Experimental method Control of variables
Experimental method: manipulation of an independent Extraneous variables: any variable other than IV that may
variable to measure the effect on the dependent variable have an effect on the DV if not controlled
Aims: general statement of what the researcher intends - Do not vary systematically to IV
to investigate – purpose of the study. Cofounding variables: variable other than the IV that may
Hypotheses: clear, precise, testable statements have affected the DV – cannot be sure of true source of
predicting/stating the relationships between variables change in DV
being investigated. - Vary systematically to IV
o Must be stated at outset. Demand characteristics: cue from researcher/situation
- Directional: states direction of difference or that may be interpreted as revealing the purpose to ppts
relationship. - Lead to change in behaviour (under-perform/
o One-tailed under-perform) – unnatural behaviour
- Non-directional: does not state the direction - Ppts not passive – participant reactivity
o Two-tailed Investigator effects: effect of investigators behaviour on
Variables: anything that can vary/change within an research outcome
investigation - Unconscious/conscious
- Independent: aspect that is manipulated by the - Can be in design, sample selection, interaction
researcher (can also change naturally) with ppts
o To measure effect on DV - Researcher bias
- Dependent: variable measure by researcher – Randomisation: use of chance in order to control effect of
changes bias when designing materials/deciding order of
o Should be a result of the IV conditions.
Operationalise: clearly defining variables in terms of how Standardisation: using the exact same formulised
they can be measured. procedures/instructions for all ppts in research study
Experimental design Types of experiment
Experimental design: different ways in which the testing Laboratory experiments:
of participants can be organised in relation to - Conducted in highly controlled environments (not
experimental conditions always a lab) – researcher manipulates IV, records
Independent measures: effect on DV.
- Participants allocated to different conditions – - High control over extraneous variables –
take part in one condition (compare conditions) demonstrates high internal validity, can
- Not the same conditions – individual differences determine change in DV is the result of the IV.
- Less economical – each participant represented - Easier replication – ensure findings are valid.
one result – need more participants - Lack of generalisability – often involve artificial
- No order effects – less likely to guess aims, get task/lack real-life experience – low ecological
tired or bored. validity and mundane realism.
Random allocation: remove individual differences in Field experiments:
independent groups by randomly assigning participants to - IV manipulated in more natural/everyday setting
different conditions - Higher mundane realism – natural environment
Repeated measures: - Produce authentic/more valid behaviour
- All participants take part in all experimental (especially if covert)
conditions (compare sets of data – like with like) - Loss of control – cause and effect more difficult to
- Order effects – likely to get bored, tired or wise establish (low internal validity – replication is
- Bigger chance of demand characteristics hard)
- Individual differences are controlled (none) Natural experiment:
- Fewer participants needed = cost effective - Researcher takes advantage of a pre-existing IV
Counterbalancing: control order effects in repeated (change in IV would happen researcher) – records
measures – half take part in one order, half the other. effect on DV.
Matched pairs: - Provides opportunities for research that would
- Participants are matched on some variable(s) that otherwise be impractical/unethical
may affect the DV – one goes to A the other to B - High external validity – involve study of real-life
, - Requires a pre-test = costly and time consuming issues/problems/events
- Participants cannot be matched exactly (even - May only happen rarely – limits generalisability
identical twins - No random allocation = socially sensitive.
- Order effects and demand characteristics less of a Quasi-experiments:
problem - IV based on pre-existing differences between
people (IV simply exists) – e.g. age, gender,
anxiety
- No manipulation
- Controlled conditions = high internal validity
- Cofounding variables from not randomly
allocating
Sampling Sampling
Population: group of people who are the focus of the Stratified sample:
researcher’s interest – target population - Composition of sample represents/reflects the
Sample: group of people who take part in research proportions of people in subgroups in target pop.
investigation – drawn from target population - Identify strata, work our proportions and select
Sampling technique: methods used to select people from using random sampling.
the target population - Avoids researcher bias
Random sample: - Representative – accurate reflection – generalise
- Sophisticated – all members of target population - Not perfect – cannot reflect all differences
have equal chance of being picked. Opportunity sample:
- Compile list, assign numbers and randomly - Researcher takes a chance to ask whoever is
generate sample around at the time of their study
- Free from researcher bias – no influence over who - Convenient – save time and effort
is selected - Unrepresentative of target population – drawn
- Time consuming/difficult to obtain – some from one area/likely psychology students
participants may refuse (end up being volunteer) - Researcher bias – may avoid people
- May end up unrepresentative. Volunteer bias:
Systematic sample: - Participants self-selected themselves to be the
- Every nth number of target population is selected sample
- List of target pop, alphabetise, nominate system, - Place an advert/ask people to raise hands
work through list until sample is complete. - Easy – minimal effort required
- Free from researcher bias – beyond influence - Volunteer bias – attract certain profile of person
- Usually representative (is possible but unlucky to (helpful, keen and curious) – affects
be unrepresentative). generalisability
Ethical issues Ethical issues
Ethical issues: arise when conflict arises between rights of Deception:
ppts and goals of research to produce authentic data. - Deliberately misleading/withholding information
Informed consent: from ppts – not revealing adequate information.
- Making ppts aware of the aims, procedures, rights - Overcome by giving full debrief at end of study
and what data will be used for. and provide chance to withdraw data.
- Informed judgement to participate – sign consent Confidentiality:
- Can produce unnatural behaviour as aims are - Ppts have the right over information about
known themselves – right to privacy
- Overcome by issuing letter or consent detailing - Confidential information should be protected
relevant information (under 16 = parental (Data Protection Act)
consent) - Use the act to protect ppts, location of study and
- Presumptive consent: ask similar group if study is institutions
acceptable – consent of original ppts presumed. - Overcome by using numbers/initials instead of
- Prior general consent: ppts give permission to names – remind them of their data being
various studies – one involving deception. protected
Experimental method: manipulation of an independent Extraneous variables: any variable other than IV that may
variable to measure the effect on the dependent variable have an effect on the DV if not controlled
Aims: general statement of what the researcher intends - Do not vary systematically to IV
to investigate – purpose of the study. Cofounding variables: variable other than the IV that may
Hypotheses: clear, precise, testable statements have affected the DV – cannot be sure of true source of
predicting/stating the relationships between variables change in DV
being investigated. - Vary systematically to IV
o Must be stated at outset. Demand characteristics: cue from researcher/situation
- Directional: states direction of difference or that may be interpreted as revealing the purpose to ppts
relationship. - Lead to change in behaviour (under-perform/
o One-tailed under-perform) – unnatural behaviour
- Non-directional: does not state the direction - Ppts not passive – participant reactivity
o Two-tailed Investigator effects: effect of investigators behaviour on
Variables: anything that can vary/change within an research outcome
investigation - Unconscious/conscious
- Independent: aspect that is manipulated by the - Can be in design, sample selection, interaction
researcher (can also change naturally) with ppts
o To measure effect on DV - Researcher bias
- Dependent: variable measure by researcher – Randomisation: use of chance in order to control effect of
changes bias when designing materials/deciding order of
o Should be a result of the IV conditions.
Operationalise: clearly defining variables in terms of how Standardisation: using the exact same formulised
they can be measured. procedures/instructions for all ppts in research study
Experimental design Types of experiment
Experimental design: different ways in which the testing Laboratory experiments:
of participants can be organised in relation to - Conducted in highly controlled environments (not
experimental conditions always a lab) – researcher manipulates IV, records
Independent measures: effect on DV.
- Participants allocated to different conditions – - High control over extraneous variables –
take part in one condition (compare conditions) demonstrates high internal validity, can
- Not the same conditions – individual differences determine change in DV is the result of the IV.
- Less economical – each participant represented - Easier replication – ensure findings are valid.
one result – need more participants - Lack of generalisability – often involve artificial
- No order effects – less likely to guess aims, get task/lack real-life experience – low ecological
tired or bored. validity and mundane realism.
Random allocation: remove individual differences in Field experiments:
independent groups by randomly assigning participants to - IV manipulated in more natural/everyday setting
different conditions - Higher mundane realism – natural environment
Repeated measures: - Produce authentic/more valid behaviour
- All participants take part in all experimental (especially if covert)
conditions (compare sets of data – like with like) - Loss of control – cause and effect more difficult to
- Order effects – likely to get bored, tired or wise establish (low internal validity – replication is
- Bigger chance of demand characteristics hard)
- Individual differences are controlled (none) Natural experiment:
- Fewer participants needed = cost effective - Researcher takes advantage of a pre-existing IV
Counterbalancing: control order effects in repeated (change in IV would happen researcher) – records
measures – half take part in one order, half the other. effect on DV.
Matched pairs: - Provides opportunities for research that would
- Participants are matched on some variable(s) that otherwise be impractical/unethical
may affect the DV – one goes to A the other to B - High external validity – involve study of real-life
, - Requires a pre-test = costly and time consuming issues/problems/events
- Participants cannot be matched exactly (even - May only happen rarely – limits generalisability
identical twins - No random allocation = socially sensitive.
- Order effects and demand characteristics less of a Quasi-experiments:
problem - IV based on pre-existing differences between
people (IV simply exists) – e.g. age, gender,
anxiety
- No manipulation
- Controlled conditions = high internal validity
- Cofounding variables from not randomly
allocating
Sampling Sampling
Population: group of people who are the focus of the Stratified sample:
researcher’s interest – target population - Composition of sample represents/reflects the
Sample: group of people who take part in research proportions of people in subgroups in target pop.
investigation – drawn from target population - Identify strata, work our proportions and select
Sampling technique: methods used to select people from using random sampling.
the target population - Avoids researcher bias
Random sample: - Representative – accurate reflection – generalise
- Sophisticated – all members of target population - Not perfect – cannot reflect all differences
have equal chance of being picked. Opportunity sample:
- Compile list, assign numbers and randomly - Researcher takes a chance to ask whoever is
generate sample around at the time of their study
- Free from researcher bias – no influence over who - Convenient – save time and effort
is selected - Unrepresentative of target population – drawn
- Time consuming/difficult to obtain – some from one area/likely psychology students
participants may refuse (end up being volunteer) - Researcher bias – may avoid people
- May end up unrepresentative. Volunteer bias:
Systematic sample: - Participants self-selected themselves to be the
- Every nth number of target population is selected sample
- List of target pop, alphabetise, nominate system, - Place an advert/ask people to raise hands
work through list until sample is complete. - Easy – minimal effort required
- Free from researcher bias – beyond influence - Volunteer bias – attract certain profile of person
- Usually representative (is possible but unlucky to (helpful, keen and curious) – affects
be unrepresentative). generalisability
Ethical issues Ethical issues
Ethical issues: arise when conflict arises between rights of Deception:
ppts and goals of research to produce authentic data. - Deliberately misleading/withholding information
Informed consent: from ppts – not revealing adequate information.
- Making ppts aware of the aims, procedures, rights - Overcome by giving full debrief at end of study
and what data will be used for. and provide chance to withdraw data.
- Informed judgement to participate – sign consent Confidentiality:
- Can produce unnatural behaviour as aims are - Ppts have the right over information about
known themselves – right to privacy
- Overcome by issuing letter or consent detailing - Confidential information should be protected
relevant information (under 16 = parental (Data Protection Act)
consent) - Use the act to protect ppts, location of study and
- Presumptive consent: ask similar group if study is institutions
acceptable – consent of original ppts presumed. - Overcome by using numbers/initials instead of
- Prior general consent: ppts give permission to names – remind them of their data being
various studies – one involving deception. protected
