RESEARCH METHODS UNIT 1 PSYCHOLOGY REVISION

Validity - the extent to which research measures what it claims to measure.

THERE ARE TWO TYPES OF VALIDITY, INTERNAL AND EXTERNAL VALIDITY!

Internal Validity - the extent to which research generates information in the reseach area, i.e. if it is controlled well with no extraneous variables.

External Validity - the extent to which the findings of research can be generalised and applied to the public.

THERE ARE 3 TYPES OF EXTERNAL VALIDITY!

Ecological Validity - if it can be generalised to every day life

Population Validity - if it can be generalised to all groups of people

Historical Validity - if it can be generalised to different historical times

GENERALLY IF A STUDY HAS HIGH INTERNAL VALIDITY IT HAS LOW EXTERNAL VALIDITY AND VISA VERSA

Reliability

Reliability - refers to the consistency of a measure. A measure is considered reliable if we get the same results repeatedly.

IN TERMS OF OBSERVATIONS THERE IS TWO TYPES OF RELIABILITY!

Intra-Observer Reliability - If the single observer is consistant in applying the criteria.

Inter-Observer Reliability - If a team of observers are applying the criteria to the same standard. Are they STANDARDISING their criteria

Variables

These are the types of variables:

Independent - the characteristic of a psychological experiment that is manipulated or changed.

Dependent - the variable that is being measured.

Confounding - variables that the researcher fails to control or eliminate.

Extraneous - any variable other than the independent variable that could cause a change in the dependent variable.

Situational - things in the environment that may impact an experiment.

Participant - indavidual characteristics that may impact how they respond in an experiment (this includes SOCIAL DESIRABILITY).

Control - something that is held constant to eliminate extraneous.

Experimental Designs

THERE ARE THREE TYPES OF EXPERIMENTAL DESIGN:

Independent Measures Design - where different participants are used in each condition of an experiment.

Repeated Measures Design - where the same participants participate in all conditions of an experiment.

Matched-Pairs Design - where different but similar participants are used in each condition of an experiment. These participants can be matched on things like age and ability.

INDEPENDANT MEASURES DESIGN

+ Less chance of order effects

- Time consuming

- No control over participant variables

REPEATED MEASURES DESIGN

+ Fewer participants needed

+ Participants variables eliminated

- Order effects (Practice and boredom)

MATCHED PAIRS DESIGN

+ Less chance of order effects

- Time consuming to match participants

- Can't control all participant characteristics

Hypothesis

Hypothesis - a testable statement predicting the precise relationship between variables. In experimental designs a hypothesis will suggest that there will be a difference between two or more groups as a results of manipulation of one of them (i.e. changing the dependent variable).

Null Hypothesis - a testable statement predicting that there will be no significant difference between the variables identified in the experimental hypothesis.

A HYPOTHESIS CAN BE DIRECTIONAL OR NON-DIRECTIONAL!

Sampling

THERE ARE SIX TYPES OF SAMPLING:

Random Sampling - All have an equal chance of selection normally methodised using a name generator.

Quota Random Sampling - Estimating the ratio and randomly choosing accordingly.

Stratified Random Sampling - Finding the ratio and randomly choosing accordingly (e.g. 2:3 male to female).

Systematic Random Sampling - Selecting on a regular structured basis (e.g. every fifth person).

Oppurtunity Random Sampling - Randomly selecting those closest or most convenient for you to use.

Volunteer Sampling - Placing an advertisement and selecting the first participants you get needed to make up sample (less order effects but more participant effects).

Ethics

THESE ARE THE ETHICAL ISSUES WE ARE CONCERNED WITH IN PSYCHOLOGY:

Consent - whether a participant has given permission to take part in the research.

Deception - whether a participant has been lied to at any point prior or during the research.

Confidentiality - whether the information and identity of participants are being protected.

Debriefing - whether participants are aware of the involvment and tasks within the research prior to giving their consent.

Right to Withdraw - a participant's right to no longer take part in the study at any point in the research.

Protection of Participants - whether the participant's mental and physical health is protected and maintained throughout the research process.

Measures of Central Tendency

THESE ARE THE MEASURES OF CENTRAL TENDENCY:

Mean –  determined by adding all scores together and dividing by the number of scores.

Mode -  uses the most frequently occurring score. A distribution with two or more scores that are equal and occur most frequently is called multi-modal.

Median – uses the middle score in a distribution (the score that occurs at exactly the 50th percentile).

Range – uses the difference between the lowest and the highest score in a distribution.

Standard deviation – a measure of spread within distribution surrounding the mean. 

MEAN

+ uses all data so provides a strong valid estimate

- skewed by extreme values

MEDIAN

+ not affected by extreme values

- not all data is used so can be a misrepresentative value

MODE

+ can be used for categorical data unlike other measures

- there can be multiple modes or no mode (vague)

RANGE

+ easy to analyse

- affected by extreme values

STANDARD DEVIATION

+ precise way to include all data without taking into account extreme values

Methods of Research

THESE ARE THE METHODS OF RESEARCH:

EXPERIMENTS

CASE STUDIES

SURVEYS/QUESTIONNAIRES

INTERVIEWS

OBSERVATIONS

CORRELATIONAL STUDIES

LAB EXPERIMENTS

+ easy to repeat so you can check results

+ able to control confounding variables

- lacks external validity because so much is controlled

- demand characteristics knowing they're being studied

NATURALISTIC EXPERIMENTS

+ can investigate an IV that may be unethical to manipulate

+ demand characteristics lessen because they are unaware that they are being watched

- can't control extraneous variables

- time consuming

CASE STUDIES

+ able to get detailed information

+ lots of internal validity and account for indaviduality

- massively lacks population validity

- time consuming

SURVEYS/QUESTIONNAIRES

+ cheap/convenient

+ people can take their time answering

+ people can answer in their own home

- can misunderstand questions

- people can miss out/incorrectly answer questions

- not everyone will return the form 

Interviews

There are three types of interviews:

Structured - set format of standardised questions

Semi-structured - set questions but these can be varied and expanded upon (normally used by psychologists)

Unstructured - no set format for the questions, allows interviewer to follow the natural progression of the conversation

INTERVIEWS

+ allows interviewer to clarify questions so that the answers are valid

+ allows participants to freeky express

+ can collect qualitative and quantitative data

- interpersonal variables

- time consuming

Observations

There are several different types of interviews:

Covert observation - The participants don't know they are being observed.

Overt observation - The participant knows they are being oberserved. (DEMAND CHARACTERISTICS)

Participant observer - The researcher joins the participant either overtly or covertly.

Non-participant observer - The researcher remains seperate from the participant.

Naturalistic observation - The behaviour being studied  isn't interfered with in any way.

Controlled observation - The behaviour being studied is observed in a specific situation.

Unstructured observation - Everything the participant does is noted down.

Structured observation - There is a standardised behavioural criteria that the observation is based on.

Time sampling - The participant's behaviour is noted at pre-decided times (every minute).

Event sampling - Every time the behaviour being studied occurs it is noted down.

OBSERVATIONS

+ can get lots of detail from observations

+ high internal and external validity often because its in a natural but manipulated environment

- time consuming

- demand characteristics if they know they are being observed

CORRELATIONAL STUDIES

+ allows us to easily establish the relationship between two variables

+ high ecological validity because nothing is manipulated

- can't find the cause of this relationship