- Is it true that people 'like' those who are similar to them?
- Why is it so? Is there an evolutionary explanation?
Since you mentioned that you want an evolutionary explanation, there is one available. In biology the effect of providing benefit towards potential non-kin based on an arbitrary marker is known as the green-beard or armpit effect. In a social human setting, if the marker is arbitrary social construct it is usually known as ethnocentrism. This sort of behavior is studied in game theory in the context of cooperate-defect games (typical example: Prisoner's dilemma) and usually called conditional altruism.
It has been shown that conditional altruism evolves in a simple spatial agent-based model and promotes cooperative behavior (Hammond & Axelrod, 2006a 2006b). The effect does not create cooperation, but if there is another mechanism for creation of cooperation (say spatial factors in the H&A model) then ethnocentrism helps maintain it and extend the range of parameters under which cooperation can occur (Kaznatcheev & Shultz, 2011).
In humans, this ability to cooperate only with others of similar culture is believed to require a significant amount of cognitive ability. In fact, some even suppose that it could have been one of the factors that drove towards the increasing complexity of our brains. Unfortunately, Kaznatcheev (2010a) shows that the ethnocentrism of the sort present in the H&A models is not robust to increase in the cost of cognition. Thus, in humans (or simpler organisms) the mechanism allowing discrimination has to have been in place already (and not co-evolved) or be very inexpensive.
The above examples dealt with the prisoner's dilemma (PD) which is a typical model of a competitive environment. In the PD cooperation is irrational, so there conditional altruism allowed the agents to cooperate irrationally (thus moving over to the better social payoff), while still treating those of a different culture rationally and defection from them. This doesn't seem as bad, but Kaznatcheev (2010b) shows that the mechanism of ethnocentrism is robust across different games (not just PD) including ones where cooperation is rational. In those games conditional altruism produces an irrational defection from the out-group. Thus, from an evolutionary stand-point this is a two-edged sword: it can cause unexpected cooperative behavior, but also irrational hostility.
You may want to read up about "homophily". It is often summarised with the phrase "birds of a feather flock together". "Heterophily" relates to when people are attracted to those that are different to them.
There was a review article by McPhereson et al (2001) which you might like to read. To quote the abstract: