Field experiment

A field experiment applies the scientific method to experimentally examine an intervention in the real world (or as many experimental economists like to say, naturally-occurring environments) rather than in the laboratory. Field experiments, like lab experiments, generally randomize subjects (or other sampling units) into treatment and control groups and compare outcomes between these groups. Clinical trials of pharmaceuticals are one example of field experiments. Economists have used field experiments to analyze discrimination, health care programs, charitable fundraising, education, information aggregation in markets, and microfinance programs.

History and Philosophy

Of course, the use of experiments in the lab and the field have a long history in the natural and life sciences. Geology has a long history of field experiments, since the time of Avicenna.cite web|author=Munim M. Al-Rawi and Salim Al-Hassani|title=The Contribution of Ibn Sina (Avicenna) to the development of Earth sciences|publisher=FSTC|url=http://www.muslimheritage.com/uploads/ibnsina.pdf|date=November 2002|accessdate=2008-07-01] Social psychology also has a history of field experiments, including work by pioneering figures Philip Zimbardo, Kurt Lewin and Stanley Milgram. In economics, Peter Bohm, University of Stockholm, was one of the first economists to take the tools of experimental economic methods and attempt to try them with field subjects. Not much work ensued quickly but the use of field experiments in economics has grown tremendously recently with the work of John A. List, Jeff Carpenter, Michael Kremer, Glenn Harrison, Colin Camerer, Bradley Ruffle, Esther Duflo, Dean Karlan, Edward "Ted" Miguel, Sendhil Mullainathan, David H. Reiley, among others.

Applications

Recent work by Glenn W. Harrison (University of Central Florida) and John A. List (University of Chicago) who has put forward a taxonomy of field experiments. See their paper in the December 2004 issue of the "Journal of Economic Literature" for a complete treatment or List's website ( [http://www.arec.umd.edu/fieldexperiments/] ) for a quicker overview. Their taxonomy partitions field experiments into three categories ranging from those that most closely resemble traditional laboratory experiments or compensated survey questions to those that are truly "natural" field experiments in the sense that the subjects involved are unaware of any treatment taking place. (Note that artificial experiments in social psychology often use deception, so that subjects are also unaware of the true treatment).

See List's website for many applications of the field experiment method, including the analysis of public good contributions, charitable giving, market anomalies, discrimination, education, health care and microfinance.

Methodology

Compared with Laboratory Experiments

The major difference between field experiment and laboratory experiment is that in a field experiment there is a limited scope to variable controls, while a typical laboratory experiment has adequate scope to rigorously control variables. The randomization and manipulation of an independent variable is difficult in field experiments. In fact the control of variables is one hallmark of the laboratory experiment, the other being the manipulation of variables independent variable(s). Apart from these characteristics the two methods are very similar - field experiments also require investigators who are specially skilled in a particular area of study. Field experiments can also establish a correlation relationship (direction of relationship included).

Compared with Natural Experiments

A natural experiment is halfway between a field experiment and an observation. In a natural experiment the independent variable is not introduced by the experimenter - for example, the introduction of television on the island of St Helena.

Compared with Non-Experimental Field Data

Caveats

* Fairness of randomization (e.g. in 'negative income tax' experiments communities may lobby for their community to get a cash transfer so the assignment is not purely random)
* Contamination of the randomization
* General equilibrium and "scaling-up"
* Difficulty of replicability (field experiments often require special access or permission, or technical detail-- e.g., the instructions for precisely how to replicate a field experiment are rarely if ever available in economics)
* Limits on ability to obtain informed consent of participants

Prospects

ources

* [http://home.uchicago.edu/~jlist/ John List's homepage]
** [http://www.arec.umd.edu/fieldexperiments/fieldexperiments/Field%20Experiments.pdf"Field Experiments."] GW Harrison, JA List. "Journal of Economic Literature." Vol. XLII (December 2004) pp. 1009–1055.
* [http://www.economics.harvard.edu/faculty/kremer Michael Kremer's homepage]
* [http://econ-www.mit.edu/faculty/eduflo/index.htm Esther Duflo's homepage]
* [http://www.deankarlan.com Dean Karlan's homepage]
* [http://www.hss.caltech.edu/~camerer/camerer.html Colin Camerer's home page]
* [http://www.povertyactionlab.com Abdul Latif Jameel Poverty Action Lab's homepage]
* [http://www.poverty-action.org Innovations for Poverty Action's homepage]

References