Define Replication Method.
Repeating your experiment to make it reliable.
Define Confound.
Uncontrolled Independent Variable.
Define Independent Variable and which axis it is usually measured on.
Variable that is manipulated by us to see if affects the dependent variable. This is measured on the X axis.
Define Dependent Variable.
Measure of the behaviour we are interested in, it is affected by the independent variable. This is measured on the Y axis.
Define what makes an experiment reliable in a psychology context. Provide an example.
If you can repeat the experiment over and over and get the same response then you have reliable results.
For a study to be reliable the same experiment must be conducted under the same conditions to generate the same results.
Describe what makes your data valid in an experiment and provide an example.
Your data is valid if it answers what you wanted to know.
Define Population in a research context.
All info/behaviours that we are interested in e.g. height, weight, score.
Define what a sample is.
A representative group of your population.
Describe 2 ways sampling can go wrong.
- Sampling Bias
2. Sampling Error
Define sampling bias with an example.
Sampling Bias is when your sample is systematic.
For example, if I wanted to know the strength of all the people in my lecture but only chose men, this would be sampling bias.
For Example, if I want to know whether Blue or Pink is more liked by children I need to have a mix of of ages, races, sexes in order to avoid sampling bias.
Does Sampling Bias effect the validity or reliability? and why?
Sampling Bias effects the validity because you are not getting the true answer to your question.
Define Sampling Error.
Sampling Error is when random samples are drawn from the same population and give different results. Happens by chance and is unavoidable but can be minimised by using large sample groups.
3 things to ask when looking at a graph.
- What are you measuring? (y axis)
- How are you measuring it? (x axis)
- What is the shape of the results?
Describe Observational Design.
Observational Design is when we are observing and looking for a correlation between 2 dependent variables. There is no independent variables.
List some critiques of Observational Design.
- You cannot conclude anything because it has not been manipulated.
- It is often the only choice for ethical and practical reasons.
- Does not require sampling so it is more sound.
Correlation does not imply causation, explain this.
Just because you see a correlation or link between 2 things does not mean they are the reason or cause of one another.
Explain the basic principles of Experimental design.
- You manipulate the Independent Variable to see if the effect on the dependent variable.
- This can show or imply causation.
- It is the more powerful design because you can identify causation.
- It is not always the ethical or practical option but where possible we should use it.
Name the 3 types of experimental design.
- Within Subject Design
- Between Groups Design
- Matched Pairs Design
Explain Within Subject Design. Provide an example.
This is when the subjects is exposed to all experimental condition of the independent variable. For example if I want to test the meteor of someone on the memory enhancing drug i would have subjects take the real pill and then at a different session take the placebo pill.
List Critiques of Within Subject Design.
- It is good for controlling some confounds because the subject is exposed to all options. E.g. They may have better memory naturally.
- There is a strong possibility of external confounds as there is time between each session. E.g. sleep.
- There can be a Risk of Order Effect. This is when you would have everyone take the same pill at the same time. You can remove the external confound but breaking it up and having some take it and one point while other take the second pill.
Explain Between Groups Design
This is when each subject only participates in 1 level of the experiment and then we compare the groups.
List Critiques of the Between Groups Design
- This is good for controlling external confounds because there is no time difference and less change.
- Possibility of subject confounds because you are using a different group of people for each level however you can minimise this by using random allocating.
Explain Matched Pairs Design
This is when you give some sort of pretest before the experiment then match subjects up according to their results. You then assign them to alternate groups.
List Critiques of the Matches Pairs Design.
- This keeps the external confounds constant and nearly keeps the subject confounds constant too.
- Labour intensive.
