Message to Prospective Graduate Students

With support from Harvard University, I have been able to admit, on average, one graduate student every other year or so to the lab. Students admitted to the program have full funding that is not tied to faculty. They arrive with a mentor or two in whose lab they will settle, but their funding is theirs and as such they are free to locate and relocate in the lab that provides the best education and launching pad. In spite of the small number of graduate students in my lab at this time, the lab unit that meets weekly consisting of all the individuals involved in research consists of about a dozen people, that additionally includes post-docs, College Fellows, graduate students with primary appointments in other labs at Harvard, thesis writers and precocious undergraduates, and visitors from other universities.

Research revolves around core issues concerning implicit social cognition (ISC) – attitudes, beliefs, and identity. How does the human mind learn about the social world in seemingly automatic and effortless fashion? How closely are our evaluations tied to our conscious sense of our efforts and reflective of our values? I am interested in solving these research questions by focusing on a few specific questions:

How universal are the processes of ISC?

What are the variations in ISC across individuals and cultures?

What is the nature of the association between implicit and explicit social cognition?

What are their origins in childhood and how invariant are they across development?

How accurate are the processes that underlie ISC, at the level of the individual and at the level of decisions made across many individuals?

Are they consistent with Bayes rule, especially when updating information over many iterations, a common occurrence as we interact with the same individuals over long periods of time?

How malleable are the contents of ISC and what inputs are they sensitive to?

Can propositional inputs serve to flip existing implicit attitudes?

Can the output of human minds, visible in language corpora, reveal the content of implicit cognition and if so, can corpora representing the past provide access to social cognition of the past to allow comparisons to attitude change across time?
As an extension, can AI overcome the biases of ISC to provide superior decisions – those that are more accurate and more in line with one’s values?

These questions are meant to be an entry point to understand the minds we have and the biological and social worlds that build and shape such minds. They allow me to focus on a range of constraints and flexibilities inherent in human nature and human experience. The answers we seek are of course psychological, but it is obvious that the quality of our understanding can be enhanced by acknowledging the moral aspects of human behavior: preferences for some over others, giving help to some over others, believing in, assigning credibility, to some over others.

I focus mainly on the troublesome aspects of human minds, less so on the benign, even though I’m interested in measuring improvements in the direction of goodness – decisions that are more accurate and decisions that are more consistent with values. We are not focused on any particular application, as the goal is basic research. It’s important for prospective graduate students to be aware of this and understand that solutions to problems they may care about are unlikely to be realized in their life time. However, because of what we study, the world provides us with occasion to test and speculate about events that occur nearly every day. It is safe for students to assume that they will be taken seriously if they are committed to deep understanding of psychological theory, are proficient at being able to translate ideas into experiments, are exceptionally strong in quantitative methods, and open to mastering new methods and techniques as their needs demand and as technologies become available. Students from this lab routinely serve as the primary TFs in undergrad and graduate statistics courses taught in the department.

The laboratory no longer exists only on campus. Many projects involve data collected via the Internet. It is a good idea for a prospective student to be familiar with web-based programming and handling large datasets. Data are also collected in places where children gather, such as science museums and schools; and from adults at universities, in train stations, bars, inner cities, Fenway Park, emergency rooms, and music conservatories – wherever the appropriate samples for theory testing may be found. Students who excel in this lab tend to select problems of interest within the broad spectrum of what we study, they set their sights high about the sheer amount of data that will be needed to answer their questions, they are usually excited about questions of method and method development in addition to theory, and they are aware that under normal circumstances, publishing at a steady rate of at least wo papers a year while in graduate is a reasonable goal if they wish to return the investment made in them. They are not easily detracted by the problem of studying in a university without a large subject pool. They are creative and downright assertive about finding ways to get their samples and applying for grants that make atypical resources available early in their career.

Assuming that all students who are admitted to Harvard have already demonstrated proficiency in analytic intelligence (using traditional measures that admission committees are provided), it is the unmeasured forms of talent that I am seeking evidence of (not stories). I am not looking for students who have merely been good research assistants in a previous lab. Rather somebody who has used their time, post-college, to develop a mature sense of why they are seeking a PhD in experimental psychology. I would like evidence of creativity, fearlessness, pragmatism, and fire in the belly. I look for students who will be in all places at all times – that is to say actively take advantage of the resources that are available in this department and university beyond the lab. I am certain that students who want to read for a full year before doing research are not a good match. Thinking is for doing, as William James said, and the students who fit best are those who realize that doing experiments is a path to learning which experiments need to be done just as much as reading is.

It is difficult, from the paper record and even from an interview, to gauge a student’s degree of intensity and immersion in an academic career. Although I am open to the possibility that there are geniuses who can sustain a successful academic career while appearing to merely dabble in it, I have never met one. It would be wise to say that those who will be happiest in this lab are those who view their entry into graduate school as a calling rather than as a job or even a career. This is a good place for somebody who understands that the path to success (not assured) is “time on task”.

In addition to the usual skills needed for graduate study in psychology, students in my lab need to be relatively sophisticated about technology. It is necessary to have some background in one or more of the following methods: computer-based experiments, web-based experiments, physiological measurement, measures of brain activation. It helps to arrive knowing R. I have a soft spot for those who have studied in another discipline and are then drawn to psychology; some of the strongest students I’ve had have followed this path. However, with psychology’s increased specialization, such students are welcome if they can demonstrate that they can “think like a psychologist”. Because you are a very good engineer , computer scientist or legal scholar doesn’t mean you’ll be a very good psychologist without learning to think like one. This is visible in the ability to recognize a great idea and separating it from a good one, from the ability to turn an idea into an experiment, and from the skill to get into the weeds of the data and draw accurate interpretations. Much of this is, of course, learned in graduate school, but those who have not immersed themselves in a psychology laboratory during the college years are best off doing so for a couple of years after college, prior to applying to work in my lab.

It is quite often the case that a student who has sophisticated research skills is not necessarily a good writer. This is unfortunate because weak writing can dramatically slow down the pace of a student’s research. So, while good science writing can be taught, it is a sufficiently fundamental skill, that it is best if it is mastered prior to arrival.

I can say without hesitation that my greatest academic pleasure derives from reflecting on the contributions and careers of the group of PhD students and postdocs from my lab. They occupy academic and non-academic positions in the U.S. and other parts of the world. They are leaders in what they have chosen to do. Here you can find a list of them and where they are currently employed. You should feel free to find out from them about what life in Banaji Lab is especially now, as own mentorship style has evolved over time.

It is clear that not all the training you will receive will come from me, perhaps not even most of it. Rather, you’ll benefit from your time here if you recognize that you are in a university that is one of more than 60 colleges and universities in the metropolitan Boston area. That you are in an institution that offers among the strongest support both financially and intellectually. That the department consists of faculty with strong ideas and opinions, but who will bow to your evidence. That the greatest gift you’ll have while here are the other graduate students and what you will each do for the other. I wish you the best in your application to graduate school and as you take these first steps towards your calling – a life of the mind.

COVID-19 update: To the extent that Harvard University supports the graduate program in psychology for 2021 I will likely admit a student. Fortunately for us, our research has been largely online with little need to interact with subjects face-to-face. The only such work we do is to study the social cognition of young children. COVID makes that work difficult so for the next year I will not accept a student interested in the development of social cognition. Rather, we will focus on adult cognition, using the web and the website to collect data.