week 2 Doing Social Psychology Flashcards
Hypothesis
A well-defined, testable prediction.
Theory
An organised set of principles used to explain observed phenomena.
Basic research
has goal of increasing understanding, often by testing a hypothesis, which is based on a theory.
Applied research
has the goal of increasing understanding AND to find practical solutions to problems
Conceptual variable
when first trying to develop an hypothesis,the variables are vague and abstract, and “just concepts”.
Operational definition
to be tested, a conceptual variable must be precisely defined in terms of how it shall be manipulated or measured, and how it is to be defined.
Construct validity
degree to which the measures or manipulations utilised, do in fact measure or manipulate what they are supposed to.
Rosenberg Self Esteem Scale
a particular self-report scale of self esteem. Considered to have good construct validity.
Bogus Pipeline Technique
Where participants are told their lies will be infallibly detected by a machine (this is false) and it causes the participants to give answers they may have otherwise preferred not to
self report questionaire
may be inaccurate due to inacuuracy of recall, bias of recall, trying to present as socially acceptable, may get different responses if wording or order of questions is altered, etc etc,
interval-contingent self report
respondent reports on experience regularly, eg. daily. In an effort to reduce time lapsed since remembered event , so hopefully report is more accurate
signal-contingent self report
respondent responds when signalled eg via sms
event-contingent self report
respondent reports asap after a pre-determined event, has occurred.
inter-rater reliability
degree to which multiple observers rate an event the same
eye-tracking technology
allows tracking of where in a scene participant looks and for how long
descriptive research
records how frequently or typically people think or behave. This may be achieved via observation, studying archives or conducting surveys.
random sampling
method of selecting participants for a survey where everyone in population has equal chance of selection
qualitative research
records data in non numerical form. more concerned with an individual’s experience. may use non structured interviews etc
quantitative research
records data in numerical form. concerned with quantifying a problem.
correlational research
measures the relationship between variables. Usually this type of research only measures variables, and does not manipulate them. an eg was a finding that regions where tweets used angrier language, tended to also have higher incidence of heart disease mortality.
Advantage of correlational research is that can explore many relationships of variables. disadvantage is that correlation DOES NOT equal causation.
correlational coefficient
degree to which 2 variables are related. range from -1 to 1. Zero= no correlation. Positive= both variables increase together. negative = 1 variable goes up when other goes down.
experimental research
can demonstrate causal relationship. experimenter has control of the experimental procedure such that only the variable is manipulated and all else is uniform. Participants are randomly assigned.
Random assignment
assigning participants already selected for trial, to the various experimental conditions, such that each participant has equal chance of assignment to any group. Thus unlikely that conditions of groups vary due to participant differences.
experiment on norm-deviation 2009 Ashton James and assoc’s
Shown group of pens, mainly black, some blue. Those from more individualistic culture more likely choose blue, and vice versa for collectivist culture. But hypothesised if in good mood, more likely to buck one’s culture trend
which was demonstrated after inducing “good” or “bad” mood with 2 diff classical music pieces.
independent variable
controlled by experimenter
dependent variable
experimenter measures to see if has been affected by the manipulations of the independent variable.
subject variable
neither truly independent nor truly dependent. Variables which are of the individual. not able to be manipulated, nor are they influenced by the independent variable. In above experiment on norm deviation, a subject variable was whether from individualistic or collectivist culture.
internal validity
how much reasonable certainty there is that changes in the independent variables of an experiment, caused the changes observed in the dependent variables.
confound
a factor other that the independent variables, which differs between experimental conditions, such that it threatens internal validity
experimenter expectancy effect
experimenter’s expectations affect how behaves towards participant (eg if experimenter knows if participant has placebo or treatment) and can affect experiment results.
external validity
extent to which can reasonably expect results of an experiment to be replicated in other world situations. ie findings can be generalised to a population
Mechanical Turk
an Amazon online research tool allows for wider ranges of sampling of individuals.
Mundane realism
degree to which experiment reflects real life situation
experimental realism
degree to which experiment’s design leads participants to react/behave naturally
deception
false information given to participants involved in an experiment. Often done to ensure realism
confederate
accomplice of an experimenter, seemingly a participant, but pre-trained how to behave/talk etc
meta-analysis
reviews a set of many different experiments statistically, to evaluate most effective etc
informed consent
should be obtained prior to research. Participant given enough info to be informed. Informed able to withdraw participation at any stage.
debriefing
should occur after any experiment/study. Inform what happened and why. Inform re any deception. Make participant feel good about having participated.
