Identification, please

Show me your structural modelWhat are the keys to success in the art world? Some combination of hard work, smart early career choices, artistic talent, access to the connected and the gate-keepers, and a bit of serendipity. The same could be said of many professions. But it’s very hard to discern the relative importance of these ingredients. Research finds that those who have attained the greatest successes in creative fields tend to be more likely to think that merit is what got them there (in this they are not much different from the successful in other arenas), while the less successful think luck and other factors have more to do with it.

An optimistic view tends to believe that for all the random factors that influence success, in the end the very successful must in fact be the most worthy. Although the QWERTY design of keyboards arose from an effort to avoid the keys jamming in manual typewriters, and that a more rational arrangement of letters might better serve the key-less computer age, it might well be that QWERTY survives because it’s really not so bad,  and that if it really were bad, we would have moved on to some other keypad. The pessimist thinks we get stuck with all sorts of mediocre outcomes because of what happens to get adopted first – that QWERTY is crummy, that Justin Timberlake is a star because of the cumulative advantages that accrue to his childhood lucky breaks.

We have divergent views because we don’t have a good way of modelling success that could distinguish between the “actual talent” and the “lucky breaks” hypotheses. In the parlance of empirical social science, our models are underidentified, and the results we observe are consistent with multiple explanations. This is not really a problem that so-called “big data” can solve. Obtaining data on hundreds of thousands of artistic careers won’t get us around the fundamental problem. That many artists with a lifetime of success tended to have early success does not indicate why they had early success.

And so the recent publication in Science by a team of five data scientists, “Quantifying reputation and success in art”, has to be read with the caution that although they work with an immense data set on artists’ careers and the galleries that show their work, finding that “early access to prestigious central institutions offered life-long access to high-prestige venues and reduced dropout rate”, they haven’t actually solved the question that opens this post. This is not to say that some artists do not face genuine barriers – geographic, socio-economic – in gaining access to the art world, or other professions that have a structure of gatekeepers, and that policies to help level that playing field would be just, and inspire more young people to higher aspirations. But the paper (which after all, is by specialists in analyzing networks and data, not the specific deep culture of the gallery world) doesn’t delve into who has that early access, and why were their works chosen for exhibition. That’s not a criticism of the methods used in the paper, which are truly innovative (their visual of the “coexhibition network” of galleries and museums is quite something) in terms of what they are trying to illustrate. But for now, we haven’t resolved the optimist vs pessimist question; that would require not even bigger data sets on artists careers, but a different sort of data altogether.

Leave a Reply

Your email address will not be published. Required fields are marked *