Psychology- Research Methods

pop psychology

simple and appealing explanations that are not backed up by empirical evidence.

Empirical evidence is based on

observation

Falsification/falsifiability

It is the principle that a proposition or theory can only be considered scientific

Replication

Is the process of repeating research to determine the extent to which findings can be generalised across time and across situations.

Quantative Methods

data to characterise behavior is numerical

Qualatative methods

data to characterise behaviour is written

nomothetic approach

a simple set of rules that characterise behaviour of all material objects

Quantatative Research

works with variables

Variables

a quantity or quality that varies across individuals or situations

operationalised variables

turning abstract concepts measurable

constructs

any theoretical defined variable given a definition to delineate it from similar constructs

Hypothesis

statement describing relationship between variables

directional hypothesis

states expect difference to be found

Non-directional hypothesis

does not state what the expected difference is

empirical evidence

has been gathered using scientific methods including the use of observation and experimentation.

anecdotal evidence

has been observed from someone's personal experience

independant variable

manipulated by the experimenter

dependant variable

expected to change

correlation

change not based on but related to something else

extraneous variables

any variable not invesigated

co-variables

correlating variables that could affect results

idiographic approach

in depth approach not seeking to apply universal rules

confounding variables

variables that could distort iv and dv

A sample

A group of individuals taking part in a research study.

Participant

An individual taking part in research.

Sampling

The process of finding and recruiting participants.Different techniques have their own strengths and limitations.

Credibility

The degree to which the results of the research can be trusted.

Generalisability

Extent to which the results/findings of research can be applied outside of the sample.

Bias

If some sort of bias influences the findings it undermines the credibility.

Volunteer sampling

a sample made of only people who choose to partake

Oppertunity sampling

A sample of only people who are readily avalible

Random Sampling

Everyone available- usually a large amount

Stratified sampling

a sample which is directly proportional to the population

Quota Sampling

Researcher decides how many people should participate and what characteristics they should have before the study begins

Purpose Sampling

Main characteristics are decided but sample size and proportions of characteristics are not

Theoretical Sampling

Stops when data saturation is reaches so new participants don't change the data

Snowball Sampling

Participants recruit other participants- usually people in a similar situation

Convenience Sampling

using the easiest available people e.g. uni students
not generalisable!

Experimental Study

At least one IV (manipulated) and DV (expected to change)

Correlation Study

No variables are manipulated

Descriptive Studies

Relationships between variables are approached seperatly and the distribution of answers is noted.

Experimental Designs

the organisation of how one variable is measured and another is manipulated and how participants are organised

Conditions

independent measures design

groups are randomly allocated and compared. the only difference in the groups is how the IV is manipulated

matched pairs design

groups are matched based on the matching variable

repeated measures design

groups are exposed to multiple conditions and compared to themselves

counterbalancing

participants use another group of participants with the conditions in the reverse order

participant variability

extent to which participants are different

matching variable

variable used to organise groups

order effects

data is changed due to practise or fatigue

Construct Validity

Focused on the quality of operationalised constructs

Internal Validity

characterises the quality of the method

External Validity

refers to extend of generalisability

Population Validity

refers to extend to generalise to general population

Ecological Validity

refers to extend to generalise to different settings

Selection

Occurs if for some reason the groups are not equivalent at the start of the experiment. Apart from the planned IV difference, they differ on some other variable.

History

Refers to outside events that happen to participants in the course of the experiment.

Maturation

During the course of the experiment the participants go through natural developmental processes, such as fatigue or simply growth.

Testing effect

The first measurement of the DV may affect the second and any further measurements.
Instrumentation

Regression to the mean

This is a concern when the initial score on the DV is extreme (high or low) and become more average

Experimental mortality

This refers to the fact that some participants drop out during an experiment, which could be a problem if the dropouts are not random.

Demand characteristics

This is where the participant understands the purpose of the experiment and subconsciously changes their behaviour to fit that interpretation. They behave in a way they think the experiment expects.

Experimenter bias

This refers to when the researcher unintentionally exerts an influence on the results of the study

Quasi Experiment

groups are not randomly allocated but there is a control group

Field Experiments

conducted in real life setting so have high ecological validity but less control over confounding variables

Natural Experiments

conducted in real life settings but researcher does not manipulate IV.

Scatterplot

graphical illustration of data where 1 dot= 1 participant

Negative Correlation

right-left. increase in one but decrease of another

Positive correlation

left-right. increase in both

Zero correlation

scores are not related and no line of best fit can be drawn

Line of best fit

demonstrates a correlation or causation on a graph

Causation

when one variable directly affects another

Statistical Significance

whether or not the difference between groups can be attributed to chance or if the difference is likely the result of experimental influences

Null Hypothesis

no relationship between the two variables

Effect size

absolute value of correlation coefficient from 1 to -1

Method triangulation

aim to get the same results with a different method

Data Triangulation

aim to get the same results with data from different sources

Researcher Triangulation

using other researchers to carry out and review data

Theory Triangulation

using multiple perspectives or theories

Acquiescence bias

tendency to give positive answer

Social desirability bias

tendency to give answers that make the participant seem more likeable

Dominant Respondant

when one participant influences others or leads the group

Sensitivity

distorting answers to questions to hide secrets

Confirmation Bias

researcher has a belief that affects the participant

Leading Question Bias

the researchers question encourages a certain answer

Question order bias

when the answer to one question influences the answers to the others

Sample Bias

when the sample is not suitable

Biased reporting

findings are not represented equally in the report

Sample to population generalisation

sample is representative of the population (usually random sampling) but can be hard to generalise if its qualitative

Theoretical generalisation

Quant: construct Validity, Qual: achieved through rigorous analysis

Case to Case Generalisation

Quant: ecological valididty, Qual:responsibility of reader and researcher

Observations

focused on how people interact in a non-artifical environment so the researcher can become immersed. Can be: overt,covert, structured, unstructure, laboratory, naturalistic

The interview

in depth 1-1 interviews to gain insight to a subjective experience. can be structured, semi structured or unstructured.

Case Study

in depth investigation of a person or group that can involve other methods. NOT SAMPLING, instead only a certain group. Group can be hard to access and bias can be an issue

Focus Group

6-10 participants all are asked questions at the same time. This can be good to get lots of data quickly but can be limited by dominant respondents and is hard to provide anonymity or confidentiality

What are the ethical concerns when designing research?

- informed consent
- protection from harm
- anonymity and confidentiality
- withdrawal from participation
- deception
- debriefing
- cost benefit analysis

informed consent

A participant's agreement to take part in a study after being told what to expect ( possible risks and consequences )

Protection from harm

the right of research participants to be protected from physical or psychological harm

anonymity&confidentiality

both protect privacy

anonymity

no data that allows researchers to match the data with the person