Research Methods

Laboratory

Definition: "The researcher deliberately manipulates the independent variable while maintaining strict control over extraneous variables through standardised procedures in a controlled environment"

Example Studies: - Loftus+Palmer; Bandura; Samuel and Bryant

Strengths: - Shows cause and effect relationship. 
- Has control over extraneous variables.
- Easy to replicate, i.e. highly reliable due to standardisation.

Weaknesses: - Higher risk of demand characteristics
- Lacks ecological validity
- Ethical issues more likely.
- Can have unrepresentative samples. 

Field

Definition: "The researcher deliberately manipulates the independent variable, but does so in the participants own natural environment."

Example Studies: Piliavin et al; Rosenhan experiment one.

Strengths: - Shows cause and effect relationship.
- High in ecological validity.
- Less risk of demand characteristics.

Weaknesses: - Less control over extraneous variables.
- Harder to replicate. 

Quasi/Natural

Definition: "The independent variable is changed by natural occurrence; the researcher just records the effect on the dependent variable. Quasi experiments are anywhere control is lacking over the IV."

Example Studies: Baron-Cohen; Maguire; Griffiths; Sperry.

Strengths: - Can infer cause and effect relationship but is not as strong as IV is not controlled. (e.g. Baron-Cohen)
- High ecological validity. (e.g. Griffiths)
- Can study scenarios which would be unethical/impractical to create. (e.g. Sperry)

Weaknesses: - Less control over extraneous variables. (e.g. Griffiths)
- Harder to replicate. (e.g. Baron-Cohen)

Repeated Measures Design

Definition: A repeated measures design involves using the same participants in each condition of an experiment. E.g. giving a group of participants a driving test with no alcohol, followed at a later time by the same test after a pint of lager.

Example Studies: Samuel and Bryant

Strengths: - No participant variables as you're being compared against yourself.
- Less participants needed as they are used in both conditions.

Weaknesses: - Higher risk of demand characteristics as they see both conditions.
- Risk of order effects: where their performance on a second test will be affected by the first test. i.e. someone gets bored on the second test. 

Independent Measures Design

Definition: Involves using different participants in each condition of the experiment, e.g. giving one group of participants a driving test with no alcohol then a different group of participants the same test after a pint of lager.

Example Studies: Loftus+Palmer; Baron-Cohen; Maguire.

Strengths: - Less demand characteristics as they're only in one condition.
- No order effects as they're only in one condition.

Weaknesses: - Participant variables: being compared to someone different.
- More participants needed as they're only used in one condition. 

Matched Pairs Design

Definition: A matched pairs design involves using different but similar participants in each condition of an experiment. An effort is made to match the participants in each condition in any important characteristics that might affect performance. e.g. in driving ability, alcohol tolerance, etc. (matched on things that may affect the DV)

Example Studies: Raine; Bandura.

Strengths: - Less participant variables that affect DV as they're matched.
- Less demand characteristics as they're only in one condition.
- No order effects.

Weaknesses: - Impossible to completely match participants.
- More participants needed as they're only used in one condition.
- Most time consuming. 

Definition: The researcher observes naturally occurring, spontaneous behaviour in participants own environment.

Strengths: - High ecological validity
- If covert, no demand characteristics.

Weaknesses: - Lack of control so hard to replicate.
- Cannot infer cause and effect. 

Definition: Researcher observes natural behaviours in a controlled/artificial environment.

Examples: Dement and Kleitman; Milgram.

Strengths: - Lots of control involved so easier to replicate, therefore more reliable.
 - Can allow more accurate observations.

Weaknesses: - Behaviour may not be natural so low in ecological validity.
- Risk of demand characteristics. 

Definition: Researcher becomes part of the group/activity they are observing.

Examples: Rosenhan

Strengths: - High ecological validity.
 - If covert, no demand characteristics.
- Detailed, in depth information gained.

Weaknesses: - Difficult to record observation without being found out. 
- Could be risk of demand characteristics.
- Cannot be replicated.
- Subjective. 

Event Sampling
'Recording an event everytime it occurs'

Strengths: - Enables frequencies of behaviours to be recorded; quantitative data. 
- All important events can be recorded and nothing is missed.

Weaknesses: - Time consuming.
- Quantitative data lacks detail. 

Time Sampling
'Records behaviours that occur at timed intervals'

Strengths: - Objective record/Reduces bias
- Quantitative data easily analysed.

 Weaknesses: - May miss behaviours occurring outside of time slot.
- Quantitative data lacks detail.

Questionnaire
'Respondents record their own answers' 

Strengths: - Questionnaires are standardised and can easily be replicated.
- Quantitative data can be generated which is easy to analyse.
 Weaknesses: - Risk of social desirability bias where people lie to make themselves look good, reducing validity.
- Responses are subjective as participants may interpret questions differently.

 Interview
'Face-to-face conversations, can be structured; semi-structured; unstructured' (Structured = pre-determined questions, Unstructured = informal chat)

Strengths: - Structured interviews are standardised.
- Unstructured allow collection of qualitative data which is highly detailed.
Weaknesses: - Structured limit responses so full information not received.
- Unstructured are difficult to analyse due to qualitative data. 

Open
'Allows participants to elaborate their answers and give more detail'

Strengths: - Produces qualitative data which provides detail so participants can express opinions, increasing validity.
- Analysis retains detail of participants answers, so information is not lost.

Weaknesses: - Qualitative data is difficult to analyse.
- Interpretation can be subjective, leading to bias/lacks validity.

Closed 
'Force participants to choose a pre-determined option'

Strengths: -Easy for people to respond to, large amounts of data collected.
- Quantitative data easy to analyse.

Weaknesses: - Quantitative data lacks detail, cannot express opinions fully.
- Response bias (always saying yes) /social desirability bias

Definition: - A measure of how strongly two variables are related.

Positive Correlation: - High values of one variable are associated with high values of the other. (i.e as one variable increases, so does the other)

Negative Correlation: - High values of one variable are associated with low values of another. (i.e. As one variable increases, the other decreases)

Correlation Coefficient: - Found using Spearmans Rank Test with ordinal data.

Hypotheses: One Tailed - Say whether there will be a positive or negative correlation.
Two Tailed - Say there will be a correlation or relationship.
Null - Simply state there will be no correlation.
NEVER SAY IV/DV/DIFFERENCE OR EFFECT
Always operationalise variables (i.e. There will be a relationship between no. words recalled and the amount of time spent sleeping)

- Variables must give a range of numbers; - You can just analyse data and don't need participants; - Do not manipulate variables.

Population - The group of people that we want to investigate.
- A sample should be representative (i.e. gender balance, age range/characteristics should be typical of target population) and large (to reduce risk of distortion by anomolous/freak values)

Opportunity - Selecting participants who are available at the time of doing the study.
Strengths: - Quick and easy.
Weaknesses: - Selection bias (i.e. selecting people you think you'll benefit from asking)
- Unrepresentative (as all participants are similar as found from same environment)

Random: - Every person has an equal chance of being chose; start with a list of everyone in the target population, get computer and use database to choose names.
Strengths: - Should be representative but won't be if using a small sample.
- No selection or researcher bias as everyone has an equal chance of being chosen.
Weaknesses: - Time consuming
- Some participants may not take part. 

Self-Selected: - People volunteer to take part in response to advertisements/invitation.
 Strengths: - Less chance of sample attrition; More convenient as they come to you.
Weaknesses: - Atypical participants may want to help; Unrepresentative. 

Nominal Data: - Involves counting the number of subjects that fall into categories. The numbers refer to the frequency with which something occurs. (e.g. The number of students who have blue, brown or green eyes) Uses bar charts.

Ordinal Data: - Numbers represent positions within a group. Participants are given scores and ranked. (e.g. The number of words recalled from a memory test between year 10 and year 7 students) Uses bar charts.

Interval Data: - An interval scale uses equal intervals. (e.g. The length of time a person takes to complete a test, or the IQ scores a person gets) Uses histograms.

Statistical tests are used to find out what the probability is that a result has occured by chance. If P < 0.05 then there is a less than 5% probability that the results are due to chance and therefore the null hypothesis is rejected.

Experiments: Independent Measures Design:
With Nominal data (Chi 2); With Ordinal data (Mann-Whitney U)

Repeated Measures Design:
With Nominal data (Sign Test); With Ordinal data (Wilcoxon)

Mode: - The value or event that occurs the most frequently.
Advantages: - Not influenced by extreme scores, useful to show most popular values.
Disadvantages: - May be no mode, or more than one.

Median: - The middle value when all scores are placed in rank order.
Advantages: - Not distorted by extreme freak values.
Disadvantages: - Can be distorted by small samples and are less sensitive.

Mean: - The average value of all the scores.
Advantages: - Most sensitive measure of central tendency.
Disadvantages: - Can be distorted by extreme freak values.