Experiments and Scientific Processes

Lab experiment

IV is manipulated in a controlled environment

• + extraneous variables can be controlled, standardisation makes it easy for researchers to replicate the study and check for consistent results

• - low eco value (findings can’t be generalised to everyday)- unnatural, unrealistic behaviour, experimenter effects, demand characteristics

Field experiment

Natural environment with IV manipulated

• + high eco validity (findings can be generalised to everyday situations), less demand characteristics

• - situational and participant variables, difficult to replicate, ethical principles (lack of consent)

Quasi experiment

IV is naturally occurring but takes place in a lab

• + useful when its unethical to manipulate the IV, increase external validity (existing groups makes it easy for results to be generalised to larger population)

• - non-random selection of participants could introduce bias (difference in characteristics), less control over extraneous variables

Natural experiment 

The effect of a natural occurring event or situation on the DV

• + high eco validity (real world settings reflect actual behaviour), reduced demand characteristics

• - participant and situational variables may bias results (hard to replicate), sample bias (ppts not randomly allocated) 

Independent measures (groups) design

Using different participants for each condition

• + reduced order effects, same material in each condition, reduced demand characteristics

• - participant variables increased (confound results), twice as many participants needed

Repeated measures design

Same participants taking part in each condition

• + participant variables won’t confound results, fewer participants needed

• - results influenced by order effects, demand characteristics increase, can’t use same materials

Matched participants design

Different people matched on certain participant variables (sample has split characteristics)

• + reduces participant variables, less demand characteristics, educed order effects, same material used in each condition

• - difficult to establish matches and time consuming, twice as many participants needed

Confounding variable definition 

When an extraneous variable has not been properly controlled and interferes with the dependent variable 

Independent variable definition 

A variable manipulated by a researcher to bring about change 

Dependent variable definition

What occurs as a result of the manipulation of the IV

Extraneous variables definition 

A variable which can influence the dependent variable- must be controlled by researchers to ensure the validity and replicability of our results

Cause and effect definition 

Manipulating an independent variable (cause) to measure its effect on the dependent variable (the effect)

Operationalised definition

Where a variable is defined by the researcher and an abstract concept is turned into a measurable variable

Aims definition

The intent of the study and the purpose

Hypothesis definition

A precise, testable statement of what researchers expect to happen in a study and why- suggests expected relationship between variables

What are the 3 types of hypothesis

1. Directional (one-tailed) makes a specific prediction about the outcome

2. Non-directional (two tailed) makes a non specific prediction about the outcome

3. Null suggests there is no relationship between variables- used to test ideas and the statement a researcher wants to disprove 

Experiment definition

A structured investigation designed to establish a cause and effect relationship between variables- manipulation of IV to measure the DV- while controlling extraneous variables. A hypothesis is scientifically tested

Sampling definition

Strategies used to select a subset of individuals (a sample) from a larger population to study and draw inferences about the entire population

Opportunity sampling 

Use people who are available at the time

• + quick and convenient, participants give informed consent

• - biased- participants share many of the same variables and characteristics

Random sampling 

Everyone in the target population has an equal change of being picked e.g. using a computer

• + representative reflection of the target population, prevents bias

• - difficult to obtain a truly random sample, group may be too large to give everyone an equal chance

Volunteer (self-selected) sampling

Volunteers have agreed to take part or responded to an advert

• + quick and easy to do, reach large number of participants, ethical (given consent and aware of risks)

• - type of volunteers may not be representative of target population, may have similar characteristics and traits 

Systematic sampling

Selecting the nth person on a list- n chosen by dividing population by sample size

• + unbiased selection achieves representative sample, not biased by preconceptions of ability or behaviour 

• - participants may refuse to take part- cost more time + money, list may not be fully randomised

Stratified sampling

Population divided into subgroups- individuals randomly selected from each stratum- proportionally represent each aspect of the target population 

• + more representative- ensures each sector is reflected and so results can be generalised, bias minimised and enhanced validity 

• - very complex and time consuming, some populations may have too many differences to be organised into subgroups 

Target population definition 

The group that researchers draw the sample from and want to be able to generalise the findings too 

Counterbalancing definition 

Removing order effects in repeated measure design. Two groups of participants would perform each condition, but in reverse order to balance order effects. AB I BA 

Standardisation definition

A control which keeps things the same for all e.g. standardising instructions given to participants or standardising the amount of time participants are given

Participant variables

Personal variables involving participants e.g. age, personality, health

• Screen participants before they participate

• Randomly allocate participants to conditions

• Use the same participants in each condition

• Use matched pairs design 

Situational variables

Aspects of the research situation which confound the DV if not controlled e.g. noise, temperature, interruptions

• Use lab setting

• Constancy- ensure all participants are exposed to it

Demand characteristics

Features of the study encourage participants to behave in a certain way and guess what the aim is

• Deceive participants about the aim- keep them naive (single-blind technique)

• Use field setting so participants are unaware of involvement

• Use distraction tactics 

Investigator effects

Researcher influences participants performance e.g. raises and eyebrow and participants behave accordingly

• Use a third party who is unaware of the aim ‘double blind technique’

• Ensure researcher is unaware which condition participants are in

Order effects 

Performance affected by the order participants perform each condition e.g. 2nd time they are more practiced or relaxed

• Use different participants in each condition

• Use time breaks between condition

• Counterbalance so some do A then B and some do B then A

Social desirability

Participants act in ways which make them seem more desirable- their answers are more socially acceptable then their true opinions or behaviours 

• Create comfortable environment where participants feel relaxed sharing honest opinion

• Ensure anonymity- no fear of judgement

• Choose a self completion mode 

Informed consent and how its dealt with

Participants must be fully informed about the purpose, risks, benefits of research and should voluntarily agree to participate

• Prior general consent (if participant agrees they wouldn’t object to being deceived in future research, it is assumed they have given consent)

• Presumptive consent (if a random sample of population consent this can be generalised)

• Children as participants must have gained consent of parents

Right to withdraw

Participants should have the right to withdraw them and their data at any time without negative consequences

Privacy and confidentiality

Participants should be identified with a number or pseudonym- their identities should not be disclosed and data should not be individually identifiable

Competence 

Psychologists must maintain the highest standards and ensure work is of scientific value

Protection from harm

Participants must not be exposed to physical or psychological harm- distress should be minimised and support provided

Debriefing

Participants should be debriefed after the study about the true nature, clarifying misunderstandings and allowing participants to ask questions- ensures participants leave in the same state they entered

Deception and how it’s dealt with

Prevents participants from giving informed consent and participants may find themselves in research against their wishes- lack of trust for psychologists

• Debriefing on the true aim, return participants to the state they were in when they entered, retrospective informed consent (give right to remove data after true nature revealed)

Internal reliability definition 

Whether the experiment is consistent within itself e.g. did all participants have the same experience 

External reliability definition

Whether the experiment is replicable e.g. if it happens again would it be done in the exact same way

Validity definition

Whether the experiment/correlation/self-report measures what it intends to

Internal validity definition

The extent to which the manipulation of the IV directly affects the DV (no confounding variables)

External validity definition

The extent to which the results generalise beyond the study

Population validity definition

The extent to which the results generalise beyond the immediate sample

Ecological validity definition

The extent to which the results generalise beyond the experiment to a real-life setting

Temporal validity definition

Generalisation across time

Face validity definition

Whether a measure appears to measure what it should

Concurrent validity definition

Measure agrees with an established test

How to increase validity in an experiment

• Control extraneous variables - reduces confounding variables increasing the internal validity

• Standardised procedures - ensures consistency

• Random allocation - reduces participant variables

• Single / double-blind procedures - reduces demand characteristics or researcher bias