Return to:   The Experimental Method
The experimental method is a controlled procedure that sees the manipulation of an independent variable (IV) in order to observe or measure its effect on a dependent variable e.g. levels of alcohol and measured effect on reaction time.

The ExperimentThe experimental method of research is a controlled procedure that sees the manipulation of an independent variable (IV) to observe and measure any effect this has on a dependent variable (DV). The essential features of the experimental method are then the control, observation and measurement of variables. Illustrating the hallmarks of a science, the use of the experimental method in psychological research makes us more confident about the validity of any cause-effect relationship established between an independent and dependent variable. The experimental method also makes us more confident about the generalisation of its results to a wider population. The experimental method's emphasis on strict procedures helps others replicate the experiment to confirm, or otherwise, the original research findings. This is how scientific knowledge grows. The experimental method sets out to test a null hypothesis, which if rejected allows the researcher to accept the experimental, or research, hypothesis. Hypotheses will, more often than not, be tested at a level of significance. A level of significance is the wager, probability, or p value a researcher places on themselves that their results happened by chance. Students of psychology generally set themselves a level of significance of 1:20, which means they are happy to accept that in the rejection of their null hypothesis, there is still a 5% chance that any cause-effect relationship established by the acceptance of the experimental hypothesis is because of chance or random factors. These chance or random factors, which can be an alternative explanation for results in psychological research, are called extraneous variables and are of two kinds. Random variables just happen, making them impossible to control, and confounding variables, which can be anticipated and controlled. Examples of confounding variables would be any situational variable found in the experimental setting, experimenter variables, such as experimenter and expectancy effect, and participant variables, which are those participant peculiarities that can influence results. These would include participant expectancy and demand characteristics. To control for extraneous variables the experimental method uses a design procedure, which are of two kinds. A related, or within-subjects design like repeated measures or matched pairs, and alternatively, a between-subjects design and unrelated independent group / sample / measures design. Each category, and each type of design has particular features. For example, the repeated measures design, where participants experience all conditions/manipulations of the independent variable, can as we know attract order effect. Counterbalancing can of course control order effect. What design is adopted is often decided in anticipation of confounding variables that c an arise in the different experimental situations, which are the laboratory, field and quasi-experiment. They differ in their location, and thus the degree of control a researcher has over confounding variables. The laboratory experiment often attracts criticism on the grounds of ecological validity. They are sometimes 'No' real '! The further away from the laboratory a researcher goes however, the more likely that the more ecologically valid quasi- or field experiment attracts random and confounding variables. As a research method, the experimental method is assuredly the most rigorous of methodologies in psychology.

Type of Experiment
Laboratory Experiment
  • Cause and effect can be better established

  • Objectivity and precision measurement of DV

  • Control of confounding variables

  • Replication due to standardised procedures (designs/instructions)
Field experiment
  • Difficult to control extraneous variables
  • Difficult to replicate
  • Precision measurement problematic
  • Ethics (consent, deception, privacy etc.)
Quasi experiment
  • Cause-effect difficult to establish between IV and DV
  • No control over IV
  • Little control over extraneous variables
  • Impossible to replicate
  • Only spatio-temporal validity
  • Ethics (consent, deception, privacy etc.)

<<< Back to Research Methods Home

<<< Back to the correlational technique

Forward to the Observational Method >>>

Talk about A Psychology Research Investigation in the forum!
Buy books in Gerard's Bookstore, or keep up to date in music trends with Bobo's Beat!
Switch to:
Research methods and the Correlation
Psychological Processes
Fun Learning and Teaching Stuff
<<< Back
Back to top
Back to top
Back to top
Area 51

[ Sign my Guestbook] - [Read my Guestbook ]