big five personality taxonomy: conceptualization and psychological measurement oliver p. john...

Big Five Personality Taxonomy:Conceptualization and

Psychological Measurement

Oliver P. John

University of California, Berkeley

Correlations with School Performance at 12: C and O

John, Caspi, Robins, et al. (1994). The Little Five. Child Development. Data are for 450 12-year old boys from the Pittsburgh Youth Study.

Conscien-tiousness

Openness

Good Grades

Conscien-tiousness

Agreeableness

LowDelinquency

Antecedent traits at 21 predicting peak work outcomes at 52: Betas

Women at age 52 Open- Extra- Conscien- ness

version tiousnessJob types (selection)

Artistic/investigative .38** -.01 .08

Enterprising/social -.10 .28* .01

Work achievement (P x E fit)Occupational creativity .49** .25* .04High status job .34** .40** .06

Work satisfaction -.01 .34**-.08

(appraisal/construal)

Source: George, Helson, & John (2011)

Socio-cultural factors and timing of work involvement

Work and social Open- Extra- Conscien-role measures ness version tiousness

Work participation -.04-.09 .12(ages 21-27)

Work satisfaction (27) .01 .02 .08

Early career begun and maintained to 43 .35** .32**

.07Work identity late 50s .12 .29*

.35**

Socio-cultural factorsTraditional role -.29** -.12

.23*Early divorce -.06 .17

-.25*Women’s movement .19 .34**

.05

Personality psychologists study how individuals…

• Work with others, and even more – Play, love, and bond (teams, families, …)

• Achieve complex and long-term goals• Savor and manage emotions when we

face – Success or failure (pride)– Opportunities/rewards or risk (excitement)– Stress and adversity (fear, anxiety)

Personality psychologists have identified…

• Individual characteristics that – Drive individual well-being (satisfying life)– Predict “success” at school & work (grades, $$)– Foster close, satisfying relationships (teachers)

• Our findings show they– Have some temporal consistency (reliable) but– Also malleable through the environment

• Today: can we organize these characteristics– Into a taxonomy with content domains and facets

• How can we measure them?

Overview: Three Questions

• What should we measure?– What are the major content domains? Big Five

• How should we measure it?– Questionnaires; issues with verbal reports– Data sources: self, parents, teacher – Construct validity: Prediction of life outcomes

• When should we measuring them?– What age groups are measured? In what

context?

Question 1

• What should we measure?– What are the major content domains?– 30 years ago, personality researchers looked at

long bewildering lists of characteristics – Like the ones Charles Fadel showed us– Here’s another much shorter one

What School Boards, Teachers and Parents Want in Kids

• Explore • Innovate• Be kind• Be a leader• Be creative• Cooperate• Engagement• Take risks

• Take responsibility• Confidence• Work as a team• Stay cool under pressure• Have empathy• Persist after failure• Try something new• Respect rules



• Too much!• Create some order!• Simplify!




• Too much!• Create some order!• Simplify!


• Factor analysis:• Find largest number of• independent dimensions




Openness




Conscientiousness




Agreeableness

Emotion regulation

Task performance

Open mindedness

Engaging with others

Collaboration

1 Stress resistance 2 Self-confidence 3 Emotional control 4 Self-esteem 5 Self-compassion

1 Self-discipline 2 Organisation 3 Dependability 4 Goal orientation 5 Task initiation

1 Curiosity2 Creativity3 Aesthetic Interests4 Appreciation5 Self-reflection/Awareness

1 Social connection 2 Assertiveness 3 Enthusiasm

1. Compassion 2 Respect and politeness 3 Trust4 Relationship harmony

Oliver John and Filip de Fruyt, 2015

Proposed OECD Study: Framework of S/E skills

Taxonomy: Five Psychosocial Systems and Tasks of Living

• Openness: Exploration– Interests

• Conscientiousness: Self-regulation– Standards

• Engagement/Extraversion: Approach– Rewards/gains

• Amity/Agreeableness: Belonging – Close bonds/social support

• Emotional stability: Coping– Adversity: Failures/losses/punishments

Five Reasons Why the Big 5 “Stick”

1. Relatively independent of IQ

2. Found by many independent investigators

3. Universal? Surprise: Same five domains across cultures and language communities

4. Replicate: Hallmark of good science

5. They work: Predict important outcomes

Still, you may find the Big Five a bit odd/unfamiliar Like “Green eggs and ham.” Dr. Seuss:

“You do not like them. So you SAY.

Try them, try them, and you may.”

Question 2: How to measure?

In children and adolescents:• Gold standard was behavior in specific tasks • Delay test: objectively recorded in minutes

– Too specific (one-shot); impractical

• Observational measures: People who have observed child across situations and time– Parent and Teacher reports– Peer nominations; Self-reports

• Considered “subjective” measures – Because scaling done by human judges but – May better reflect natural range of behavior– And predict important outcomes (doing it right?)

Subjective Ratings of OpennessOriginal, curious, imaginative, complex

Correlate with:Better performance on creativity testsInterest and success ($$) in investigative and artistic careersUnconventional attitudes (and hair)Intense interest/curiosityBored is worse than poor

Lower on these skills

Openness:Exploration System

• Interest, imagination, aesthetic reactions• Mental states, experiential life (oops, cognitive)• Manifestions in kids:

• curiosity, “pretend” play, imaginary friends

• Functions: very 21st century– Flexible adaptation to changing environments– Innovation through learning– Critical for change and growth (Barbie, age 45)

California Child Q-Sort Items

Parent and teacher ratings of children forOpenness: Is curious and exploring; likes to learn new things Has a vivid imagination Is creative in the way s/he thinks, plays, or works Daydreams; often lost in thoughts, fantasy world

Source: John, Caspi, et al. (1994) (20th birthday)

Conscientiousness

Self-regulation system: Meeting standards

Functions: “executive control”Initiate, coordinate, monitor, and complete complex, long-term, and goal-directed behavior

California Child Q-sort items:Has high standards for him/herself.Plans things ahead.Does not give up easily; persistent. [Grit]Makes things happen; gets things done.

Inner-city Pittsburgh boys (N=450): Hard outcome measures (grades, j-d, beh. problems)

Conscien-tiousness

Openness

Better Grades

Conscien-tiousness

Agreeableness

Less ConductProblems

Questionnaire Method and Items

• History: Q-sorts, like one-on-one interviews

Slow, inefficient; big interviewer effects

• Abbreviated to questionnaire method: efficient

Items and response format are fixed; limited

• Research on best items (John & Robins, 1993)

Concrete and observable better than abstract

“I like to go to large noisy parties”

Neutral (descriptive) better than evaluative

“I am no good at social interaction”

Short, simple, affirmative statements better

Avoid long, complex compounds and negations

Big Five Inventory (BFI): Self-report

Instructions: Here are a number of characteristics that may or may not apply to you. For example, do you agree that you like to spend time with others? Please write a number next to each statement to indicate the extent to which you agree with that statement.

I am someone who… • ______ Is talkative• _____ Is helpful and unselfish with others• _____ Does a thorough job• _____ Is relaxed, handles stress well • _____ Is curious about many different things• _____ Is easily distracted• _____ ….

Big Five Inventory (1991)Items: Grade 5 reading level (age 10)Short phrases (better than single adjectives)

“Keeps working until the task is finished”“Likes to think, play with ideas”“Remains calm in tense situations”

Rate on a 5-point scale“Yes, that’s me!” versus “No way, I don’t do that”1=“Disagree strongly”; 5=“Agree strongly”

PsychometricsRetest reliability high over 3-6 months Validated across self and observer ratings

Large data baseTranslated and adapted into 30 languagesCompleted by 2 million people (about 10 min)

Reliability and Construct Validity

• Reliability is about replicability: • Can we reproduce the same test scores

across equivalent measurements? Deals with random error.

Reliability and Construct Validity

• Validity addresses whether test scores measure what we intended, it’s about the meaning. Deals with systematic errors.

• Convergent validation across data sources Multi-method design

• Do different judges agree on their ratings of the target person, like 360 degree assessments in business world

• Round Robin design: everybody rates self and each of the other group members

Self vs. Consensual Rating by Others

K

S

J P

OConsensual Rating• How a person is generally

rated by others.

Self-Reports• How the person sees

him or herself.

I do all my HW before I play.S does all his HW before he plays.P does…

Modeling Judge Differences

K

S

J P

OPerceiver Effect• How a person generally

rates others.

Consensual Rating• How a person is generally

rated by others.


him or herself.

More Complex Judge Differences


him or herself.

Consensual Rating• How a person is generally

rated by others.

K

S

J P

OPerceiver Effect• How a person generally

rates others.

Relationship Effect• How one person uniquely

rates another.

Error

Observer ratings: Mostly idiosyncratic and biased?

• High agreement about the target among observers• Substantial agreement between self and consensual

observer rating– Correlations between .50 and .70, that is:– Improvement from chance prediction (50:50) to odds from

75:25 (3/4 correct) and 85:15 (6/7 correct)

• Findings vary by– Big Five domain (Highest agreement for E)– Length and type of acquaintanceship

• Conclusion: Subjective ratings are NOT random or idiosyncratic but reflect individual differences that can be observed and rated consensually

Other Types of Validity Evidence• Concurrent criterion validity: High C research Ss!

Arrival times (5 min. earlier)

Spend more time on the rating task

Complete more longitudinal segments (over 4 years)

• Predictive validity: Predicting future outcomesAddresses causality issue (George et al., 2011):

C at age 21 (in college) predicts more stable marriages

at 52 and financial security at 70

• Value of longitudinal studies Ultimate outcomesFriedman et al.: Terman boys study mortality

Hampson, Goldberg, et al. (2006, 2009): Hawaiian kids

Teacher ratings predict health outcomes in middle age

Extensive research on potential problems and biases with ratings

• Detect random responding (a few don’t read the items)• Correct acquiescence (nay-sayers vs. yeah-sayers)

– Esp. in 10-13 olds (Soto & John, 2008, 2009)• Minimize impression management/social desirability

– Decades of research (e.g., Paulhus & John, 1989)– Authenticity: Most people want to be themselves/real,

rather than “great” (Swann, 1989)– Items: descriptive and neutral, not evaluative

• Control for self-deception: Use multiple data sources– Self and teacher/parent ratings

• Kids under 10: Too young for questionnaire self-reports– Use individual interviews, puppet interview for 4-8

Part 3. When should we measure?

• Two critical school transitions– Into school and middle childhood: Learning

the “good student” identity: A, C, and ES up– Getting through adolescence: Learning the

skills to form an adult identity and launch into adult world

• General “typical” developmental trends– Versus individual trajectories

5 6 7

4

5

6

Agreeableness

Age (years)

Sco

re

averagetrajectory

catching up

early gain, thenslow down

Agreeableness and Conscientiousness

0.05

0.15

0.25

0.35

0.45

0.55

10 11 12 13 14 15 16 17 18 19 20

Age

Mea

n

Agreeableness

Conscientiousness

Note: Points are observed means. Trends are quadraticregression curves (Soto, John, Gosling, & Potter, 2011).

Limitations: So much we do not know

• Most research cross-sectional– Different kids at different ages (hard to compare)

• But we need longer-term longitudinal research– Study same kids over time: map individual trajectories

• Much of the research conducted in– USA, England, Germany, and Scandinavia– Now Brazil!

• But we need a true cross-country effort • We’ve learned much about “subjective” measures• No longer scary (green eggs with ham) but more to do• Measurement issues in comparing groups/nations

A Hierarchical Measurement Model: 5 Domains and 15 Facets

Big Five can be too broad

Need one level lower in hierarchy: facets

Big 5 x 3 Specific Facets = 15 Concepts

Incorporates/consistent with OECD report

Self-reports for ages 10-18• 20 min. per student: 45-60 short BFI items• 3 (or 4) items for the 15 concepts

Teacher reports for ages 5 to 12 (or 18)• 5 hours (300 min) for class with 20 students• 15 min per student: Matching 45-60 items

Parent reports? Ideal to get for all ages…

BFI with 3 Specific Facets

OpennessIntellectual interest (Likes to think, play with ideas)

Aesthetic Sensitivity (Is fascinated by art, music, or literature)

Imagination (Is original, comes up with new ideas)

ConscientiousnessOrderliness (Keeps things neat and tidy)

Perseverance (Is productive, gets things done)

Self-control (Is reliable, can always be counted on)


Agreeableness• Compassion (Considerate and kind to everyone)• Politeness (Is respectful; treats others with respect) • Trust (Assumes the best about people)

Extraversion• Sociability (Is outgoing, comfortable around people)• Assertiveness (Takes on leadership roles)• Activity/Engagement (Is full of energy, enthusiasm)


Emotional Stability• Self-confidence (Feels secure, comfortable with self)• Stress Resilience (Calm/relaxed, handles stress well)• Emotion Regulation (Keeps temper/frustration/anger

under control)

Thank you

If you are still curious …

• (and truly high in Openness) • You can find more info in• the 3rd Edition of the• Handbook of Personality

Some Specific Skills to Learn (8th Grade)

1. Help direct the discussion but don’t derail it (Make notes about issues you’d like to come back to later)

2. Respond to each other (Not just the teacher)

3. Stay curious and open to new ways of thinking (You don’t have to know “the right answer,” and there may not even be one.)

End of Version 4 of Talk

Correlations with School Performance at 12: C and O

John, Caspi, Robins, et al. (1994). The Little Five. Child Development. Data are for 450 12-year old boys from the Pittsburgh Youth Study.

Conscien-tiousness

Openness

Good Grades

Conscien-tiousness

Agreeableness

LowDelinquency

The blue Big Five: Paradigm shift in publications only since 1995

0

250

500

750

1000

1250

1500

1750

1980-1984

1985-1989

1990-1994

1995-1999

2000-2004

2005-2009

Publication year

Nu

mb

er

of

pu

blic

ati

on

s

Cattell/Eysenck

Big Five/FFM

Add graph that shows individual trajectories

What should we measure?

• Start with very specific skills?– Unlimited numbers– Complex– Highly contextualized– Often age specific

Some Specific Skills to Learn (8th Grade)

1. Help direct the discussion but don’t derail it (Make notes about ideas you’d like to come back to later)

2. Respond to each other (Not just the teacher)

3. Stay curious and open to new ways of thinking (You don’t have to know “the right answer,” and there may not even be one.)

Big Five: Broad domains of social-emotional differences

• Some brief (rushed) background!!• Let’s start with Openness, which you will find

easy to recognize• Because you HERE, at this time, far away from

home, in part because you are high in O• E.g., you like to experience other cultures, try

foods that are new to you, you are interested in ideas, willing to try out new or different ideas that you might not agree with or that challenge your own views

“Construct Validation”

• Generalizability (convergence) across different data sources, such as– Teacher, Parent, and Self-report

• Hypothesis tests: Does the measure predict expected behavior in specific settings?– C Arrival time in minutes– C Time spent working on task – C Assessments completed in longitudinal study (ouch!)

• Discriminant validity—previous examples – C and O Good grades – C and A Better conduct

E

C

O S

B

Multiple Perspectives on the Person: 360 Degree Assessment

K

S

P J

O

Modeling Convergent Validity: Agreement among Data Sources

John & Robins, 1994; Robins & John, 1997; Kwan, John, et al., 2004, 2008

Self-ratings and Ratings of Others

K

S

J P

O

Rating self• How the person sees

him or herself.• "I finish all my HW"

Rating all others• How the person sees

each of the other group-members.

• "Person S finishes all his homework"

During early school years (middle childhood)

• Transition from preschool into the first grades must be very important

• On average, children acquire increasing skills related to

• Agreeableness• Conscientiousness• Emotional stability • Bot not all kids show those gains, and follow

different trajectories

Part 3. When should we measure?

• General “typical” developmental trends– Versus individual trajectories

• During transitions, start and end point, and trajectories differ sharply across kids– Like onset and end of physical growth spurt

• Individual child: Start ahead or fall behind? • Two critical school transitions

– Into school and middle childhood: Learning the “good student” identity: A, C, and ES up

– Getting through adolescence: Learning the skills to form an adult identity and launch into adult world

Additional material (unused)

big five personality taxonomy: conceptualization and psychological measurement oliver p. john...

Documents

work satisfaction

work identity

school work grades

timing of work involvementwork

peak work outcomes

personality taxonomy

personality researchers

personality psychology