Types of Experiment
- Laboratory Experiment: One conducted in laboratory conditions is high in Internal Validity as lots of Extraneous Variables can be controlled. Having such control makes the study easier to replicate, however such controlled conditions may not be a reflection of real life.
- Field Experiment: One conducted in a more realistic environment, it may be possible to control some extraneous variables but not as much as in a lab experiment. Participants are usually not aware they are being studied therefore the effect of an experimenter is less. However Demand Characteristics may still occur one the IV is changed.
- Natural Experiment: One conducted upon a natural setting with existing IV's in place, for example on treatment for a mental disorder. Causal relationships and conclusions cannot be drawn from this sort of experiment.
Repeated Measures: Participants can be made to do both conditions. However Demand Characteristics can happen, to avoid this use Counterbalancing where you swap the order of conditions
For example half the people do Cond1 then Cond2 the rest do Cond2 then Cond1
Independent Groups: Participants are split up and only do one condition each.
Matched Pairs: Participants get paired up with others who have similar things to themselves which can be assesed through many things, IQ, demographics etc.
This can be very expensive however and time consuming and there's a possibility the pairs are matched with the incorrect things.
Simply observing people..
They use different types of sampling techniques to measure behaviour such as Time Sampling, say for instance measuring something every 30 seconds or Event Sampling say every time a behaviour occurs tally it.
Observational studies can have problems where the observer may not be looking at the relevant things or may focus on the wrong thing, they may even have Observer Bias whereby their own expectations effect the results they record.
This is a good method as it allows the psychologist to see what people actually do rather than what they say they do.
It allows a researcher to establish where a relationship may exist between two variables, however there may be other extraneous variables which are not taken into account which will have an effect on the two that are examined causing faulty data.
These sort of studies lack internal and external validity.