On the diet and exercise, does anyone have links to some proper studies? This is exactly the sort of thing where someone could have run a study in which they chose 500 men, asked them about their exercise habits and then measured their testosterone and found the two were correlated. That does not prove causation, though, in fact it would seem more likely that cause and effect were the other way round, i.e. those whose testosterone was naturally high felt more need for exercise.
A study would ideally recruit men, measure their starting testosterone level, then divide into three groups: one doesn't change anything, one increases the amount of exercise and the other decreases the amount of exercise then the researchers measure again to see what change there has been. Likewise with each dietary change that is thought to have an effect.