1
May-22
6. Research methods
Experimental method:
Aim: general statement that describes the purpose of an investigation
Derived from theories.
Hypothesis: statement made at start of a study stating relationship between variables dictated by theory
Directional (H1) = states the sort of difference anticipated between conditions.
o Used when there is previous research to indicate a direction.
Non-directional (H1) = states that there is a difference (but not what it is) between conditions.
o Used without previous research or if research is contradictory.
Null hypothesis (H0) = there is no difference.
Variables:
Independent variable = researcher changes and manipulates to see effect on DV.
o Levels of IV: control condition, experimental condition. Both should be controlled to
create a clear and testable hypothesis (must be measurable).
Dependent variable = researcher measures effect of IV on DV.
All other variables that are not the IV and DV should be controlled.
Variables and control:
Variables: extraneous, confounding
Extraneous (‘nuisance’) variables = do not vary systematically with IV but may affect results.
Confounding variables = do vary systematically with IV – could explain changes in DV thought to
be a result of manipulation of IV.
Other influences on results: demand characteristics, investigator effects
Demand characteristics = clues in experimental situation as to its aim – may cause changes in
participants’ behaviour (they may do what they think is ‘expected’ of them.
Investigator effects = the unwanted influence of a researcher’s behaviour on the results.
o E.g., expectancy effects, unconscious cues, leading questions, participant selection etc.
Control of variables: randomisation, standardisation
Randomisation = use of chance to control for effects of bias when designing material, choosing
order of conditions etc.
Standardisation = all participants subject to the same environment, information, and experience.
o E.g., standardised instructions in verbatim format.
o Therefore, non-standardised changes do not act as extraneous variables, affecting
results.
Experimental design:
Experimental design: independent groups, repeated measures, matched pairs
Independent groups = two separate groups of ppts experience two conditions. Their
performances are compared.
Repeated measures = all ppts experience both conditions one after the other and the two
performances are compared.
Matched pairs = ppts paired based on variables relevant to the study. May include a pre-test to
determine pairs.
,2
May-22
EVALUATION:
Independent groups:
+ No order effects.
o Therefore, ppts less likely to guess the aim of the study, affecting behaviour (demand
characteristics).
- Participant variables.
o Changes to DV may be due to differences between ppts, rather than IV manipulation.
o Therefore, cannot determine cause and effect. Random allocation needs to be used to attempt
to control for this.
- Less economical.
o 2x the amount of ppts needed compared to repeated measures – each ppt only contributes one
set of data.
o Therefore, less cost effective than repeated measures.
Repeated measures:
+ Participant variables controlled.
o Can be sure that changes to DV are not a result of differences between participants.
o Therefore, easier to determine cause and effect.
+ More economical.
o Each ppt contributes two sets of data which can be directly compared with one another.
o Therefore, more cost-effective.
- Order effects.
o May create boredom/fatigue which may affect results. May also guess the aim of the study,
affecting behaviour and subsequently the results (demand characteristics).
o Therefore, order acts as a confounding variable – cannot determine cause and effect.
Counterbalancing needs to be used to attempt to control this (half of ppts do condition A first,
half do condition B first).
Matched pairs:
+ Order effects and demand characteristics more controlled.
o Two sets of ppts, therefore no effect of order and less likely to guess the aim of the study and
exhibit demand characteristics.
o Therefore, easier to determine cause and effect.
HOWEVER: - Ppts cannot be matched exactly.
o Results may still be affected by ppt variables, despite being matched.
o Therefore, remains difficult to establish cause and effect.
- Not economical.
o Each ppt only produces one set of data. Also, time-consuming, and expensive to do pre-test to
match the pairs.
o Therefore, not cost-effective compared to repeated measures.
, 3
May-22
Types of experiment:
Laboratory experiment: IV manipulated in a highly controlled environment
+ Control of extraneous variables.
o Can be sure that the change in DV is due to the manipulation of IV.
o Therefore, can determine cause and effect; high internal validity.
+ Replicable.
o High level of control = less chance of new extraneous variables when repeated.
o Therefore, easy to replicate in order to see if findings are valid.
- Low generalisability.
o Artificial environment and task, meaning ppts may behave according to demand
characteristics because they are aware of testing.
o Therefore, low external validity because findings cannot be used to explain everyday
behaviour.
Field experiment: IV manipulated in a natural/everyday setting
+ Higher mundane realism due to natural environment.
o Behaviour is likely to be more authentic because ppts may not know that they are being
studied.
o Therefore, high external validity – can explain everyday behaviour.
- Lack of control over extraneous variables.
o Cannot establish cause and effect as easily as in a lab experiment. Also, replication is
more difficult.
o Therefore, harder to establish internal validity due to extraneous influences.
- Ethical issues.
o E.g., if ppts do not know that they are being studied, they cannot consent.
o Therefore, may be difficult to carry out field experiments without it being a breach of
privacy.
Natural experiment: IV is pre-existing and is taken advantage of by the researcher – it would have
changed even if they were not present.
+ Opportunities for research that may have been impossible.
o E.g., for ethical reasons (such as investigation of the effect of a natural disaster on stress
levels).
o Therefore, high external validity because real-life issues are being studied.
- Rare events.
o Limits opportunity for further research and generalisation of results.
o Therefore, low generalisability and usefulness.
- Ppts not randomly allocated.
o May be confounding variables affecting results.
o Therefore, cannot establish cause and effect due to low internal validity.
May-22
6. Research methods
Experimental method:
Aim: general statement that describes the purpose of an investigation
Derived from theories.
Hypothesis: statement made at start of a study stating relationship between variables dictated by theory
Directional (H1) = states the sort of difference anticipated between conditions.
o Used when there is previous research to indicate a direction.
Non-directional (H1) = states that there is a difference (but not what it is) between conditions.
o Used without previous research or if research is contradictory.
Null hypothesis (H0) = there is no difference.
Variables:
Independent variable = researcher changes and manipulates to see effect on DV.
o Levels of IV: control condition, experimental condition. Both should be controlled to
create a clear and testable hypothesis (must be measurable).
Dependent variable = researcher measures effect of IV on DV.
All other variables that are not the IV and DV should be controlled.
Variables and control:
Variables: extraneous, confounding
Extraneous (‘nuisance’) variables = do not vary systematically with IV but may affect results.
Confounding variables = do vary systematically with IV – could explain changes in DV thought to
be a result of manipulation of IV.
Other influences on results: demand characteristics, investigator effects
Demand characteristics = clues in experimental situation as to its aim – may cause changes in
participants’ behaviour (they may do what they think is ‘expected’ of them.
Investigator effects = the unwanted influence of a researcher’s behaviour on the results.
o E.g., expectancy effects, unconscious cues, leading questions, participant selection etc.
Control of variables: randomisation, standardisation
Randomisation = use of chance to control for effects of bias when designing material, choosing
order of conditions etc.
Standardisation = all participants subject to the same environment, information, and experience.
o E.g., standardised instructions in verbatim format.
o Therefore, non-standardised changes do not act as extraneous variables, affecting
results.
Experimental design:
Experimental design: independent groups, repeated measures, matched pairs
Independent groups = two separate groups of ppts experience two conditions. Their
performances are compared.
Repeated measures = all ppts experience both conditions one after the other and the two
performances are compared.
Matched pairs = ppts paired based on variables relevant to the study. May include a pre-test to
determine pairs.
,2
May-22
EVALUATION:
Independent groups:
+ No order effects.
o Therefore, ppts less likely to guess the aim of the study, affecting behaviour (demand
characteristics).
- Participant variables.
o Changes to DV may be due to differences between ppts, rather than IV manipulation.
o Therefore, cannot determine cause and effect. Random allocation needs to be used to attempt
to control for this.
- Less economical.
o 2x the amount of ppts needed compared to repeated measures – each ppt only contributes one
set of data.
o Therefore, less cost effective than repeated measures.
Repeated measures:
+ Participant variables controlled.
o Can be sure that changes to DV are not a result of differences between participants.
o Therefore, easier to determine cause and effect.
+ More economical.
o Each ppt contributes two sets of data which can be directly compared with one another.
o Therefore, more cost-effective.
- Order effects.
o May create boredom/fatigue which may affect results. May also guess the aim of the study,
affecting behaviour and subsequently the results (demand characteristics).
o Therefore, order acts as a confounding variable – cannot determine cause and effect.
Counterbalancing needs to be used to attempt to control this (half of ppts do condition A first,
half do condition B first).
Matched pairs:
+ Order effects and demand characteristics more controlled.
o Two sets of ppts, therefore no effect of order and less likely to guess the aim of the study and
exhibit demand characteristics.
o Therefore, easier to determine cause and effect.
HOWEVER: - Ppts cannot be matched exactly.
o Results may still be affected by ppt variables, despite being matched.
o Therefore, remains difficult to establish cause and effect.
- Not economical.
o Each ppt only produces one set of data. Also, time-consuming, and expensive to do pre-test to
match the pairs.
o Therefore, not cost-effective compared to repeated measures.
, 3
May-22
Types of experiment:
Laboratory experiment: IV manipulated in a highly controlled environment
+ Control of extraneous variables.
o Can be sure that the change in DV is due to the manipulation of IV.
o Therefore, can determine cause and effect; high internal validity.
+ Replicable.
o High level of control = less chance of new extraneous variables when repeated.
o Therefore, easy to replicate in order to see if findings are valid.
- Low generalisability.
o Artificial environment and task, meaning ppts may behave according to demand
characteristics because they are aware of testing.
o Therefore, low external validity because findings cannot be used to explain everyday
behaviour.
Field experiment: IV manipulated in a natural/everyday setting
+ Higher mundane realism due to natural environment.
o Behaviour is likely to be more authentic because ppts may not know that they are being
studied.
o Therefore, high external validity – can explain everyday behaviour.
- Lack of control over extraneous variables.
o Cannot establish cause and effect as easily as in a lab experiment. Also, replication is
more difficult.
o Therefore, harder to establish internal validity due to extraneous influences.
- Ethical issues.
o E.g., if ppts do not know that they are being studied, they cannot consent.
o Therefore, may be difficult to carry out field experiments without it being a breach of
privacy.
Natural experiment: IV is pre-existing and is taken advantage of by the researcher – it would have
changed even if they were not present.
+ Opportunities for research that may have been impossible.
o E.g., for ethical reasons (such as investigation of the effect of a natural disaster on stress
levels).
o Therefore, high external validity because real-life issues are being studied.
- Rare events.
o Limits opportunity for further research and generalisation of results.
o Therefore, low generalisability and usefulness.
- Ppts not randomly allocated.
o May be confounding variables affecting results.
o Therefore, cannot establish cause and effect due to low internal validity.
