Ecological Validity

Internal Validity

A critical feature of psychological research—and in general of behavioral research—is that all relevant factors of experiments are under the control of the experimenter, such that unambiguous conclusions can be drawn from data regarding the relation between dependent and independent variables. In other words, the researcher’s goal is a clear, neat, and clean relation between experimental manipulation and measurements. This is called internal validity.

To achieve interval validity, all potential confounding factors need to be eliminated, both subjectwise and experiment-wise. For instance, subjectwise selection might include assuring that participants are physically, psychiatrically, and psychologically healthy and who do not take recreational drugs, smoke, and so on. Furthermore, it includes controlling for variation due to testing at different time points of the day, to testing in different phases of the menstrual cycle, or to testing in men and women. Experiment-wise control includes testing in well-controlled rooms (e.g., temperature-controlled rooms with little distraction) and according to standardized protocols designed to eliminate any environment-related and experimenter-related influence. Finally, with the advent of computers, tasks are easily administered on computers, further eliminating a potential contribution of the experimenter and allowing optimal control of administration of stimuli and other relevant test features.

External Validity: Across Laboratories

[Page 700]A next step in experiments is the question of external validity, which boils down to two features: within and between laboratories and real life. The first important step is whether results are replicable within the same laboratory and across different laboratories. After all, results may be related to the specific sample of participants despite all precautions taken, even given a sufficient sample size, or related to the experimenter, despite all precautions taken to control for this. When results are replicated within and between laboratories, perhaps even across different continents and/or cultures, then by and large, it seems that something robust has been captured. Still, one issue remains, and that is whether results do have any bearing on real life, which pertains to the concept of ecological validity, of which there are three aspects.

External Validity: Ecological Validity

The first aspect is related to the context per se. Although the results may be replicable within and between laboratories, they may be entirely a product of the interaction between the participant and the manipulation in the laboratory; the more so when testing is done under very restricted conditions. For instance, in experiments using functional magnetic resonance imaging scanners, participants are positioned in the scanner, restricting their behavior and movements and limiting test possibilities. Not all participants like to participate in such experiments. In addition, there are a number of exclusionary criteria such as metal in the body, as with orthopedic implants. Finally, many experiments are conducted with university student participants who may not be representative of the general population. Results may thus only apply to a subpopulation participating in a restricted experimental setting, limiting extrapolation to real life.

...