If genes change in proportion because they make people more or less likely to reproduce, that's "natural selection".
If the genes change in proportion just due to randomness (they're all equally fit, but some were passed on more times than others as a fluke), we call it "genetic drift" [0].
Evolution involves mutation and gene flow [1] adding new genes to a population, and natural selection and genetic drift changing the relative proportion of various genes within the population.
I didn't mention any changes due to random sampling. I was careful to qualify that the choice of mate is influenced by genes. Wouldn't that qualify as natural selection more than genetic drift? Although, now that I think of it, the two are probably more like directions on a continuum, since choice of mate would be influenced by both genes and random sampling. Is that accurate?
If choice of mate is influenced by genes, and there's a difference in fitness level due to those genes, then it would be natural selection. That is, if Alice-Bob have more kids than Claire-Dave for genetic reasons, natural selection is favoring AB over CD.
If choice of mate is influenced by genes, but there's no difference in fitness level, then you're looking at a natural drift scenario. That is, if Alice-Bob have more kids than Claire-Dave due to some random effect like happening to meet much earlier in life, AB's gains over CD are classified as drift.
Both effects are usually present, but sometimes one is much more significant than the other.
If the genes change in proportion just due to randomness (they're all equally fit, but some were passed on more times than others as a fluke), we call it "genetic drift" [0].
Evolution involves mutation and gene flow [1] adding new genes to a population, and natural selection and genetic drift changing the relative proportion of various genes within the population.
[0] http://en.wikipedia.org/wiki/Genetic_drift
[1] http://en.wikipedia.org/wiki/Gene_flow