Flashcards in Week 2 - Traits: Measurement Deck (18)
Inferences about personality can be made from a number of different data sources: self-report (S-data), informant-report (I-data), life-outcome (L-data), test (T-data)
Most frequently used type of data when making inference about personality --> clinical interviews, self-report questionnaires
Strengths: individuals have access to information about themselves that is inaccessible to anyone else
Limitations: respondents must be willing and able to answer honestly and consistently [unsystematic error --> carelessness and indifference, systematic error --> acquiescence (tendency to simply agree to items), impression management (faking/lying esp with undesirable traits), self-deception (social desirability)]
Another way to learn about an individual's personality is to father the impressions or evaluations of knowledgeable informants
Strengths: some traits crucially depend upon the report of others ie popularity, for those with limited ability to self-report
Limitations: view of personality for the outside, restricted to a limited number of situations, informant bias
Life outcomes data
Biographical data obtained from archival records that are available to public scrutiny - the "residue" of personality
Strengths: information in archival records is typically accurate and not prone to the potential biases of S-data or I-data.
Limitations: life outcomes are determined by a multitude of different factors and may reveal very little about personality
Participants are either placed or found in a "testing situation" and their behaviour is directly observed.
Strengths: gathering own information, allows for precise objective and quantifiable measurement of variables
Limitations: uncertain interpretation
Constructing self-report scales
1. Identifying potential questionnaire items (rational approach, empirical approach, theoretical approach)
2. Factor analysis of questionnaire items (correlation matrix and factor loadings)
3. Evaluation of questionnaires --> reliability (test-retest, internal consistency, inter-rater) & validity (face, criterion-related, construct)
Common sense based around the lexical hypothesis: that all important personality traits must have been encoded within the natural language.
4.5% of total English vocabulary trait terms
Four categories --> real traits of personality; present activity, temporary states of mind and mood; character evaluations; behaviour, physical qualities, talents
A largely a theoretical approach which employs criterion groups extreme on the trait or characteristic of interest (clinical and normal) --> criterion-keying
Uses a given theoretical perspective to determine content for the initial item pool
Trait data for nomothetic research
Personality questionnaires (self-report, informant-report) are typically used. T-data with children & L-data as supplementary.
Often in nomothetic research, questionnaire measures of S or I-data used to predict: L-data (eg divorce) & T-data (eg physiological reactivity).
Evaluating self-report scales
Reliability: test-retest, internal consistency, inter-rater
Validity: face validity, criterion- related, construct
Precision in measurement.
Precision is determined by the consistency of scores obtained by the same person on repeated measures.
The correlation between scores obtained by the same persons on an identical test administered on two separate occasions (usually 1mth apart).
If correlation is low, test might be measuring a transient state rather than a stable trait.
When interval between tests is greater than 6mths, correlation coefficient is typically referred to as a stability coefficient (rather than a reliability coefficient).
Split-half reliability (involves correlating one half of the test with the other)
Cronbach's alpha (the average of all possible split-halves --> both halves of the test should rank people in a similar way, if not the consistency attributed to error in content sampling)
The degree of agreement/consistency between 2+ raters/scorers
Requires some relationship be established between test scores and some criterion external to that test
Eg scores on a test of sociability should predict the number of conversations a person initiates.
Refers to whether or not a test measures the unobserved theoretical construct it purports to measure
Convergent & discriminant validity