Suppose you are conducting a two-way ANOVA on a dataset with three treatments and four blocks. If the last observation in the third treatment and the second block is altered to be an extreme outlier, what is the most likely effect on the results of the two-way ANOVA?

The between-group sum of squares will increase, making it easier to detect significant differences.

The $F$-statistics will increase, making it more likely to find significant differences.

The results will remain unaffected because ANOVA is robust to outliers.

The within-group (error) sum of squares will increase, causing the $F$-statistics to decrease and reducing the chance of finding significance.