The metrics are bad even before the gaming. Take volunteer work, for example. In theory, they're selecting for people who show a concern for others. In reality, they emphasize organized, secular volunteer work. So they give points for the sort of stupid, futile service projects upper middle class kids do for photo-ops in Africa, but not points for taking care of aging family members, community involvement through the church, etc.

Moreover, they create a huge bias in favor of kids who can afford to spend their free time doing unpaid work, and against kids that need to work to earn a little extra spending money, or kids that need to help with the family business. We're not even talking about "poor" people here. The people hardest-hit by these policies are first-generation middle class people, who don't stand out with a sob story about how they grew up in poverty, but aren't in a position where they can do volunteer work and also have their parents buy them a car. Emphasizing these factors also creates a bias against immigrants. A lot of first-generation Americans just don't feel comfortable engaging in the community in that way.

I'd say it's part of the same discussion if colleges continue to use metrics after they have been gamed of predictive value and now measure something else.