Human Experiments. Egor Lazarev’s Impressions about the participation in workshops at the school ICPSR

Egor Lazarev told about his academic trip to the summer school ICPSR(Ann Arbor, Michigan, USA), where he took part in two short intensive workshops: “Design, Conduction and Analysis of Field Experiments in Social Sciences” (August, 1-5) and “Spatial Econometrics” (August 15-19), which have been delivered by professor Donald Green (Columbia University) and by professor Robert Franzese (University of Michigan), respectively.

Egor Lazarev

Egor Lazarev told about his academic trip to the summer school ICPSR(Ann Arbor, Michigan, USA), where he took part in two short intensive workshops: “Design, Conduction and Analysis of Field Experiments in Social Sciences” (August, 1-5) and “Spatial Econometrics” (August 15-19), which have been delivered by professor Donald Green (Columbia University) and by professor Robert Franzese (University of Michigan), respectively.

After receiving a grant to travel to school ICPSR (it is the mecca of political scientists and sociologists studying the quantitative methods of data analysis) I initially signed up to take the courses “Regression III” and “Advanced Topics in Maximum Likelihood Estimation”. I had no idea what it is, but it sounded impressive. It has turned out that both of these courses included serious math, which has almost no relation to social sciences. However, fate had spared me: the American Embassy did not allow me to get a visa for a couple of days and as a result I had to sign up for the short intensive seminars instead of the general school program. This time I was more farsighted and reasonable in my choice of the workshops and I chose the programs which I was really interested in. In addition, I chose the courses of the most famous lectors with sky-high citation indeces who are professors at Yale and Michigan Universities.

Ann Arbor, Michigan, USA

Donald Green, who delivered the course on experiments, began his first lecture saying that all the participants of his seminar have been deceived because the organizers have slipped them the professor of Columbia University instead of the professor of Yale University. It happened so that not long ago professor Green decided to change job after 23 years in Yale where he has leaded the huge institution, and to move to New York just to live in a big city.

One of Professor Green's former colleagues described him as a “preacher” - a person who believes in the rightness of his own research with a religious force and because of the gift of persuasion converts students to his “religion”. Such description turned out to be extremely accurate: during all five seminar’s days from 9 am to 5 pm all students were sitting and listening to Green’s lectures as bewitched.

Green’s current “religion” is experiment as social science technique. These experiments are based on studies of the cause-effect relation between the phenomenon, which the researcher is interested in, and some random interference generated by scholar. Ideal experiments can be found in medical science, when patients are randomly selected into experimental and control groups (the first group receives medicine, the second one undergoes placebo effect or receives nothing), and then the doctors trace the effect whether this or that drug works or not. The effectiveness of this approach is in randomization to one basis (for example, receiving of medication) erases all the significant differences between groups, or, to be more precise, it makes possible to ignore them.

In recent years, the research in the framework of randomized experiments has become very popular in economics of developing countries. Scientists led by Esther Duflo from MIT began to randomize everything: distribution of micro-credits, vouchers for university studies, construction of roads and wells, garbage disposal, the appointment of women as heads of villages, etc. At the same time, scholars began to exploit random distribution that was organized by the government or by history. Various lotteries and even national borders became the object of these experiments.

Ann Arbor, Michigan, USA

In political science experiments became popular as well. Professor Green, for example, won fame by experimenting with factors which cause voting turnout. Their idea was the following: some of the voters were invited to come to the polls by phone, someone was invited by mail, someone was invited personally, and someone was not disturbed. They compared the results of different ways of warning and look what happened.

At the first glance the results do not seem fascinating. The scholars found out that personal calls raise voting turnout by 2.4 percent. But, on the other hand, these results were obtained in a strictly scientific way. No wonder that all the other ways of non-experimental research in social sciences (even if they are based on complicated Regression methods and Maximum Likelihood Estimation) are not taken seriously by professor Green and his colleagues.