Research Methods in Social Relations ch. 5 – G. Maruyama &
C.S. Ryan (2014)
Research Question: When are laboratory settings appropriate for research, and how do different
types of laboratory studies affect the validity of research conclusions?
Research Challenges: Some questions are difficult to investigate in labs, and participants may be
aware they're part of a study, impacting validity.
Laboratory Study Types:
Impact Study: Tests direct effects on behaviour.
Judgment Study: Participants respond to stimuli.
Observational Study: Allows controlled observation.
Validity in Labs:
Internal Validity: High due to control and random assignment.
Construct Validity: Strong if variables reflect real-life concepts. It is indicated by
experimental realism, the extent to which the manipulations are perceived by participants in
the intended way. Not so much by mundane realism, the degree to which experimental
settings resemble real-life situations. Construct validity can be threatened by artifacts
distorting results:
o Experimenter Expectations: Unintentional bias from researchers, influencing
participants behaviour.
o Demand Characteristics: Participants altering behaviour due to perceived study
goals.
External Validity: Generally low due to artificial settings. Universalistic research looks for
broad truths that apply to many situations, like testing if a theory is true in general.
Particularistic research looks at specific cases or settings, like studying a particular company's
situation. Important there is similarity to real-life settings.
Key Elements:
Setting: Use cover stories to guide participants and minimize demand characteristics.
Independent Variable: Ensure manipulation is believable and clear.
Manipulation Checks: Verify the manipulation's effect but be cautious with excluding
participants based on manipulation checks as they can lower sample size, affect validity, and
introduce bias if not done transparently..
Dependent Variables: Use 2-3 reliable and valid measures to capture different aspects. Avoid
too many measures to prevent results being significant by chance (capitalising on chance).
Self-report measures need careful wording. Behavioroid measures (promises of future
actions) are less reliable but easier to use. Performance measures assess how well
participants complete tasks. Ensure all measures are reliable, valid, and minimize demand
characteristics.
Debriefing: Researchers can identify inattentive or suspicious participants, check procedure
clarity, and ensure participants leave with an understanding of deception and positive
feelings.
C.S. Ryan (2014)
Research Question: When are laboratory settings appropriate for research, and how do different
types of laboratory studies affect the validity of research conclusions?
Research Challenges: Some questions are difficult to investigate in labs, and participants may be
aware they're part of a study, impacting validity.
Laboratory Study Types:
Impact Study: Tests direct effects on behaviour.
Judgment Study: Participants respond to stimuli.
Observational Study: Allows controlled observation.
Validity in Labs:
Internal Validity: High due to control and random assignment.
Construct Validity: Strong if variables reflect real-life concepts. It is indicated by
experimental realism, the extent to which the manipulations are perceived by participants in
the intended way. Not so much by mundane realism, the degree to which experimental
settings resemble real-life situations. Construct validity can be threatened by artifacts
distorting results:
o Experimenter Expectations: Unintentional bias from researchers, influencing
participants behaviour.
o Demand Characteristics: Participants altering behaviour due to perceived study
goals.
External Validity: Generally low due to artificial settings. Universalistic research looks for
broad truths that apply to many situations, like testing if a theory is true in general.
Particularistic research looks at specific cases or settings, like studying a particular company's
situation. Important there is similarity to real-life settings.
Key Elements:
Setting: Use cover stories to guide participants and minimize demand characteristics.
Independent Variable: Ensure manipulation is believable and clear.
Manipulation Checks: Verify the manipulation's effect but be cautious with excluding
participants based on manipulation checks as they can lower sample size, affect validity, and
introduce bias if not done transparently..
Dependent Variables: Use 2-3 reliable and valid measures to capture different aspects. Avoid
too many measures to prevent results being significant by chance (capitalising on chance).
Self-report measures need careful wording. Behavioroid measures (promises of future
actions) are less reliable but easier to use. Performance measures assess how well
participants complete tasks. Ensure all measures are reliable, valid, and minimize demand
characteristics.
Debriefing: Researchers can identify inattentive or suspicious participants, check procedure
clarity, and ensure participants leave with an understanding of deception and positive
feelings.
