What is an Economic Experiment?
An experiment in economics is based on the possibility to reproduce one part of a real economic behaviour in a laboratory. In this way, the experimenter can control and replicate the observed behaviour in order to identify the role played by one or more
definite elements, named variables.
So, an economic experiment is similar to a scientific experiment.Carrying out an experiment requires subjects and tools; for the experiment to be effective, a protocol of rules needs to be strictly followed. In economics, the experimental subjects are the economic behaviours. As a result, usually the experimental subjects are real people. So, economic experiments can be viewed as a “synthetic” economic reality. Tools can be divided into material tools and conceptual tools. Material tools include a laboratory (physical location), a group of subjects, and a set of techniques (mainly computer based, sometimes on paper, rarely oral). Conceptual tools can vary according to the kind of experiment, but a common denominator is the possibility of “comparison” whereby the importance of a particular variable is studied by comparing behaviours with and without the studied variable. So, an economic experiment requires different subjects experiencing different situations, which are called treatments. It may happen that two (or more) treatments are experienced by two (or more) groups (between-subjects procedure); or it may happen that two (or more) treatments are experienced by a single group (within-subjects procedure). In any case, only one of the treatments includes the variable under examination, so that the environment is isolated and controlled).
Moreover, an economic experiment requires strict consistency with precise procedures, named protocols. In particular, instructions, incentives and fairness towards experimental subjects are considered compulsory.
Instructions must be clear and complete. Incentives should be monetary and well balanced to the required efforts. This is considered to be important to ensure an adequate motivation in experimental subjects; the practice of paying participants emerged as early as the 1950s and, according to Roth (1995), it can be read as the answer to Wallis and Friedman’s (1942) criticisms. According to these criticisms, the hypothetical nature of experimental choices made experimental results weak. Recently, the idea emerged that if it is true that the reliability of experimental data requires a good motivation, at the same time, it is also true that a good motivation does not necessarily require monetary incentives.The diffusion of economic experiments has also brought some criticisms about the relationship between experiments and reality (Sugden 2005). It is a problematic relationship in both directions: when one tries to reproduce reality in the laboratory and when one tries to apply the laboratory results to reality. The central issue of this relationship is the basic difference between the simplicity of experimental situations and the complexity of the correspondent reality.