Task 2. Research types & measures Flashcards
Factors to consider when choosing your variables for your study
- Research Tradition -> if topic follows a previous one, you can choose the same variables
- Theory -> your P.O.V may influence your chosen variable
- Availability of new Techniques -> use new methods to investigate unknown variables
- Availability for Equipment
Reliability
- ability of a measure to repeatedly produce similar results
Interrater reliability
(Pearson r)
- statistical measure for establishing degree of agreements amongst observers (when - — measure consists of judgements or rating of multiple observers)
Test-retest reliability
- repeat the same test twice, long break in-between and given to the same individuals, proves stable traits
Parallel-forms reliability
- same as test-retest, except the forms are not exactly the same, but equivalent (prevent recall)
Split-half reliability
- both parallel forms are put together in one test and done in the same sitting
Accuracy
- produces results that agree with an established standard
Validity
DOES THE MEASURE MEASURE WHAT ITS SUPPOSED TO MEASURE!
Face validity
- the extent to how well a measurement APPEARS to measure what its designed to measure
Content validity
- how well does the measure cover the entire topic
Construct validity
- measure that accurately measures an abstract construct (non-measurable variable)
Criterion validity & subtypes
(concurrent & predictive)
- how well a test score measures an individuals value on some criterion
- concurrent (own test & criterion measured at the same time)
- predictive (own test predicting a later standard)
Adequacy of a dependent measure
- sensitivity: some measures may be insensitive to manipulation, whereas others show an effect
- range effects: occurs when a variable has an upper/lower limit (FLOOR & CEILING)
Behavioural measures
(dependent variable)
- recording the behaviour of the subjects while exposing to different situations, etc., offers no insight into causality
- frequency: # of occurrences over t
- latency: amount of t it takes to respond
- # of errors
Physiological measures
(dependent variable)
- monitoring bodily functions (heart rate, brain waves, etc.)
Self-report measures
Advantages & Disadvantages?
- variety of forms where participants self-report
- Likert Scale, Rating scale (0-5/10), Q-sort methodology (qualitative measurement, establishing evaluating categories & sorting them)
+ -> a lot of info, combine w/ other measures - -> open-ended tedious to analyze, socially desirable answering, truthfulness, reliability of memory, no ecological validity
Implicit measures (IAT)
Implicit Association Test - rapid testing, put of conscious control
Demand characteristics
(performance cues, role attitude, preexisting attitude)
- cues provided by the researchers about the purpose of the study and expected behaviour
- performance cues: participant formulates a hypothesis on what the idea is about -> influence behaviour
- role attitude cues: participant get cues on how theyre supposed to behave
- preexisting attitudes: cooperative or apprehensive or negative
Experimenter bias
( +expectancy effects
+reducing the bias )
- experimenter can unintentionally influence participant (e.g. gestures, mimic, clothes, etc.)
- expectancy effects: researchers preconcieved ideas about how participants should behave are subtly communicated to subject, and in turn affect the participants behaviour
- SOLUTION: single- or double-blind, automatizing the experiment
Manipulation check (of a dependent variable)
- tests whether or not independent variable had the intended effect on the participant or not
Behavioural categories
(observational research)
general and specific categories of behaviour you are interested in observing
Methods for quantifying behaviour
(observational research)
- frequency: # of times a behaviour occurs within a period
- duration: how long a behaviour lasts
- intervals: dividing observation into time intervals
Methods of coding behaviour
(observational research)
- paper-and-pencil
- audio recorder
Observer bias
->solution?
- occurs when observers know the goal of the study or the hypotheses you are testing & their observations are influenced by this info
- SOLUTION: blind observers
Quantitative data
- counting/quantifying data of behaviour (rating scales)
+ -> wide range of statistical tests available
Qualitative data
- written records of observed data that are analyzed statistically
- -> large amounts of data, needs specialized computers to analyze the data
Quantitative vs qualitative?
both.
Naturalistic observation
(definition, advantages & disadvantages)
- observing subjects in their natural habitat; nonexperimental
+ -> good insight into behaviour, extremely high external validity - -> time consuming & expensive, no insight into underlying causes
Unobtrusive observation
(observational research, nonexperimental)
observations that dont alter the natural behaviours of the subject (e.g. you hiding from view)
Etnography
(observational research, nonexperimental)
- participant/nonparticipant; overt or covert
- researcher becomes immersed in the behavioural/social system of the studied group
- used to study cultures, “undercover agent”
- overt (participants know)
- covert (unobtrusive)
Sociometry & Sociogram
(observational research, nonexperimental)
- identifying & measuring interpersonal relationships in a group
- sociogram: graphic representation of the results (measured interpersonal relationships)
Case history
(observational, nonexperimental)
- descriptive technique of studying (observing) one or more previous cases and reporting on them
Archival research
(observational, nonexperimental)
- studying existing records with a specific question in mind
Kinds of questionnaire items
- open-ended: respond in own words, highly accurate
- restricted: both positive &negative, short, mutually exclusive
- partially open-ended: similar to restricted + added “other: “ category
- rating scales
6 ways of administering questionnaires
- mail survey
- internet survey
- telephone survey
- group administered survey
- face-to-face interview
- mixed-mode