Well, time to start this blog off right!
The tagline of my blog is “Anthropology, Modelling, and Statistics.” While the order does not always follow causally, they are ranked in relative importance.
Why do anthropology?
Gilles Deleuze, in Negotiations, wrote that he saw a thread of joy cultivated in the empirical tradition; I believe that this is fundamental to anthropology. We start by observing, learning another language, starting to participate in whatever it is that people do. To hastily start building models from data without understanding where the data came from limits the scope of good research. In anthropology, it is a given that the researcher changes and is changed by the subject they study. This is known as reflexivity, and is a unique focus of cultural anthropologists.
However, I think that since humans are certainly not indeterminably reflexive (we change according to certain inputs or rules), anthropology must seek out links with formal models.
Why do modelling?
Savage Minds, an anthropology blog, had an interview last year with Sarah Kendzior of Al Jazeera English, who has a PhD in cultural anthropology from Washington University. It resonates with my impressions of what cultural anthropology is and how to communicate to others about its findings. One quote stuck out to me:
No one outside the discipline cares about your jargon, your endless parenthetical citations, your paywalled portfolio, your quiet compliance. They care whether you have ideas and can communicate them.
The need for interdisciplinary work and public outreach are vital to anthropology’s mission. And anthropologists are not simply trying to refute others’ beliefs but to understand them. In turn, this has implications for how to communicate anthropology. The anthropologist must use concrete language and rigorous argumentation to support their assertions. Fortunately, applied mathematics has made much of everyday experience describable in terms of math and mathematics is, at its heart, rigorous argumentation.
Since anthropologists must stake out normative determinations, especially in regards to counterfactual policy and everyday interactions, we are moralistically bound to taking the mathematical revolution of the 20th and 21st centuries seriously in our research.
Why do statistics?
We need not give up our stance on reflexivity to build mathematical models. My hunch is that statistics in social science “takes care” of some forms of reflexivity by catching its effects in the error term of the model, or that certain reflexivities are built into the initial model. However, since anthropology looks to describe the complex interactions of symbolism, we know that certain models are bound to fail when they are guided by a limited subset of predictors. Anthropologists also know more about how human behavior can vary across the globe and through time. Hence, anthropologists are in a position to quantify this variance. In other words, anthropology is well-suited for statistical modelling.
Bringing it together.
Not only should anthropologists build models, we should continue critiquing the model building process. I think ethnography can be a great way to get the ground truth about how a model performs – in real life, academia, and business. Through the constant churn of anthropology, modelling, and statistics, we can build models that have real substance and the capacity to predict complex human behaviors. That is the mission of EthnoSitu.