Experiments
An experiment is systematic observation and measurement of a number
of variables of interest within a controlled environment. An experiment
will frequently have two or more groups:
- a test or experimental group
- a control group
In the experiment, the test group is subjected to some intervention
(for example, a treatment or exposure to some information); the control
group receives no intervention.
The purpose of the control group is to ensure that any changes in key
variables resulting from the intervention can be measures in the test
group.
Problems:
- Must have clear measures for comparing during experiment
- Time interval is critical
- Tests before and after can be learned
- Selecting comparable control groups poses difficulties
- People react to experimental situation, both subjects and
researchers.
Issues to Consider
- What method of randomisation to groups is to be used?
- Different design - which should I choose?
- Blinding - should the subjects know they are part of the test
group?
- Ethical Issues - what should I tell people who are in the
experiment
An example of the use of an Experiment in a 'Lifestyle Survey'
setting could be the use of exercise referral programmes. Here selected
subjects are assessed, various baseline measurements made, prior to
receiving an intervention (a planned programme of exercise or other
behavioural change activities). Measurements are again made at the end
of the programme. A matched control group (not subject to the
intervention) also receives the baseline and post-activity measurement. |
