DOING EXPERIMENTAL RESEARCH

 

I.                    Introduction.

A.     The goal of experiments is to create a controlled environment in which the effect of the independent variable on the dependent variable can be carefully measured and documented.  Experiments typically seek to document the effect of some factor or condition on the behaviors or attitudes of individuals or small groups.  Few experiments use large organizations, institutions, communities, cities, or entire societies as their unit of observation.

B.     Experiments are not widely used by sociologists.

 

II.                 The logic and design of the classical experiment.

A.     Operationalization of the independent variable.

                                                1.      The independent variable, sometimes called the causal variable, the stimulus, the experimental factor, or the treatment, is carefully controlled and manipulated by the researcher.

                                                2.      Usually a single indicator of the independent variable is identified and manipulated by the researcher.

B.     Operationalizing the dependent variable.

                                                1.      The independent variable's effect on the dependent variable, sometimes called the effect, response, or outcome variable, is determined by carefully measuring the dependent variable before the independent variable is manipulated (pre-test) and again afterward (post-test).  The two measures are compared to determine if the dependent variable has changed.

                                                2.      Questionnaires, indices and scales are often used to operationalize the dependent variable, but observational measures are also common.

C.     Controlling for extraneous factors.

D.     In order to insure that any change in the dependent variable is due to the introduction of the independent variable and not something else, all other factors are carefully controlled.

                                                1.      Physically controlling the environment of the experiment.

                                                2.      Control groups.

a.       The importance of control groups.

b.      Sampling, randomization, and assignment.

c.       Matching.

d.      The use of placebos.

 

III.               Modifications of the classical design.

A.     Improvements and elaborations on the classical design.

                                                1.      The double-blind experiment.

                                                2.      The Solomon Four Group Design.

                                                3.      Multiple experimental groups.

                                                4.      Factorial designs.

                                                5.      Latin square designs.

B.     Quasi-experimental designs.

                                                1.      Equivalent control group, post-test only designs.

                                                2.      Non-equivalent control group, pre-test/post-test designs.

C.     Pre-experimental designs.

                                                1.      One group, post-test only designs.

                                                2.      One group, pre-test/post-test designs.

                                                3.      Non-equivalent control group, post-test only designs.

 

IV.              Other types of experiments.

A.     Time series and regression discontinuity analysis.

B.     Field experiments.

                                                1.      Naturally occurring experimental conditions.

                                                2.      Manipulating the field.

C.     Breaching experiments.

 

V.                 Threats to the validity of experiments.

A.     Internal.

                                                1.      Selection bias.

                                                2.      History, maturation, and attrition.

                                                3.      Testing sensitivity.

                                                4.      Instrumentation.

                                                5.      Statistical regression.

                                                6.      Diffusion, contamination, and imitation.

                                                7.      Compensatory behavior, competition, and demoralization.

                                                8.      Researcher expectations.

B.     External.

                                                1.      Lack of realism.

                                                2.      Problems with generalizability.

                                                3.      Reactivity.

 

VI.              Ethical issues in survey research.

A.     Informed consent, deception, and voluntary participation.

B.     Anonymity and confidentiality.

C.     Manipulation and the risk of harm.