go to the School of Psychology home page Go to the UNE home page
Chapter 7 - Analysing the Data Part IV - Analysis of Variance Chapter 1 - Behavioural Science and research Chapter 2 - Research Design Chapter 3 - Collecting the Data Chapter 4 - Analysing the Data Part I - Descriptive Statistics Chapter 5 - Analysing the Data Part II - Inferential Statistics Chapter 6 - Analysing the Data Part III - Common Statistical Tests Operationalism Experimental and non-experimental designs Internal and external validity Between groups Vs repeated measures designs Ethical issues

 

Chapter 2: Research Design

 

Experimental designs

An alternative to the correlational design is the experimental design, or true experiment. The major advantage of an experimental design is that, if properly conducted, it can demonstrate causality. In the example above, we have no way of knowing whether children who watch more violent television differ in some fundamental way (say, in terms of disposition or parental upbringing) from children who donÕt watch violent television. These differences may be the ultimate reason why the children differ in how aggressive they are in the playground. In an experiment, we can get around this problem in two ways. First, we could actually manipulate how much violent television each child was exposed to. Some children in the experiment could be exposed to violent television (the experimental group), while others may be exposed to no violence at all (the control group). So, one feature that distinguishes an experimental study from a correlational study is the degree to which the researcher manipulates exposure to the independent variable (in this case, exposure to violent TV). Second, to eliminate the possibility that children who are exposed to violent television may differ in some way from those who are not (such as age, disposition, upbringing), we would randomly assign children to each of the two conditions in the experiment (exposure or no exposure). After the experimental manipulation, we would then observe the children in a context in which they could display aggression (like we did in the correlational study perhaps). If children exposed to violent television aggressed more than children not exposed to violent television, and all other confounding variables were eliminated as possible explanations for this difference, then we could indeed say that it was the violent television that caused the increased violence. These two features, manipulation of the independent variable and random assignment of participants to the conditions or levels of the independent variable, are what distinguish true experiments from correlational studies. In the absence of these two features, the study is not a true experiment.

While experiments are quite useful for demonstrating cause and effect relationships, they suffer from some major disadvantages. First, good experiments are difficult to conduct. They require a lot of human energy and resources. Second, it takes a lot of ingenuity, cleverness, and experience to design experiments well. Third, experiments often take the behaviour we are interested in out of context. This sometimes produces considerable artificiality and some question how readily we can generalise any findings to other contexts. Finally, in some contexts, there are questions as to how ethical it is to manipulate peopleÕs exposure to the things being studied. For example, if we really believed that exposing children to violent TV does increase aggressiveness are we justified in deliberately exposing such children to the violence?

 

 

 

© Copyright 2000 University of New England, Armidale, NSW, 2351. All rights reserved

UNE homepage Maintained by Dr Ian Price
Email: iprice@turing.une.edu.au