Four Factor Design ~ Homoscedasticity Problem


For school, I have decided to do a project on airline reliability; specifically, looking at on-time departures. I decided to use a four factor design with my factors being: airline, airport, day, and time.

The data has all been collected. I'm verifying that the conditions of ANOVA hold. However, the condition of common variance does not for airport, day, and time. I used Levene's test to determine this for each factor. I know I must use a variance stabilizing transformation, however, no matter which one I try it doesn't seem to work. I have used the Box-Cox (?) to help me determine which transformation to use. The Lambda's I have obtained are: 0.3083, 0.12361, and 0.67 (for airport, day, and time respectively).

I'm not sure where I am going wrong. Am I right to use Levene's test for each factor? Any help would be greatly appreciated as I am currently at a loss.