S-data: Self-reports
- Questionnaires
Benefits:
- Easy to get large subject group
- Time-efficient
- Interview: allows thorough exploration
- Some information only known by individual
A Likert-rating scale should be used to nuance the research results.
General disadvantages:
- Alexithymia (Alexithymie)
- No self-knowledge or introspection
- Dependent on motivation, capacity of person Bias:
- Social desirability
- Memory biases
O-data: Others-reports
- Professional observers
- Acquaintances
Benefits:
- Some biases less frequent
- Access to other kinds of information
- Multiple observers: interrater reliability
No idiosyncratic elements
T-data: Reactions to certain stimuli
- Standardized tests
Experimentdl methods Disadvantages:
- Risk for faking
- Ecological validity
- Effects of tester on participant
Implicit tests Developed in the 1990 by Greenwald et al.
- The (I)mplicit (A)ssociation (T)est.
- Reason: The assessment of attitudes is often influenced by social desirability, selfrepresentation, etc. and the IAT convents these problems.
Measurement of associations between concepts (reaction time)! e.g., racism, attitudes regarding elderly, political standpoints, etc.
- This circumvents self-representation – Is successful, often used in research.
Mechdnicdl registrdtion equipment
- g. an actometer
Physiologicdl ddtd
- The autonomic nervous system is used.
- g. difference between psychopaths and control group when confronted with scary pictures.
Projective techniques
- Ambiguous stimulus is described.
- Often the ‘own’ structure, the inner features and the conflicts on the stimulus are projected.
- Rorschdch-test
L-data: Data about an individual’s life
- g. education, work, accidents, marriage, internet usage. – S- and O-data are used to predict L-data.
Convergence of data
It is of great importance that the different types of data are converged.
Low convergence:
– Sources measure other phenomena/contexts
Solution = triangulation: Averaging specificity of the source e.g., conflicts in the family.
Evaluation and Buality of a study
Reliability
= Sensitivity of medsurements for unwdnted fluctudtions or vdridtions in medsurement results in the sdme individudl (medsurement errors).
Danger of response tendencies
- Acquiescence = reversed coding
- Social desirability
- Extreme answers
Test-retest relidbility
- If you do the same test several times, will the results be the same?
Interndl consistency
- Calculated using coefficient alpha
Inter-rdter relidbility
- The degree of agreement among raters
Validity
= The extent to which d study is well-founded dnd corresponds dccurdtely to the redl world. A study can be reliable, but not valid. A study can’t be valid but not reliable.
Fdce vdlidity
- Does, at first sight, the test measure what it is supposed to measure?
Content vdlidity
- Are the essential elements from the phenomenon you want to measure present?
Predictive vdlidity
- Can the test predict an external criterion?
Convergent vdlidity
- Does the test correlate with other tests that aim to assess the same phenomenon?
Discrimindnt vdlidity
- Does the test also measure things you do not intent to measure?
Construct vdlidity
- Does the test measure the theoretical construct that it is supposed to measure?
Generalizability
= The degree in which d test remdins vdlid dcross different circumstdnces.