Here’s an article on how human eyes might be indicators of personality. More pits in the iris means correlates with “tender, warm, and trusting” personalities, and more curves around the edge correlates with neurosis and impulsiveness.
The most important objection here, it seems to me, is that correlation does not necessarily indicate causation. However, the study did provide at least some provisional notes on causation. One gene, PAX6, controls embryonic iris development. Scientists have also found that PAX6 mutating correlates with “impulsiveness and poor social skills.”
I suppose this just puts the correlation ≠ causation charge at one remove. In other words, the crucial decision to make is whether PAX6 mutations cause social issues. If you grant that, then this study of irises makes perfect sense.
Even if the study is sound, I think, this particular “biomarker” would vary in importance based on the individual and on the culture. David Schmitt has argued that “humans evolved a pluralistic mating repertoire that differs in adaptive ways across sex and temporal context, personal characteristics […], and facultative features of culture and local ecology” (259).1 That argument seems readily portable to questions of more general personality evaluation (that is, those in which mate choice is not necessarily the genre of decision-making).
Anecdotally, I find myself mistrustful of people with small pupils.
1 “Fundamentals of Human Mating Strategies.” In The Handbook of Evolutionary Psychology. Buss, David M., ed. Hoboken, NJ: John Wiley & Sons, Inc., 2005. 258-91.