Research Methods

Aim

A general statement to describe what the study intends to investigate

Variables

Any object, characteristic or event that varies in some way within the experiment

Operationalisation

The act of putting the IV and DV into practice by making them measurable

Independent variable (IV)

The variable that is chosen by the experimenter to be manipulated

Dependent variable (DV)

The variable that is chosen by the experimenter to be measured

Extraneous variables

Any variable which, if not controlled by the experimenter, could affect the DV

Confounding variables

Any variable which was not controlled by the experimenter that has affected the DV in a systematic way

Hypothesis

A precise, testable statement that predicts the findings of research

Null hypothesis

A precise, testable statement that predicts that there will be no difference between the two conditions

Experimental hypothesis

A precise, testable statement that predicts that there will be a difference between the two conditions

Directional/one-tailed hypothesis

A hypothesis that predicts the direction of the results by stating exactly how the DV will be affected by the IV

Non-directional/two-tailed hypothesis

A hypothesis that does not predict the direction of results as it merely states the DV will be affected by the IV but does not suggest why

Reason for a directional hypothesis

The direction of results is clear because findings from previous research predict their likely direction

Reasons for a non-directional hypothesis

1. The direction of results is unclear because previous research has not been carried out
2. The direction of results is unclear because previous research has shown conflicting findings
3. The direction of results is unclear because previous research has been heavily criticised

Correlation coefficient

A number between -1 and +1 that indicates the direction and strength of the relationship between two variables

Laboratory experiment

Carried out in a controlled experiment; IV is directly manipulated; participants can be randomly allocated to conditions; all other variables can be controlled = cause and effect can be established (only experiment type where this can occur)

Field experiment

Carried out in a natural environment; IV is directly manipulated; participants can be randomly allocated to conditions

Quasi experiment

Can be like any other type of experiment in any respect, except participants cannot be randomly allocated to conditions because they fall into a specific group e.g. male and female

Natural experiment

A type of quasi experiment where the IV is not manipulated because researchers use a naturally occurring difference in the IV

Strengths of a laboratory experiment

High levels of control over variables and easy to establish cause and effect

Weaknesses of a laboratory experiment

Low mundane realism and high demand characteristics

Strengths of a field experiment

High mundane realism and low demand characteristics

Weaknesses of a field experiment

Low levels of control over variables and lack of informed consent

Strengths of a natural experiment

The research is ethical with low psychological or physical harm and high ecological validity

Weaknesses of a natural experiment

Low levels of control over variables and difficult to establish cause and effect

5 ethical issues in research

1. Deception
2. Informed consent
3. Harm
4. Privacy and confidentiality
5. Right to withdraw

Ways to overcome ethical issues

1. Debriefing
2. Prior general consent
3. Presumptive consent
4. Informed consent
5. Reminders to ensure right to withdraw

Repeated measures

The same participants are used in each condition so they experience all conditions

Strengths of repeated measures

Removes individual differences and clearly shows the effect of IV on DV

Weaknesses of repeated measures

Demand characteristics are more likely and order effects become likely

Independent groups

Different participants are used in each condition so they only experience one condition

Strengths of independent groups

Removes the possibility of order effects and demand characteristics are less likely

Matched pairs

Different but similar participants are used in each condition - an effort is made to match the participants in each condition pre-assessed factors which are relevant to the study

Strengths of matched pairs

Limits individual differences, removes the possibility of order effects and demand characteristics are less likely

Weaknesses of matched pairs

Individual differences may still occur and twice as many participants are needed as independent groups

Order effects

When participants experience practise or tiredness which can affect their performance in the second condition, making results invalid

Counterbalancing (ABBA)

Participants are split into two groups, who experience the conditions in reverse order (Group 1 = AB and group 2 = BA) which balances the effect of order effects

Randomisation

The same as counterbalancing but participants are randomly assigned to whether they experience AB or BA, which balances the effect of order effects

Control

Achieved when all variables other than the IV are held constant

How is control achieved?

Random allocation of participants and standardisation

Demand characteristics

When participants perceive the demands of the study and act accordingly, lowering the internal validity of the research because of a change in their natural behaviour

Single-blind technique

Participants do not know which condition they have been allocated to, which means they are less likely to perceive the demands of the experiment

Experimenter effects

When some characteristic of the experimenter causes participants to behave unnaturally, lowering the internal validity of research

Double-blind technique

Both the participants and the experimenter do not know the aims or conditions of the research, so the researchers are unable to influence the participants in any way

Pilot study

A small-scale version of the real experiment that can be used to check the design works, if participants can easily understand instructions, and if there are any sources of bias that require control, which saves time and money

Random sampling

A sample generated by identifying a list of the entire target population and randomly selecting participants from the list in an unbiased way

Stratified sampling

A sample generated by identifying a list of the entire target population as well as the characteristics which are important to the research (e.g. gender) before randomly selecting a ratio that reflects these groups

Systematic sampling

A sample generated by identifying a list of the entire target population before specifically selecting every nth participant

Strengths of the 3 unbiased sampling methods

High population validity and low experimenter bias

Weaknesses of the 3 unbiased sampling methods

Sometimes low population validity and impractical

Opportunity sampling

A sample generated when participants based on their availability at the time of the study, where the researcher asks available people if they are willing to take part

Volunteer sampling

A sample generated when participants self-select or volunteer to take part in research, normally in response to an advert

Strengths of the 2 biased sampling methods

Practical and ethical

Weaknesses of the 2 biased sampling methods

Low population validity and high experimenter/volunteer bias

Observations

A method where the researcher makes use of a naturally occurring change in the IV to observe the behaviour of participants

Controlled observations

Participants are observed in a controlled setting e.g. a laboratory

Naturalistic observation

Participants are observed in a natural setting e.g. in an office or on a street corner

Covert observation

Participants are not aware they are being observed

Overt observation

Participants are aware they are being observed

Participant observation

When the researcher actively takes part in the group or situation being observed

Non-participant observation

When the researcher does not take part in the group or situation being observed

4 ways to reduce observer bias

1. Behavioural categories
2. Event sampling
3. Time sampling
4. Inter-rater reliability (+0.8 or above)

Strength and weakness of observations

Ethical but have high observer bias

Strength and weakness of controlled observations

Low confounding and extraneous variables but low mundane realism

Strength and weakness of naturalistic observations

High mundane realism but high confounding and extraneous variables

Strength and weakness of covert observations

Low evaluation apprehension but lack of informed consent

Strength and weakness of overt observations

High levels of informed consent but high evaluation apprehension

Strength and weakness of participant observations

Rich qualitative data with high external validity but high experimenter effects

Non-participant observations

Low experimenter effects but less rich qualitative data with lower external validity

Strength of using the mean

It is the most sensitive measure of central tendency because it is the only measure which is representative of all scores

Limitations of using the mean

Can give peculiar outcomes which don’t represent reality e.g. 2.6 children and may be distorted by extreme scores

Strength of using the median

Not distorted by extreme scores

Limitation of using the median

Only represents the middle of the data set

Strength of using the mode

Not distorted by extreme scores

Limitations of using the mode

Only represents the most frequently occurring score and may not be appropriate for small data sets where each score appears only once

Strength of using the range

Easy to calculate

Limitations of using the range

Does not use all the scores and distorted by extreme scores

Strength of using SD

Most precise as it uses all scores in the data set

Limitation of using SD

More difficult to calculate than the range

Case study

In-depth and detailed study of an individual or group, often used to study unique examples of phenomena, which may support or criticise psychological theories

Strengths of case studies

Allow researcher to investigate behaviour that would be unethical to study in controlled conditions and allows researchers to test existing theories

Limitations of case studies

No control over extraneous variables, low population validity and low in reliability

Quantitative data

Numerical data that can be easily and objectively analysed using inferential statistics

Qualitative data

Rich, subjective, non-numerical data, often in the form of written or recorded descriptions, which is detailed but can be more difficult to analyse

Primary data

Data which is collected first-hand by the experimenter

Secondary data

Data which is collected by somebody other than the experimenter and has normally already been published

Meta-analysis

A type of secondary data where researchers combine the primary data from a number of smaller studies to analyse it as one large data set

Self-report techniques

Non-experimental methods where participants respond about themselves, usually by answering questions about their behaviour, attitudes or feelings

Strengths of questionnaires

Low evaluation apprehension and low experimenter bias

Limitations of questionnaires

High volunteer bias and low population validity

Strength of interviews

Rich, qualitative data

Limitations of interviews

High experimenter bias and high evaluation apprehension

Nominal level of measurement

Used when the DV is operationalised by putting the data into named categories or counting the frequency or number of participants in a category

Ordinal level of measurement

Used when the DV is operationalised by putting the data into order or by ranking the data using a scale

Interval level of measurement

Used when the DV is operationalised by putting the data into equal units using a standardised measurement like time, weight or distance

When to use sign test

Test of difference, nominal level and related data

When to use chi square test

Test of difference, nominal level and unrelated data

When to use Wilcoxon T test

Test of difference, ordinal level and related data

When to use Mann-Whitney U test

Test of difference, ordinal level and unrelated data

When to use related T-test

Test of difference, interval level and related data