EXPERIMENTAL METHOD - Types of experiment, laboratory / field experiments / natural / quasi-experiments.
EXPERIMENTAL DESIGNS - Repeated measures, independent groups, matched pairs.
Laboratory experiments Field experiments
● A controlled environment where EVs and CVs can be ● Experiment conducted in a natural setting
carefully controlled. ● The researcher goes to ppts
● Ppts go to the researcher ● But variables are still well controlled.
● IV is manipulated / the effect on the DV is recorded. ● IV is manipulated and the DV is recorded
✅High degree of control of EV/CVs
● Locations can include shops / schools
● Anywhere the ppt is expected to behave naturally.
● Effect of EVs/CVs can be minimised. ● Eg - Bickman's NYC study
● High internal validity ● UNHE
✅ More naturalistic
● Confidence establishing a cause-and-effect relationship
● All variables outside of the IV and DV are kept consistent
✅
● Observed differences due to the IV manipulation ● Ppts are more likely to act naturally & more authentically
Easy to replicate ● Higher mundane realism
● Researchers can repeat studies to check results ● Results have higher ecological validity
✅
● Reduced chance that new EVs are introduced. ● More generalisable
● Standaridisation Ppts are unaware
● Findings can be confirmed via replication ● More likely to behave normally
❌
● Greater validity ● Greater generalisability
May lack generalisability ● No demand characteristics
❌
● Controlled lab environment may be artificial and overly ● Greater external validity
contrived Harder to control ev/cv's
● Low mundane realism, not emulating real life environments ● Observed changes in the DV may be due to EVs / CVs
❌
● Low external & ecological validity ● Precise replications aren't possible
Demand characteristics ● Harder to draw a cause-and-effect relationship
❌
● Environmental cues invite responses & pptreactivity ● Low control & internal validity
● Aware they are in the study Ethical issues/considerations
● Low internal validity ● Ppts may not have given informed consent
● Causal relationship is hard to confirm ● Privacy may be invaded
● Raising ethical issues.
Natural experiments Quasi experiments
● Experiment where the IV changes ● Experiment where the IV is based on a pre-existing
● BUT it is not manipulated by the researcher. difference which exists and is unmanipulated.
● Varies even if the study didn't happen. ● Ppts CANNOT be randomly assigned between levels of IV,
● The DV may be naturally occurring or measured by the ● Level of IV is usually an innate characteristic.
researcher ● DV is measured, while all other possible variables affecting
● Eg - Reactions to earthquakes / weather and mood the DV are kept constant.
● EORR ● Eg: gender, age, mental health condition, income level,
✅ May be the only practical/ethical option
educational level
✅
● CCRC
● May be unethical to manipulate the IV Highly controlled conditions
● Provides opportunities to do causal research for certain ● Studies are often carried out in controlled conditions
✅ ✅
topics that would be impossible ● High replicability and internal validity
High external validity Comparisons can be drawn between people
● Real-world issues ( natural disasters / institutionalisation) ● Participant variables & differences
❌
● Findings are more relevant to real experiences ● Comparisons can be made between different types of
❌
Rarely occuring events people
● Can't be replicated & natural events may be 'one-offs' No random allocation
● Reduces research opportunities ● Experimenter can't control pre-existing variables
● Limits the scope of generalising the findings ● Ppt variables may cause the change in the DV -> CV
❌ ❌
● Low reliability ● Harder to draw conclusions
Difficult to randomise pp's Hard to demonstrate causal relationships
● Experimenter has no control over conditions ppts are placed ● Researcher doesn't manipulate the IV
into as the IV is pre-existing ● We cannot be confident the change in the DV was due to
● CV/EVs in groups that aren't controlled the IV
Repeated measures Independent groups
● Same ppts take part in all conditions of the IV. ● Ppts only perform in one condition of the IV
, ● SOLUTION: Counterbalancing (ABBA) ● One group does condition A / the other does condition B
✅
● An attempt to control for the effects of order in a repeated ● Ppts should be randomly allocated to reduce order effects
measures design: No order effects
● Half the ppts experience the conditions in one order, and the ● Ppts are only tested once
other half in the opposite order. ● Less chance of boredom, fatigue or practice effects
✅
● Group 1 does condition A, then B ● Controls a CV
✅
● Group 2 does condition B, then A Less demand characteristics
Participant variables ● Less likely to guess the aim as they are only tested once
● As the same person is compared in 2 conditions = ppt ● Behaviour may be more 'natural'
❌
variables are eliminated ● Higher experimental realism (mundane)
✅
● Controls an important CV No control over partcipant variables
Fewer ppts ● Ppts within group A and group B may will act differently
● Half the ppts compared to independent groups ● This acts as an EV/CV
❌ ❌
● Less time consuming and more economical ● Reduces the internal validity
Order effects Less economical
● Ppts may perform better / worse if they do the same task 2x ● Twice as many pp's are required
(practice/fatigue) ● Other experimental designs need less ppt's to gather the
❌
● Reduces the validity of the results same data
Ppts may guess the aims ● Expensive and time- consuming
● Demand characteristics
● Ppts may change their behaviour
● Reduces internal validity
Matched pairs The experimental method
● Pairs of ppts are first matched on participant variable(s) that ● Manipulation of an (IV) to have an effect on the (DV)
may affect the DV. ● Demonstrates a cause-and-effect relationship
● One member of the pair is assigned to Condition A and the ● Clearly operationalised and measurable
✅
other to Condition B.
Participant variables EQ: Explain what is meant by replicability and why is it an
● Pairs are matched on participant variables important feature of science? (5)
● Eliminates participant variables as EV/CV's ● Scientific method involves defining a problem and
formulating a hypothesis which is tested with empirical
✅
● Enhances control and validity
No order effects research.
● Ppts are only tested once ● Research findings are an important part of this process.
● Less chance of boredom, fatigue or practice effects ● If we wish to draw conclusions from research studies, the
procedures and findings should be repeatable.
❌
● Controls a CV
Matching isn't perfect ● Unrepeatable results may imply flaws or lack of control
● Time-consuming and expensive within the method used and are of limited use in theory
● May not be able to control all the variables relevant construction.
❌
● Creates EV/CVs
More ppts
● A large pool of potential ppts is needed
● Data may be hard to obtain
● Expensive to recruit
OBSERVATIONAL TECHNIQUES - Types: naturalistic and controlled observation / covert and overt observation / participant & non-ppt observation.
OBSERVATIONAL DESIGN - behavioural categories; event sampling; time sampling.
Observational techniques Naturalistic observation
● A way of seeing or listening to what people do without ● Observing and recording behaviour in naturally occurring
having to ask them. situations without trying to manipulate and control the
● Often used within an experiment situation.
● A way of assessing the DV. ● In the ‘real world’
✅High external & ecological validity
● Non-experimental
● Researcher watches & records spontaneous/natural
behaviour of ppts without manipulating levels of IV. ● Behaviours are more likely to be spontaneous in one's
✅ Capturing what people do (high external validity)
● People often act differently during self-report methods
natural environment
● High mundane realism
❌
● More generalisable to real life (high ecological validity)
❌ Risk of observer bias
● Useful insight into spontaneous behaviours Low control
● EV's & CV's may be uncontrolled
, ● Researcher interpretations / expectations may impact ● Replication can be impossible
observations ● Low ecological validity if the pp's become aware they are
● More than 1 observer should be used (inter-observer being watched
reliability)
Controlled observation Overt observations
● Watching & recording behaviour within a structured ● Ppts’ behaviour is watched & recorded
environment ● With their knowledge and consent - awareness
✅ More ethically acceptable
● Aspects of the environment are controlled, in an attempt to
standardise the experience.
● Eg: Ainsworth, Zimbardo ● Informed consent is given
✅ Replicability ❌
● Right to withdrawal is given
Demand characteristics
● More easy to repeat standardised procedures ● More likely to be recording unnatural behaviour due to ppt
❌
● Findings can be checked, greater control of ev/cv's awareness of the study
Low external validity ● Reduces the internal validity of findings
● Behaviour may be contrived/unnatural due to the controlled ● Social desirability bias
environment
● Demand characteristics
● Low mundane realism & low ecological validity
Covert observations Participant observations
● Ppts’ behaviour is watched and recorded ● Researcher becomes a member of the group whose
✅
● WITHOUT their knowledge or consent. behaviour he/she is watching and recording.
Reduced demand characteristics ● Joins in with the group being observed, taking part in the
● Natural behaviour is likely to be recorded -> higher internal task or conversations
✅ Greater insight
validity
● Removes ppt reactivity issues
❌
● More likely to show naturalistic behaviours ● Researcher experiences the same situations
Ethically questionable ● Greater validity of the findings
❌
● No informed consent ● Building rapport, trust = more comfortable = disclose more
● Not ethical (deception) Loss of objectivity
● Could be invasion of privacy ● Possibility that the researcher identifies too strongly with
the research group
● ‘Going naive' - biased
Observer bias ● Threatens objectivity & internal validity of the findings
● Expectations that can influence observation and influence
the perception of reality. Non - participant observations
DEALING with observer bias ● Researcher remains outside of the group whose behaviour
● Checking inter observer reliability & using multiple he/she is watching and recording REMAINS SEPARATE
✅ More objective
observers.
● Multiple observing teams and they are using categories
created independently of each other ● Less likely to identify with the ppts
● Applying them on their own behalf ● Objective distance = less subjective
❌
● (Total no. of agreements / total no. of observations ) x100 ● Reduced chance of bias = increased validity of the findings
Data collection devices Loss of insight
● Visual, audio or written data can be obtained ● Researcher may be too far removed and open to
● Binoculars / two-way mirrors / cameras/ paper coding stereotypes
systems / audio recorders / CCTV cameras ● Reduced external validity and insight
● Lack of rapport
Time sampling Behavioural categories / coding systems
● Recording behaviour within a pre-established time frame, at ● Target behaviour is broken up into components
● Observable / identifiable and measurable
✅
regular intervals (e.g: every minute)
❌
Reduces the no. of observations ● Eg - aggression (no of punches and kicks)
● Less time consuming Lack of clarity & ambiguity
● Categories may overlap
❌
● Flexibility, observing unexpected types of behaviour
Unrepresentative ● Poor categories -> hard to distinguish between behaviours
❌
● Data is unrepresentative & important details may be missed & draw findings
● Not an accurate reflection of the entire observation Dustbin categories
● All forms of behaviour should be compartmentalised, not
Event sampling just in one 'dustbin'
● Target behaviour or event is first established ● Unrecorded & missed behaviours
EXPERIMENTAL DESIGNS - Repeated measures, independent groups, matched pairs.
Laboratory experiments Field experiments
● A controlled environment where EVs and CVs can be ● Experiment conducted in a natural setting
carefully controlled. ● The researcher goes to ppts
● Ppts go to the researcher ● But variables are still well controlled.
● IV is manipulated / the effect on the DV is recorded. ● IV is manipulated and the DV is recorded
✅High degree of control of EV/CVs
● Locations can include shops / schools
● Anywhere the ppt is expected to behave naturally.
● Effect of EVs/CVs can be minimised. ● Eg - Bickman's NYC study
● High internal validity ● UNHE
✅ More naturalistic
● Confidence establishing a cause-and-effect relationship
● All variables outside of the IV and DV are kept consistent
✅
● Observed differences due to the IV manipulation ● Ppts are more likely to act naturally & more authentically
Easy to replicate ● Higher mundane realism
● Researchers can repeat studies to check results ● Results have higher ecological validity
✅
● Reduced chance that new EVs are introduced. ● More generalisable
● Standaridisation Ppts are unaware
● Findings can be confirmed via replication ● More likely to behave normally
❌
● Greater validity ● Greater generalisability
May lack generalisability ● No demand characteristics
❌
● Controlled lab environment may be artificial and overly ● Greater external validity
contrived Harder to control ev/cv's
● Low mundane realism, not emulating real life environments ● Observed changes in the DV may be due to EVs / CVs
❌
● Low external & ecological validity ● Precise replications aren't possible
Demand characteristics ● Harder to draw a cause-and-effect relationship
❌
● Environmental cues invite responses & pptreactivity ● Low control & internal validity
● Aware they are in the study Ethical issues/considerations
● Low internal validity ● Ppts may not have given informed consent
● Causal relationship is hard to confirm ● Privacy may be invaded
● Raising ethical issues.
Natural experiments Quasi experiments
● Experiment where the IV changes ● Experiment where the IV is based on a pre-existing
● BUT it is not manipulated by the researcher. difference which exists and is unmanipulated.
● Varies even if the study didn't happen. ● Ppts CANNOT be randomly assigned between levels of IV,
● The DV may be naturally occurring or measured by the ● Level of IV is usually an innate characteristic.
researcher ● DV is measured, while all other possible variables affecting
● Eg - Reactions to earthquakes / weather and mood the DV are kept constant.
● EORR ● Eg: gender, age, mental health condition, income level,
✅ May be the only practical/ethical option
educational level
✅
● CCRC
● May be unethical to manipulate the IV Highly controlled conditions
● Provides opportunities to do causal research for certain ● Studies are often carried out in controlled conditions
✅ ✅
topics that would be impossible ● High replicability and internal validity
High external validity Comparisons can be drawn between people
● Real-world issues ( natural disasters / institutionalisation) ● Participant variables & differences
❌
● Findings are more relevant to real experiences ● Comparisons can be made between different types of
❌
Rarely occuring events people
● Can't be replicated & natural events may be 'one-offs' No random allocation
● Reduces research opportunities ● Experimenter can't control pre-existing variables
● Limits the scope of generalising the findings ● Ppt variables may cause the change in the DV -> CV
❌ ❌
● Low reliability ● Harder to draw conclusions
Difficult to randomise pp's Hard to demonstrate causal relationships
● Experimenter has no control over conditions ppts are placed ● Researcher doesn't manipulate the IV
into as the IV is pre-existing ● We cannot be confident the change in the DV was due to
● CV/EVs in groups that aren't controlled the IV
Repeated measures Independent groups
● Same ppts take part in all conditions of the IV. ● Ppts only perform in one condition of the IV
, ● SOLUTION: Counterbalancing (ABBA) ● One group does condition A / the other does condition B
✅
● An attempt to control for the effects of order in a repeated ● Ppts should be randomly allocated to reduce order effects
measures design: No order effects
● Half the ppts experience the conditions in one order, and the ● Ppts are only tested once
other half in the opposite order. ● Less chance of boredom, fatigue or practice effects
✅
● Group 1 does condition A, then B ● Controls a CV
✅
● Group 2 does condition B, then A Less demand characteristics
Participant variables ● Less likely to guess the aim as they are only tested once
● As the same person is compared in 2 conditions = ppt ● Behaviour may be more 'natural'
❌
variables are eliminated ● Higher experimental realism (mundane)
✅
● Controls an important CV No control over partcipant variables
Fewer ppts ● Ppts within group A and group B may will act differently
● Half the ppts compared to independent groups ● This acts as an EV/CV
❌ ❌
● Less time consuming and more economical ● Reduces the internal validity
Order effects Less economical
● Ppts may perform better / worse if they do the same task 2x ● Twice as many pp's are required
(practice/fatigue) ● Other experimental designs need less ppt's to gather the
❌
● Reduces the validity of the results same data
Ppts may guess the aims ● Expensive and time- consuming
● Demand characteristics
● Ppts may change their behaviour
● Reduces internal validity
Matched pairs The experimental method
● Pairs of ppts are first matched on participant variable(s) that ● Manipulation of an (IV) to have an effect on the (DV)
may affect the DV. ● Demonstrates a cause-and-effect relationship
● One member of the pair is assigned to Condition A and the ● Clearly operationalised and measurable
✅
other to Condition B.
Participant variables EQ: Explain what is meant by replicability and why is it an
● Pairs are matched on participant variables important feature of science? (5)
● Eliminates participant variables as EV/CV's ● Scientific method involves defining a problem and
formulating a hypothesis which is tested with empirical
✅
● Enhances control and validity
No order effects research.
● Ppts are only tested once ● Research findings are an important part of this process.
● Less chance of boredom, fatigue or practice effects ● If we wish to draw conclusions from research studies, the
procedures and findings should be repeatable.
❌
● Controls a CV
Matching isn't perfect ● Unrepeatable results may imply flaws or lack of control
● Time-consuming and expensive within the method used and are of limited use in theory
● May not be able to control all the variables relevant construction.
❌
● Creates EV/CVs
More ppts
● A large pool of potential ppts is needed
● Data may be hard to obtain
● Expensive to recruit
OBSERVATIONAL TECHNIQUES - Types: naturalistic and controlled observation / covert and overt observation / participant & non-ppt observation.
OBSERVATIONAL DESIGN - behavioural categories; event sampling; time sampling.
Observational techniques Naturalistic observation
● A way of seeing or listening to what people do without ● Observing and recording behaviour in naturally occurring
having to ask them. situations without trying to manipulate and control the
● Often used within an experiment situation.
● A way of assessing the DV. ● In the ‘real world’
✅High external & ecological validity
● Non-experimental
● Researcher watches & records spontaneous/natural
behaviour of ppts without manipulating levels of IV. ● Behaviours are more likely to be spontaneous in one's
✅ Capturing what people do (high external validity)
● People often act differently during self-report methods
natural environment
● High mundane realism
❌
● More generalisable to real life (high ecological validity)
❌ Risk of observer bias
● Useful insight into spontaneous behaviours Low control
● EV's & CV's may be uncontrolled
, ● Researcher interpretations / expectations may impact ● Replication can be impossible
observations ● Low ecological validity if the pp's become aware they are
● More than 1 observer should be used (inter-observer being watched
reliability)
Controlled observation Overt observations
● Watching & recording behaviour within a structured ● Ppts’ behaviour is watched & recorded
environment ● With their knowledge and consent - awareness
✅ More ethically acceptable
● Aspects of the environment are controlled, in an attempt to
standardise the experience.
● Eg: Ainsworth, Zimbardo ● Informed consent is given
✅ Replicability ❌
● Right to withdrawal is given
Demand characteristics
● More easy to repeat standardised procedures ● More likely to be recording unnatural behaviour due to ppt
❌
● Findings can be checked, greater control of ev/cv's awareness of the study
Low external validity ● Reduces the internal validity of findings
● Behaviour may be contrived/unnatural due to the controlled ● Social desirability bias
environment
● Demand characteristics
● Low mundane realism & low ecological validity
Covert observations Participant observations
● Ppts’ behaviour is watched and recorded ● Researcher becomes a member of the group whose
✅
● WITHOUT their knowledge or consent. behaviour he/she is watching and recording.
Reduced demand characteristics ● Joins in with the group being observed, taking part in the
● Natural behaviour is likely to be recorded -> higher internal task or conversations
✅ Greater insight
validity
● Removes ppt reactivity issues
❌
● More likely to show naturalistic behaviours ● Researcher experiences the same situations
Ethically questionable ● Greater validity of the findings
❌
● No informed consent ● Building rapport, trust = more comfortable = disclose more
● Not ethical (deception) Loss of objectivity
● Could be invasion of privacy ● Possibility that the researcher identifies too strongly with
the research group
● ‘Going naive' - biased
Observer bias ● Threatens objectivity & internal validity of the findings
● Expectations that can influence observation and influence
the perception of reality. Non - participant observations
DEALING with observer bias ● Researcher remains outside of the group whose behaviour
● Checking inter observer reliability & using multiple he/she is watching and recording REMAINS SEPARATE
✅ More objective
observers.
● Multiple observing teams and they are using categories
created independently of each other ● Less likely to identify with the ppts
● Applying them on their own behalf ● Objective distance = less subjective
❌
● (Total no. of agreements / total no. of observations ) x100 ● Reduced chance of bias = increased validity of the findings
Data collection devices Loss of insight
● Visual, audio or written data can be obtained ● Researcher may be too far removed and open to
● Binoculars / two-way mirrors / cameras/ paper coding stereotypes
systems / audio recorders / CCTV cameras ● Reduced external validity and insight
● Lack of rapport
Time sampling Behavioural categories / coding systems
● Recording behaviour within a pre-established time frame, at ● Target behaviour is broken up into components
● Observable / identifiable and measurable
✅
regular intervals (e.g: every minute)
❌
Reduces the no. of observations ● Eg - aggression (no of punches and kicks)
● Less time consuming Lack of clarity & ambiguity
● Categories may overlap
❌
● Flexibility, observing unexpected types of behaviour
Unrepresentative ● Poor categories -> hard to distinguish between behaviours
❌
● Data is unrepresentative & important details may be missed & draw findings
● Not an accurate reflection of the entire observation Dustbin categories
● All forms of behaviour should be compartmentalised, not
Event sampling just in one 'dustbin'
● Target behaviour or event is first established ● Unrecorded & missed behaviours
