Conducted in a controlled environment so has high internal validity as extraneous variables can be controlled.
- Good control of extraneous variables
- Casual relationships can be determined
- Standard procedures allow replication, improving reliability
- Artificial situations may make participant's behaviour unrepresentative
- Participants may respond to demand characteristics and alter their behaviour
- Investigator effects may bias results
Conducted in a more natural environment. May be possible to control extraneous variables however more difficult than in lab.
- Experimenter effects reduced as participants unaware of being in study.
- Participants in normal situation are likely to behave in representative way.
- Participants likely to be unaware they theure in study so less demand characteristics than in lab
- Controlling extraneous variables more difficult than in lab so researcher is less sure that only IV is effecting DV
- Fewer controls so harder to replicate than in lab experiments
- If particiapnts unaware theyre in study, raises ethical issues
Quasi (natural) experiment
Makes use of existing IVs, such as treatment used for people with schizophrenia. Participants aren't randomly allocated to conditions which may reduce validity.
- Can be used to study real-world issues
- If participants are in their normal situation their behaviour is likely to be representative.
- Participants are likely to be unaware that theyre in the study so demand characteristics will be less problematic
- Used to investigate variables that couldn't practically or ethically be manipulated
- Only possible when naturally occuring differences arise
- Control over extraneous variables is more difficult than in lab experiment
- Researcher isn't manipulating the IV so less sure that its causing changes in DV than in true experiments.
- Gererally can't be replicated
Repeated measures design
Each participant takes part in every condition under rest. Each condition represents one level of the IV. There may also be a control condition
- Good control for participant variables as same person is tested twice. People in group A may be friendlier and thats why they do better on task than group B. In rmd participant variables are the same across conditions
- Fewer participants needed than with independant measures design because if you have 20 participants, you end up with 20 items in final analysis.
- Order effects are produced e.g. participant may do better on second test due to practice effect or may perform less well due to boredom. (can control order effects with counterbalancing)
- Participants may guess the purpose of experiment as they do both conditions so may make aim obvious
- Condition A may be easier than in condition
Independent measures design
Particiapants are allocated to two (or more) experimental groups representing different levels of the IV.There may also be a control group.
- Avoids order effects because each participant is only tested once
- Avoids participants guessing the aims of the experiment
- There is no control of participant variabes, e.g. participants in group A may be more intelligent so they have higher scores on the test.
- Needs more participants than with a repeated measures design because if you have 20 participants there are 10 in each group so you have 10 items in the final analysis.
Matched pariticipants design
Participants who are similar on key variables are paired. One member of the pair is placed in group A and the other in group B. This means that there are two groups of particiapnts. Each group is given one level of the IV.
- Controls for participant variables because of the matching
- Avoids order effects because it is like an independent measures design.
- Very time consuming to match participants on key variables.
- May not control all participant variables because you can only match on variables known to be relevant, but it could be that other variables are important.
Each experimental condition should be presented first or last in equal measure. One way to do this is give half of the particiapnts condition A followed by B and the other half condition B followed by A
In an experiment, any variable other than the independent variable that might potentially affect thee dependent variable and thereby confound the results.
For example, if conducting an experiment on the effect of sunshine and mood, the extraneous variable would be air temperature because it varies along with sunshine.
Cues in an experimental situation that communicate to participants what is expected of them and may then unconciously affect a participant's behaviour.
A researchers expectations or beliefs may encourage certain behaviours in participants. The result is that the researcher's expectations are fulfilled.
Anything that the researcher does that has an effect on a participant's performance in a study, other than what was intended.
Concerns the ability to generalise a research effect beyond the particular setting in which it is demonstrated to other settings. In the case of an experiment the issue is whether the experimental experience represents the actual behaviour being investigated.
Concerns the extent to which the results of a study can be generalised to other groups of people besides those who took part in the study.