Experimental Design Essay ⋆ Political Science Essay Examples ⋆ EssayEmpire

This example Experimental Design Essay is published for educational and informational purposes only. If you need a custom essay or research paper on this topic please use our writing services. EssayEmpire.com offers reliable custom essay writing services that can help you to receive high grades and impress your professors with the quality of each essay or research paper you hand in.

Experiments offer a unique opportunity to examine the nature and direction of causal relationships between variables. Experiments represent a particularly effective and useful method when investigators need clear causal information, past studies have generated inconsistent or contradictory findings, multi-method validation is desired, or methodological triangulation may prove helpful in uncovering complex or subtle underlying processes.

Experiments typically refers to laboratory studies in which the investigator retains control over the recruitment, assignment to random condition, treatment, and measurement of participants. There are three important ways in which experiments differ from other forms of social measurement. First, experimenters do not wait for events to occur and then observe and measure them; rather, experimenters create such situations, capture exactly what they are interested in studying, and measure those variables. Second, experimenters manipulate conditions so that some participants receive a control condition while others receive the treatment condition; then experimenters observe any systematic differences between groups or individuals, keeping everything the same except the manipulated variable. Third, experimenters randomize participants to conditions to ensure that observed differences do not result from preexisting differences between participants. Experimenters typically report average treatment effects and remain less concerned with the effect of a particular treatment on a given individual. Therefore, in analyses of results, demonstrating no effect may reflect a true finding but may result from additional factors that exert opposite effects on different kinds of people.

Experiments offer several important advantages. In addition to the ability to determine causal connections and randomize participants to conditions, experiments also allow scholars to take precise measurements of variables. Whether such measurements involve paper-and-pencil tests, physiological or behavioral measures, or other forms of self-report, experimenters can design and test the measurement strategies best suited for the scale and topic they are investigating. In this way, experiments allow researchers to begin to explore the micro foundational processes underlying a wide variety of social and political processes.

Experiments do not come without certain disadvantages. Experiments often present seemingly artificial situations and environments to unrepresentative participant pools. Artificial environments often result because it is either impossible or unethical to create the situation that the investigator wants to study. In addition, certain participants may or may not be suitably representative of the population most relevant for application.

Another important concern relates to the consequences of noncompliance among participants. Most experiments analyze data according to intent to treat, wherein participants are randomly assigned to treatment or control and the data are analyzed regardless of adherence to treatment or subsequent noncompliance or deviation from the experimental protocol. This can bias results if a participant’s withdrawal from the experiment resulted from the treatment itself and not from some extraneous factor.

Experiments possess two kinds of validity, internal and external. Internal validity refers to whether a researcher is actually getting at the variables he or she claims to be investigating. External validity refers to whether particular findings can generalize to a larger, external reality. In general, psychologists are much more concerned with maintaining internal validity, whereas political scientists care much more about external validity.

Basic experimental design can take one of several forms. Within-subject studies examine changes within an individual as the result of a manipulation over time. Between-subjects studies explore the differences between people under these conditions. Often, it may be ethically impossible to compare individuals, as when a scientist wants to understand the effects of a certain disease; in such cases, a matching paradigm is used whereby control individuals are found who match the affected target participant in every way possible except for the existence of the illness or the other variable of interest. In matching designs, manipulations occur within the participant, but effects might then be compared across participants. Participants should never be matched to examine differences in a natural dimension such as race or sex because too many other variables are intertwined with such characteristics to make such a comparison valid or viable.

Field experiments and natural experiments are other kinds of experiments. Field experiments allow investigators to randomly assign individuals or groups to treatment and control. Although analyses must take account of additional contextual factors that can influence observed outcomes, they also permit an investigation of phenomena that might be hard to reproduce or study in a laboratory setting. In this way, there is a trade-off between a gain in external validity and a potential loss in internal validity. Natural experiments occur when an investigator takes advantage of a natural occurrence, such as a fire or hurricane, to examine particular consequences, but these experiments do not allow investigators to randomize participants across conditions.

Experimental incentives often differ. Participants in most economic experiments receive pay; participants in many experiments in psychology are given extra credit in return for participation. Finally, experimental ethics remains an important part of the process. Some experiments involve deception; others, such as those in economics, never do. All experiments should obtain informed consent from participants.

Bibliography:

Aronson, Elliot, Pheobe Ellsworth, J. Merrill Carlsmith, and Marti Gonzales. Methods of Research in Social Psychology. 2d ed. New York: McGraw-Hill, 1990.
Druckman, James, Donald Green, Jim Kuklinski, and Arthur Lupia. Handbook of Experimental Political Science. New York: Cambridge University Press, forthcoming.
Kagel, John, and Alvin Roth, eds. Handbook of Experimental Economics. Princeton, N.J.: Princeton University Press, 1995.
McDermott, R. “The Experimental Method in Political Science.” Annual Review of Political Science 5 (2002): 31–61.
“Experimental Methodology in Political Science.” Political Analysis 10, no. 4 (2002): 325–342.