Week 12 Lecture 1 - Personality Assessment

Domains of personality assessment

Organisational psychology - seeing if hiring candidate is suitable, will do job, not steal stationary

Clinical

Education - assess resilience or coping style

Counselling - problems in every day life, strengths and challenges

Forensic - assess to see if they will reoffend, criminal behaviour unfold over time

Measuring the unmeasurable?

1. Personality assessment faces serious challenges

2. Assessment appears to be subjective

3. There is no infallible source of information about the person

4. The ‘object’ knows it is being measured

5. Personality traits are not directly observable

Bias - see yourself more positively

Blind spots - don’t know how others see you

Measurement quality & confidence

Personality is measured by two main concepts:

Reliability: does the measurement yield consistent, dependable & error-free information

Validity: does the measurement assess what it is intended to assess & is it useful

Reliability: three varieties

Internal consistency

• Do the components of the test all cohere?

• All test items should correlate with one another

Inter-rater reliability

• Does the test provide the same information about the person when different people administer it?

Re-test reliability

• Does the test yield similar scores when it is administered to the same person on different occasions?

Three kinds of measurement error: within the test, between testers, and over time

High reliability = high consistency = low error

Validity

Validity has two components

Does the test measure what is intended to measure?

• Content validity

• Convergent validity

• Discriminant validity

Does the test provide practically useful information

• Predictive validity

Reliability & validity

So for a good test of trait X…

• All items should inter correlate

• The same score should occur whoever gives it

• People should get similar scores when they do it twice

• All items should clearly relate to the meaning of X

• It should correlate strongly with other measures of X

• It should not correlate with measures of Y & Z

• It should correlate with things that X is related to

Reliability & validity are both essential, but if reliability is low, validity cannot be high: a test full of measurement error can’t predict anything

Unreliability exists when there is inconsistency in what the test measures (scatter)

Invalidity exists when the test does not measure what it should (targeting the bullseye)

Kinds of personality measurement

Many different modes of personality assessment

• Interviews

• Personality inventories

• Projective tests

• Implicit personality tests

Interviews

Interviews are rarely used in personality assessment

• Time-consuming & labour-intensive

• Subjective (ie., poor inter-rather reliability)

• Interview interactions are prone to biases

  • Halo effect, self-fulfilling prophecy, confirming bias

Sometimes they are used for assessing attributes where the person may not be a reliable informant, and/or where interpersonal & nonverbal behaviour may be revealing

• Personality disorders

Forms

• Structured - same questions asked, same order, same words - reliable, but can’t follow up with questions

• Unstructured - not reliable

• Semi-structured - best of both worlds

  • Combines structure & flexibility

• ‘Provocative’

  • Type A personality

Inventories

Self report personality tests

Composed of multiple items

Items form scales

• Omnibus tests with many scales

• Single-scale tests

• Generally at least 10 items per scale

Variety of responses scales

• True/false

• Likert scales (strongly disagree → strongly agree)

• Item generation

• Pilot testing

• Item analysis

  • Check internal consistency

  • Factor analysis

• Select optimal items for final scales

• Re-test on new sample

• Correlate with other tests and prediction criteria

• Develop norms to allow score comparison

Problems of self-report

Inventories are vulnerable to response biases & limitations of self-knowledge

Longer tests include validity scales to check for this

• Lie scales (faking good)

• Infrequency scales (faking bad, random responding)

• Defensiveness scales (subtle guardedness)

• Inconsistency scales (carelessness, random responding)

Example: the MMPI

• Developed in 1940s for comprehensive clinical personality assessment

• 10 clinical scales, 3 validity scales; 566 items

• Scale development by selecting items that best differentiated known groups

• Scale scores converted to T-scores (M=50, SD=10)

• Interpretation of scale profiles

MMPI 4/8 code

• Odd, peculiar

• Non-conforming and resentful of authority

• Problems with impulse control

• Excessive drinking and drug abuse

• Deep feelings of insecurity;

• Avoids close relationships, impaired empathy;

• Withdraws into fantasy or strikes out in anger as defence against being hurt

• Most common diagnoses are schizophrenia (paranoid type), schizoid personality & paranoid personality

Projective tests

Developed to bypass problems of self-report

Aim to penetrate to deeper levels of personality

• Dynamics, object relations, core motives

Allied with the psychoanalytic approach

Involve deliberate ambiguity & open-endedness

• Ambiguous stimuli

• Unstructured responses

Based on the assumption that personality will be ‘projected’ onto stimuli without defensive distortions operating

Thematic Apperception Test

• Developed by Henry Murray

• Idiographic approach

• Series of monochromatic images

• Person tells extended story about what is happening in the picture

• Responses coded for repeated themes in the stories: motives attributed to protagonists, interpersonal conflict, ways of handling conflict etc.

Few widely accepted scoring conventions - a recipe for inter-scorer unreliability - but…

Rigorous scoring systems for defence mechanisms

• Denial & projections (Cramer)

System for scoring motives

• Need for achievement (McClelland); does not correlate with self-reported achievement striving

Rorschach Test

• Evolved from 19th Century parlour game

• Series of symmetrical ink lots

• Person says what object(s) the person sees (the “perception”) and what aspects of the blot lead them to see it

• Responses are scored on numerous dimensions

• Number of distinct precepts

• Complexity/integration of precepts

• Content themes

• Plausibility of percepts (ie. are they recognisable)

• Response to colour

• Use of shading, blank spaces

Critiques of projective tests

• Time consuming

• Encourages ‘wild’ unconstrained interpretation

• Low inter-scorer reliability

• Predictive validity is generally weak compared to self-report tests

• Often there is little ‘incremental validity’ beyond self-report tests

Implicit tests

• New form of testing based on rapid, ‘automatic’ responses

• In principle difficult to fake & less susceptible to response bias

• Early evidence suggests these methods have promise

Example Implicit Association Test (IAT)

Four sets of words:

• Self: me, my, mine

• Not self: they, them, their

• Extraversion: active, confident, outgoing

• Introversion: aloof, reserved, serious

→ Classify words quickly determining where it belongs in which category

→ Based on rapid responses can’t be defensive, can’t be faked