Experiments

Independant variable - the thing you change

Dependant variable - the thing you measure

Control group - everything kept the same

Experiment group - manipulates things in order to see the cause and effect

Lab experiments - take place in a controlled environment

Field experiments - take place in a real world environment

All causes for something are controlled by the researcher. 

These causes are variables.

Study of obediance - 1963 at Yale University

Wanted to see how people respond conform to authority figures. 

A participant was told to increase the voltage of shocks going to the actor. 30% of the participants went all the way. 

65% went all the way if they couldn't hear the screams. 

• This experiment was very unethical as there is a lot of deception. 
• It is reliable as it would be easy to repeat because it is numerical data. 
• The validity is questionable because of the Hawthorne Effect. People might have acted differently because they were being pressured or because of the environment. 
• It is a small sample size (40) so is less representative. 

Boys were dressed in girls clothers and vice versa.

The adults took care of the children:

• They offered "sophie" a doll and soft toys. 
• The girl dressed in boys clothes were taken care of with more directive behaviour. They were offered robots and cars. They were more physical. 

The research was deceptive but wasn't harmful in any way. It wouldn't be the most practical because you would need to have a facility and find volunteers, but it wouldn't be the most expensive. It could be repeated and would be preffered by positivists, because you can compare the responses on the vonunteers. The sample size was small so the representativeness is lower. 

72 nursery children divided into groups. 

24 were part of a control group which didn't see an adult using violence. The other 48 were in experimental groups, 24 seeing agressive models and 24 seeing non-agressive models. 

Children exposed to the violent model copied the behaviour. The children seeing non-agressive models were less agressive than the control group. 

This research would be fine ethically because there is no deception and no harm. It is reliable because it could be repeated, however the validity would be lower because the children are being influenced and are not neccessarily being violent but are just copying what the adults did. It is just a inflatable toy so it is hard to claim that they are violent. The research is likely to be represnatitive because there are lots of people. It would be a very time-consuming experiment. 

Take place in the real world or natural settings.

This means that they are higher in validity because the subject is unlikely to know that they are part of an experiment. This gets rid of the Hawthorne Effect which makes it more valid.

However, the representativeness is lower because there can be a lot of reasons for different results in experiments. 

An actor wearing a suit and bowler hat periodically asked for directions to Hyde Park. 

They then asked the same question but in different clothes. 

He argued that variations in responses must have been reactions to the apparent social class of the person. This is because it was the only variable that was changed. 

This could be true, but the validity is worsened by the fact that people may just be in different siuations when asked the question, such as rushing for a train. 

In a SF school (1968) they told teachers that students were taking an iq test. They chose students at random and claimed that they were the top achievers. The students chosen actually did achieve better grades which shows the power of labelling and sfp. 

This was unethical because they lied about the selections and it meant that some students weren't supported as much, whereas the randomly selected kids had a better chance to perform. It is very valid as tests were taken, but harder to repeat. This shows that it would be favoured by interpretivists. 

P - Staff, cost (lab hire), time

E- People always give consent, but they don't always consent to the true experiment. It can harm people (especially mentally) 

R - easy to repeat because it is easy to control variables and record the data.

V - Not very natural and can be difficult to isolate single causes.

R - Experiments are often small scale which makes it lower. Volunteer sampling can be biased.

T - Positivists

P - Staff, cost, time

E - Consent, potential for harm

R - Difficult to control all variable

V - Natural setting so more authentic

R - Small sample size 

T - Interpretivists