighten them up severely (Decrease Noise).