The conclusion ended up the same, but in the future you should consider normalizing against the size of the population you're drawing from.
In this case, rather than "business undergrad CEOs vs. eng undergrad CEOs" you should have "business undergrad CEOs per 1000 business undergrads vs. eng undergrad CEOs per 1000 eng undergrads".
Or the more wonky form of P(CEO | biz undergrad) vs. P(CEO | eng undergrad) ("probability of being a CEO given a business/eng undergrad degree...").
In this case, rather than "business undergrad CEOs vs. eng undergrad CEOs" you should have "business undergrad CEOs per 1000 business undergrads vs. eng undergrad CEOs per 1000 eng undergrads".
Or the more wonky form of P(CEO | biz undergrad) vs. P(CEO | eng undergrad) ("probability of being a CEO given a business/eng undergrad degree...").