james heckman_apresentação seminário educação para o século 21

Introduction Lessons History IQ Origins Modern Stability Appendix References

Hard Evidence on Soft Skills

James J. HeckmanUniversity of Chicago;

University College Dublin;Director, Human Capital and Economic Opportunity Working

Group, Institute for New Economic Thinking

Noncognitive Skills: Importance andthe Relation with Cognitive Abilities

Ayrton Senna InstituteSao Paulo, BrazilOctober 25, 2011

Heckman Hard Evidence on Soft Skills


We live in an era of widespread testing.

Achievement tests in particular have assumed a prominent role.

They are used to:

i Measure skills of persons (e.g., SAT, ACT, ENEM, GRE,GED) and certify suitability for admission and qualifications ina variety of domains of life.

ii To measure the performance of schools and entire nationalschool systems and nations (e.g. PISA scores, NCLB)



We live in an era of widespread testing.

Achievement tests in particular have assumed a prominent role.

They are used to:i Measure skills of persons (e.g., SAT, ACT, ENEM, GRE,

GED) and certify suitability for admission and qualifications ina variety of domains of life.

ii To measure the performance of schools and entire nationalschool systems and nations (e.g. PISA scores, NCLB)



These tests are not well understood.

a What do they measure?b How are they validated? What do they predict?c What do they miss?



These tests are not well understood.a What do they measure?

b How are they validated? What do they predict?c What do they miss?



These tests are not well understood.a What do they measure?b How are they validated? What do they predict?

c What do they miss?



These tests are not well understood.a What do they measure?b How are they validated? What do they predict?c What do they miss?



Achievement tests miss—or perhaps better—do not accuratelycapture—soft skills.

Personality traits, goals, motivations, and preferences that arevaluable in the labor market, in school, and in many otherdomains.



A Multiple skills are useful in predicting and causing economicand social success but with different weights in different tasks.

i People differ in their endowments of these skills.ii These differences in endowments and values of endowments

across tasks give rise to comparative advantage and sorting inthe labor market.

B Personality traits—“soft skills”—can be measured.

i They are often as predictive of many life outcomes as aremeasures of cognition.

ii Personality traits have causal effects on behavioral outcomes:changing traits through interventions changes life outcomes.




i People differ in their endowments of these skills.

ii These differences in endowments and values of endowmentsacross tasks give rise to comparative advantage and sorting inthe labor market.









B Personality traits—“soft skills”—can be measured.i They are often as predictive of many life outcomes as are

measures of cognition.







B Personality traits—“soft skills”—can be measured.i They are often as predictive of many life outcomes as are

measures of cognition.ii Personality traits have causal effects on behavioral outcomes:

changing traits through interventions changes life outcomes.



C Ignoring personality traits can deceive: conceals major socialproblems by using achievement test scores as the measure ofsuccess.

D Traits are stable across situations.

i Yet they can also be changed in a gradual way throughinvestment and experience.

ii Enhancing these traits is an important avenue for policy.

E No evidence for extreme situational specificity of the sort onceadvocated by Walter Mischel (1968).

F Situational specificity — that we are what we need to be in anyspecific situation — plays a prominent role in modernbehavioral economics.

G There are “enduring traits” that persist and govern behavioracross multiple domains of economic and social life.




D Traits are stable across situations.i Yet they can also be changed in a gradual way through

investment and experience.

ii Enhancing these traits is an important avenue for policy.







D Traits are stable across situations.i Yet they can also be changed in a gradual way through

investment and experience.ii Enhancing these traits is an important avenue for policy.






I draw on recent research on the GED, an achievement test thatsecondary school dropouts in the U.S. can take to certify thatthey are the equivalents (in cognition) of high school graduates.

Brazil has a version of this test applied on a wide scale(ENEM).



What Are the Lessons from This Research?

Success in life requires a lot more than what is measured byacademic achievement tests.

Soft skills matter.

Personality—“character” broadly defined—can be measured.

Personality can be fostered. It is an important, but neglected,component of what schools and families produce.

A focus on achievement tests fostered by movements for“accountability” distorts the incentives of students andteachers.



A Brief History of Testing

Goal of developers of IQ and achievement tests: to produce“objective” and “comparable” measures of children and schools.

Pioneers recognize their limits.

They understood that schools produced more than simple“book learning” — academic knowledge — expressedfrustration about how to measure it.



Early educators lacked a conceptual framework for thinkingabout what exactly should be measured.

They lacked the technology for implementing such tests on awide spread scale.



IQ

IQ in its modern form is an early Twentieth Century invention.

It was conceived as a measure of the capacity to learn.

Was later broadened to a concept of ability to solve abstractproblems.

IQ was tested on a wide scale.

The successful application of the IQ test suggested thepossibility of testing achievement — what was learned — notjust the ability to learn — on a wide scale.

First IQ test was designed to screen out misfits in school.

Role of the test was broadened to capture “intelligence” andquickly became associated with “g”. (Terman and Thorndike)

But the creators of IQ tests realized their limitations.



Binet:

“[Success in school] . . .admits of other things thanintelligence; to succeed in his studies, one must havequalities which depend on attention, will, and character;for example a certain docility, a regularity of habits, andespecially continuity of effort. A child, even if intelligent,will learn little in class if he never listens, if he spends histime in playing tricks, in giggling, in playing truant.”

-Binet (1916, p. 254)



“What are the chief personality traits which, interactingwith g, relate to individual differences in achievement andvocational success? The most universal personality trait isconscientiousness, that is, being responsible, dependable,caring, organized and persistent.”

-Jensen (1998, p. 575)



The modern conception of intelligence has moved well beyond“g” to a hierarchy of traits.

But “g” is still at the top of the hierarchy.



Figure 1: Modern View of “g”: An Hierarchical Scheme of GeneralIntelligence and Its Components

GeneralIntelligence

(Fluid Intelligence)Sequential ReasoningInductive ReasoningQuantitative ReasoningPiagetian Reasoning

Math ReasoningQuantitative ReasoningMath Problems

Visual PerceptionVisualizationSpatial RelationsClosure SpeedClosure FlexibilitySerial Perceptual IntegrationSpatial ScanningImagery

ClosureClosure SpeedClosure Flexibility

Perceptual SpeedNumber ComputationRT and other Elementary Cognitive TasksStroopClerical SpeedDigit/Symbol

Learning and MemoryMemory SpanAssociative MemoryFree Recall MemoryMeaningful MemoryVisual Memory

Knowledge and AchievementGeneral School AchievementVerbal Information and KnowledgeInformation and Knowledge, Math and ScienceTechnical and Mechanical KnowledgeKnowledge of Behavioral Content

Ideational FluencyIdeational FluencyNaming FacilityExpressional FluencyWord FluencyCreativityFigural FluencyFigural Flexibility

Gc(Crystallized Intelligence)Verbal ComprehensionLexical KnowledgeReading ComprehensionReading Speed“Cloze”SpellingPhonetic CodingGrammatical SensitivityForeign LanguageCommunicationListeningOral ProductionOral StyleWriting

g=

g

g

f

c

Source: Recreated from Ackerman and Heggestad (1997), based on Carroll(1993).



Operationalizing IQ



Problem similar to the Raven’s Progressive Matrices test items

Correct answer: 5



Origins of the Modern Achievement Test

Achievement tests were created in the wake of the IQ test as away to capture the knowledge acquired in schools, not just thecapacity to learn.

The distinction between knowledge and the ability to acquire itis clear.

How one separates these concepts empirically is not so clear,especially since the ability to acquire knowledge is not a fixedtrait but can be changed.



General Knowledge

The inventors of the achievement test sought to measure“useful knowledge” or “general knowledge” valuable infunctioning at work and in society, not specific knowledge of acourse.

Designed to be “objective”—not depend on teacherassessments as captured by grades. This was perceived to be away to implement meritocratic notions of education.

Iowa tests; ACT; GED; ENEM; No Child Left Behind; NAEP;PISA tests are modern versions.



Iowa Test of Educational Development (ITED)

Developed in 1930s and 1940s.

The prototype for all achievement tests.



What matters most is not how many detailed facts aperson can quote, but how well he can make use of whathe knows, and how straight he can think about personaland group problems.

The Iowa Tests, therefore, were planned to measure thefundamental ideas the student has acquired permanently,and what he is able to do with them.

This seemed much more important than just measuringthe specific information he had memorized in a course,which he would forget after the final exam.



They were planned to measure the real knowledge andreasoning ability a student had, no matter how he hadacquired them . . .

-Lindquist (1948)



Achievement tests are designed to measure acquired knowledge.

But what motivates students to acquire knowledge?

Cognition, personality, as well as environmental determinants,such as parents, teachers, and peers.



What Do Schools Produce?

Is it just the knowledge captured by standard achievementtests?

Every creator of an achievement test thought otherwise.



Mann:

“Hence to value schools, by length instead of quality, is amatchless absurdity. Arithmetic, grammar, and the otherrudiments, as they are called, comprise but a small part ofthe teachings in a school. The rudiments of feeling aretaught not less than the rudiments of thinking. Thesentiments and passions get more lessons than theintellect. Though their open recitations may be less, theirsecret rehearsals are more.”

-Horace Mann (1867, p. 420)



Other mechanisms of assessment besides the objective examwere often suggested.



“We lean heavily on written examinations, on a few typesof objective tests, and on the subjective impressions ofteachers. Many other appraisal devices could be used, suchas records of activities in which pupils participate,questionnaires, check lists, anecdotal records andobservational records, interviews, reports made by parents,products made by the pupils, and records made byinstruments (motion pictures, eye-movement records,sound recordings, and the like).”

-Tyler (1940)



But time and again, the achievement test won out as a cheaptool of assessment.

Other methods were more costly, even if they missed a lot.



How are IQ and achievement tests validated?

Usually on grades and other tests.

This is ironic in light of the fact that the achievement test wasinvented in order to avoid the “subjectivity” of grades.



Table 1: Predictive Validities of Standard IQ and Achievement Tests

Cognitive Achievement and IQ Tests

Test Domain overwhich it isvalidated

Estimated Validities Source Notes

SAT First year col-lege GPA

0.35 to 0.53 Kobrin et al. (2008)

ACT Grades inearly years ofcollege

0.42 ACT Incorporated(2007)

Stanford-Binet

Correlationswith otherintelligencetests

0.77 to 0.87 with WISC-R Rothlisberg (1987);Greene et al. (1990)

WISC(WechslerIntelligenceScale forChildren)

Correlationswith aca-demicachievement

WISC: 0.443 to 0.751with WRAT tests, 0.482to 0.788 with 1st gradegrades, 0.462 to 0.794 with2nd grade grades; WISC-R:0.346 to 0.760 with WRATtests, 0.358 to 0.537 with1st grade grades, 0.420to 0.721 with 2nd gradegrades

Hartlage and Steele(1977)

WRAT =Wide RangeAchievementTest; rangesare givenbecause cor-relations varyby academicsubject

Source: Almund et al. (2011).Heckman Hard Evidence on Soft Skills






WAIS (Wech-sler AdultIntelligenceScale)

Correlationswith otherintelli-gence tests,achievementtests, andoutcomes

0.67 (median) with ver-bal tests, 0.61 (median)with nonverbal tests, 0.69with education attained,0.38 to 0.43 with col-lege grades, 0.62 with highschool grades

Feingold (1982)

Raven’sStandardProgressiveMatrices

Correlationswith otherintelligencetests

0.74 to 0.84 with WAIS-R O’Leary et al.(1991)

GATB (Gen-eral AptitudeTest Battery)

Supervisorrating per-formancein trainingprogramsand in jobperformance

0.23 to 0.65 Hunter (1986) Large rangedue to varietyof jobs

Source: Almund et al. (2011).







ASVAB(ArmedServicesVocationalAptitudeBattery)

Performancein militarytrainingprogramsand militaryattrition rates

0.37 to 0.78 for training(mean=0.56); -0.15 for at-trition

Schmidt et al.(1988) for perfor-mance in trainingprograms; Stichtet al. (1982) forattrition rates

Large rangein trainingcorrelationsdue to avariety ofjobs

GED (Gen-eral Ed-ucationalDevelop-ment)

Test difficultyis normedagainst grad-uating HSseniors. Testscores of highschool seniorsand grades ofhigh schoolseniors

0.33 to 0.49 for HS SeniorGPA

Technical Manual:2002 Series GEDTests

Other Valid-ity Studies ofthe GED

Correlationwith otherachievementtests

.75–.79 AFQT, .78 NALS,

.81 Iowa TestsBoesel et al. (1998)

Source: Almund et al. (2011).Heckman Hard Evidence on Soft Skills






DAT (Dif-ferentialAptitudeTests)

Correlationswith aca-demicachievement

0.13 to 0.62 for collegeGPA

Omizo (1980) Large range isdue to vary-ing validity ofeight subtestsof DAT

WIAT(WechslerIndividualAchievementTest)

Correlationwith otherachievementtests; teacherratings ofstudentachievement

0.80 with grade 4 CAT/2,0.69 with grade 5 CAT/2,0.83 with grade 6 CAT/2;0.67 with teacher ratings

Michalko andSaklofske (1996)

CAT=CaliforniaAchievementTest

Source: Almund et al. (2011).



Only recently has the validity of the tests on real worldoutcomes been established.

They predict a bit, but they leave a lot unexplained.

Achievement tests and grades are usually more predictive thanIQ.



Table 2: Validities in Labor Market Outcomes from the NationalLongitudinal Survey of Youth, 1979

NLSY79 Table of Proportion of Explained Variance (tests and school performance)

Source: Borghans et al. (2011)



Personality

Personality and Motivation: missing dimensions that predictadult performance.

Some of the missing dimensions manifest in these tables thatexplain a variety of educational and labor market outcomes.

Economist’s “Dark Matter” or “Dark Energy.”

One popular measurement system: Big Five — OCEAN

OpennessConscientiousnessExtraversionAgreeablenessNeuroticism



Personality





Openness

ConscientiousnessExtraversionAgreeablenessNeuroticism



Personality





OpennessConscientiousness

ExtraversionAgreeablenessNeuroticism



Personality





OpennessConscientiousnessExtraversion

AgreeablenessNeuroticism



Personality





OpennessConscientiousnessExtraversionAgreeableness

Neuroticism



Personality





OpennessConscientiousnessExtraversionAgreeablenessNeuroticism



Table 3: The Big Five Domains and Their Facets

Big Five Personality Factor

American Psychology Association Dictionary

description

Facets (and correlated trait adjective)

Related Traits Childhood Temperament Traits

Openness to Experience

“the tendency to be open to new aesthetic, cultural, or intellectual experiences”

Fantasy (imaginative) Aesthetic (artistic) Feelings (excitable) Actions (wide interests) Ideas (curious) Values (unconventional)

—

Sensory sensitivity Pleasure in low-intensity activities Curiosity

Conscientiousness “the tendency to be organized, responsible, and hardworking”

Competence (efficient) Order (organized) Dutifulness (not careless) Achievement striving (ambitious)Self-discipline (not lazy) Deliberation (not impulsive)

GritPerseverance Delay of gratification Impulse control Achievement striving AmbitionWork ethic

Attention/(lack of) distractibility Effortful control Impulse control/delay of gratification Persistence Activity*

O C E A NSource: Table adapted from John and Srivastava (1999).



Table 3: The Big Five Domains and Their Facets

Extraversion “an orientation of one’s interests and energies toward the outer world of people and things rather than the inner world of subjective experience;characterized by positive affect and sociability”

Warmth (friendly) Gregariousness(sociable) Assertiveness (self-confident) Activity (energetic) Excitement seeking (adventurous) Positive emotions (enthusiastic)

—

Surgency Social dominance Social vitality Sensation seeking Shyness*

Activity*

Positive emotionality Sociability/affiliation

Agreeableness “the tendency to act in a cooperative, unselfish manner”

Trust (forgiving) Straight-forwardness (not demanding) Altruism (warm) Compliance (not stubborn)Modesty (not show-off) Tender-mindedness (sympathetic)

Empathy Perspective taking Cooperation

Competitiveness

Irritability*

AggressivenessWillfulness

Neuroticism/ Emotional Stability

Emotional stability is “predictability and consistency in emotional reactions, with absence of rapid mood changes.” Neuroticism is “a chronic level of emotional instability and proneness to psychological distress.”

Anxiety (worrying) Hostility (irritable) Depression (not contented) Self-consciousness (shy) Impulsiveness (moody) Vulnerability to stress (not self-confident)

Internal vs. External Locus of control Core self-evaluation Self-esteemSelf-efficacyOptimism Axis I psychopathologies (mental disorders) including depression and anxiety disorders

Fearfulness/behavioral inhibition Shyness*

Irritability*

Frustration (Lack of) soothability Sadness

O C E A NSource: Table adapted from John and Srivastava (1999).



Alternative measurement systems and their communalities



Competing taxonomies of personality

Evsenck Costa & McCrae Tellegen Zuckerman Cloninger Big NineBig Three NEO-PRF Big Five MPQ

Neuroticism Neuroticism Negative Emotionality Neuroticism-Anxiety Harm Avoidance AdjustmentAnxious Anxiety Stress reaction

VulnerabilityDepressed DepressionGuilt-feelingLow self-esteemTenseIrrationalShy Self-consciousness AlienationMoodyEmotional Impulsiveness

PsychotismAggressive Hostility Aggression Aggression-Hostility Cooperativeness AgreeablenessColdEgocentric Rugged IndividualismImpersonal AgreeablenessAnti-social AltruismUnempathic ComplianceTough-minded Tendermindedness

StraightforwardnessTrustModesty

Figure XCompeting Taxonomies of PersonalityNote: Figure reproduced from Bouchard and Loehlin (2001). Figure used with permission of the publisher

(cont.)



Competing taxonomies of personality (cont.)


Conscientiousness Constraint Self-Directedness DependabilityDeliberationDutifulness

Impulsive Self-discipline Control Locus of ControlOrderCompetence Persistence AchievementAchievement striving

TraditionalismExtraversion Extraversion

Sensation-seeking Excitement seeking Harm avoidance Impulsive Sensation Novelty SeekingSeeking

VenturesomeActive Activity ActivitySurgentCarefree

Positive emotionalityAchievement Reward Dependence

Social ClosenessSociable Gregariousness Sociability AffiliationLivelyAssertive AssertivenessDominant Social Potency Potency

Positive emotions Well-beingWarmth


Psychotism (cont.)



Competing taxonomies of personality (cont.)


Openness Self-TranscendenceFantasyAesthetics AbsorptionFeelingsActionsIdeasValues




Big Five is widely used.

Criticisms of Big Five

The five-factor model is atheoretical.

The five-factor model is relatively silent on an important classof individual differences: motivation and goals.



Defining non-cognitive traits for children

Childhood temperament is the term used by developmentalpsychologists to describe the personalities of infants andchildren.

Most of the research on temperament has examined specificlower-order traits rather than broader, higher-level factors thatcharacterize studies of adult intelligence and personality.



Validities for Personality Tests

What are the validities for personality tests?

How correlated are scores on these tests with behaviors?



Table 4: Validities for Personality Tests

Source: Almlund et al. 2011Heckman Hard Evidence on Soft Skills


Table 4: Validities for Personality Tests

Source: Almlund et al. 2011



How Predictive Are These Personality Traits?

What do they predict?

With what strength?



Figure 2: Association of the Big Five and intelligence with Years ofSchooling

-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Crystalized Intelligence

Fluid Intelligence

Openness

Conscientiousness

Extraversion

Agreeableness

Emotional Stability

Standardized Regression Coefficient

Males

Unadjusted for Intelligence Adjusted for Intelligence




Figure 3: Correlations of The Big Five and Intelligence with High SchoolCourse Grades




Figure 4: Associations with Standardized Achievement Test Scores

Duckworth, Heckman, Almlund, and Kautz 10/22/2010 126

Figure 13. Associations with Standardized Achievement Test Scores

Note: The values represent standardized regression coefficients when controlling for age at IQ test, gender, and ethnicity. IQ is measured using Raven’s Progressive Matrices. Source: Duckworth (2009)

Martin and colleagues were among the first to demonstrate that teacher and parent ratings

of early childhood persistence, (low) distractibility, and (low) activity prospectively predict both

course grades and standardized achievement test scores (see Martin [1989 for a summary]).

[JJH: Related to Big Five. Discuss causality.] Likewise, in a representative sample of

Baltimore first graders, teacher ratings of attention span-restlessness in first grade predicted both

course grades and standardized achievement test scores four years later (Alexander, Entwisle and

Dauber [1993]).

More recently, in a sample of preschool children from low-income homes, parent and

teacher ratings of effortful control accounted for unique variance in standardized achievement

test scores in kindergarten, even after controlling for general intelligence (Blair and Razza

-0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

Intelligence

Openness

Conscientiousness

Extraversion

Agreeableness

Emotional Stability

Standardized Regression Coefficient

Private School

Public School



Figure 5: Associations with Job Performance

Duckworth, Heckman, Almlund, and Kautz 10/22/2010 134

predictive as IQ. Conscientiousness, however, may play a more pervasive role than IQ. The

importance of IQ increases with job complexity, defined as the information processing

requirements of the job: cognitive skills are more important for professors, scientists, and senior

managers than for semi-skilled or unskilled laborers (Schmidt and Hunter [2004]). In contrast,

the importance of Conscientiousness does not vary much with job complexity (Barrick and

Mount [1991]), suggesting that it pertains to a wider spectrum of jobs. Causality remains an

open question. The raw correlations presented in Figure 14 do not account for reverse-causality,

and the authors do not clearly delineate when the measures of personality were taken.

Figure 14. Associations with Job Performance

Note. The values for personality are correlations that were corrected for sampling error, range restriction, and measurement error. Job performance was based on performance ratings, productivity data and training proficiency. The authors do report the timing of the measurements of personality relative to job performance. The value for IQ is a raw correlation. Source(s): The values reported for personality traits come from a meta-analysis conducted by Barrick and Mount [1991]. The value for IQ and job performance was reported in Schmidt and Hunter [2004].

0 0.1 0.2 0.3 0.4 0.5 0.6

Intelligence

Openness

Conscientiousness

Extraversion

Agreeableness

Emotional Stability

Correlation



Figure 6: Correlations of Mortality with Personality, IQ, andSocioeconomic Status (SES)




Figure 7: Ever been in jail by age 30, by ability (males)

Polarization

Argument

Skills

Evidence

Critical and Sensitive Periods

Environment

Intuitive

Estimates

Illustration

Summary

.15

.05

.10

.00

NoncognitiveCognitive

0 – 20 21 – 40 41 – 60 61 – 80 81 – 100

Prob

abili

ty

Percentile

Note: This figure plots the probability of a given behavior associated with moving up in one ability distribution for someone after integrating out the other distribution. For example, the lines with markers show the effect of increasing noncognitive ability after integrating the cognitive ability.

Ever Been in Jail by Age 30, by Ability (Males)

Source: Heckman, Stixrud, and Urzua (2006).

Source: Heckman, Stixrud, and Urzua (2006)



Figure 8: Probability of being teenage and single with children (females)

Polarization

Argument

Skills

Evidence


Environment

Intuitive

Estimates

Illustration

Summary

Probability of Being Single With Children (Females)

.08

.04

.06

.02

.10

NoncognitiveCognitive

0 – 20 21 – 40 41 – 60 61 – 80 81 – 100

Prob

abili

ty

Percentile

Note: This figure plots the probability of a given behavior associated with moving up in one ability distribution for someone after integrating out the other distribution. For example, the lines with markers show the effect of increasing noncognitive ability after integrating the cognitive ability.

Source: Heckman, Stixrud, and Urzua (2006).

Source: Heckman, Stixrud, and Urzua (2006)



How Predictive Are Achievement Tests Compared toConscientiousness for Success in Post-Secondary Education?



Table 5: The Relative Predictive Power of Conscientiousness and SATScores for College GPA



What Do Grades and Scores on Achievement TestsMeasure?

How are the scores on test and grades determined?



f

c



Figure 9: Decomposing Achievement Tests and Grades into IQ andPersonality [NLSY79]

0.48

0.23

0.43

0.190.16

0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.60

R-Sq

uare

d

IQ, Rosenberg, and Rotter IQ Rosenberg and Rotter

AFQT Grades

Achievement Grades

Source: Borghans, Golsteyn, Heckman et al. [2011].



Figure 10: Decomposing Achievement Tests and Grades into IQ andPersonality [Stella Maris]

Source: Borghans, Golsteyn, Heckman et al. [2011].



This explains in part the greater predictive power ofachievement tests and grades compared to IQ in explaining realworld outcomes.



Table 2: Validities in Labor Market Outcomes from the NationalLongitudinal Survey of Youth, 1979

NLSY79 Table of Proportion of Explained Variance (tests and school performance)

Source: Borghans et al. (2011)



Causal Inference

Causality is essential in designing policy and interpretingevidence.

Correlation is not the same as causation, although a lot ofpublic policy is based only on correlations.

The doctors in Russia.



Correlation vs. CausalityA is correlated with B

What is causal?

A B?




What is causal?

A B

B A

?

?




What is causal?

A B

B A

A B

?

?

?




What is causal?

A B

B A

A B

?

?

?

C

B

? A




What is causal?

A B

B A

A B

?

?

?

C

A

? B

C

B

? A




What is causal?

A B

B A

A B

?

?

?

C

A

? B

C

A

? B

C

B

? A




What is causal?

A B

B A

A B

?

?

?

C

A

? B

C

A

? B

C

A

? B

C

B

? A



Questions

Are any of the predictive relationships between personality andcognition and outcomes causal?

Can we change cognition and personality and affect outcomesor are they fixed traits?

Is promoting cognition and personality a useful policy option?

What do we miss by ignoring soft skills?



Difficulties in Establishing Causality

Reverse causality(A ⇐⇒ B)

Measured traits can be caused in part by the outcomes beingstudied.



All psychological measurements are calibrated on taskperformances.

A fundamental interpretive problem lies at the heart of anypsychological measurement system for any particular trait.

For these traits to be accurate measures of any particular traitor set of traits, it is necessary to standardize for incentives andthe effects of other traits in performing a task.

Examples: Incentivizing IQ tests.

Can boost IQ by 15 points by giving candies for correct answers— the Black/White gap in IQ in U.S.



Examples

IQ and Achievement Test Scores Reflect Incentives and Efforts, andCapture Both Cognitive and Personality Traits



Table 6: Incentives and Performance on Intelligence TestsAlmlund, Duckworth, Heckman, and Kautz 12/31/2010

83

Table 5. Incentives and Performance on Intelligence Tests Study Sample and Study

Design Experimental

Group Effect size of incentive

(in standard deviations)

Summary

Edlund [1972]

Between subjects study. 11 matched pairs of low SES children; children were about one standard deviation below average in IQ at baseline

M&M candies given for each right answer

Experimental group scored 12 points higher than control group during a second testing on an alternative form of the Stanford Binet (about 0.8 standard deviations)

“…a carefully chosen consequence, candy, given contingent on each occurrence of correct responses to an IQ test, can result in a significantly higher IQ score.”(p. 319)

Ayllon & Kelly [1972] Sample 1

Within subjects study. 12 mentally retarded children (avg IQ 46.8)

Tokens given in experimental condition for right answers exchangeable for prizes

6.25 points out of a possible 51 points on Metropolitan Readiness Test. t = 4.03

“…test scores often reflect poor academic skills, but they may also reflect lack of motivation to do well in the criterion test…These results, obtained from both a population typically limited in skills and ability as well as from a group of normal children (Experiment II), demonstrate that the use of reinforcement procedures applied to a behavior that is tacitly regarded as “at its peak” can significantly alter the level of performance of that behavior.” (p. 483)


Within subjects study 34 urban fourth graders (avg IQ = 92.8)

Tokens given in experimental condition for right answers exchangeable for prizes

t = 5.9


Within subjects study of 12 matched pairs of mentally retarded children

Six weeks of token reinforcement for good academic performance

Experimental group scored 3.67 points out of possible 51 points on a post-test given under standard conditions higher than at baseline; control group dropped 2.75 points. On a second post-test with incentives, exp and control groups increased 7.17 and 6.25 points, respectively

Clingman and Fowler [1976]

Within subjects study of 72 first- and second-graders assigned randomly to contingent reward, noncontingent reward, or no reward conditions.

M&Ms given for right answers in contingent cdtn; M&Ms given regardless of correctness in noncontingent condition

Only among low-IQ (<100) subjects was there an effect of the incentive. Contingent reward group scored about 0.33 standard deviations higher on the Peabody Picture Vocabulary test than did no reward group.

“…contingent candy increased the I.Q. scores of only the ‘low I.Q.’ children. This result suggests that the high and medium I.Q. groups were already functioning at a higher motivational level than children in the low I.Q. group.” (p. 22)

Almlund, Duckworth, Heckman, and Kautz 12/31/2010 84

Zigler and Butterfield [1968]

Within and between subjects study of 52 low SES children who did or did not attend nursery school were tested at the beginning and end of the year on Stanford-Binet Intelligence Test under either optimized or standard conditions.

Motivation was optimized without giving test-relevant information. Gentle encouragement, easier items after items were missed, and so on.

At baseline (in the fall), there was a full standard deviation difference (10.6 points and SD was about 9.5 in this sample) between scores of children in the optimized vs standardconditions The nursery group improved their scores, but only in the standard condition.

“…performance on an intelligence test is best conceptualized as reflecting three distinct factors: (a) formal cognitive processes; (b) informational achievements which reflect the content rather than the formal properties of cognition, and (c) motivational factors which involve a wide range of personality variables. (p. 2) “…the significant difference in improvement in standard IQ performance found between the nursery and non-nursery groups was attributable solely to motivational factors…” (p. 10)

Breuning and Zella [1978]

Within and between subjects study of 485 special education high school students all took IQ tests, then were randomly assigned to control or incentive groups to retake tests. Subjects were below-average in IQ.

Incentives such as record albums, radios (<$25) given for improvement in test performance

Scores increased by about 17 points. Results were consistent across the Otis-Lennon, WISC-R, and Lorge-Thorndike tests.

“In summary, the promise of individualized incentives contingent on an increase in IQ test performance (as compared with pretest performance) resulted in an approximate 17-point increase in IQ test scores. These increases were equally spread across subtests… The incentive condition effects were much less pronounced for students having pretest IQs between 98 and 120 and did not occur for students having pretest IQs between 121 and 140.” (p. 225)

Holt and Hobbs [1979]

Between and within subjects study of 80 delinquent boys randomly assigned to three experimental groups and one control group. Each exp group received a standard and modified administration of the WISC-verbal section.

Exp 1-Token reinforcement for correct responses; Exp 2 – Tokens forfeited for incorrect responses (punishment), Exp 3-feedback on correct/incorrect responses

1.06 standard deviation difference between the token reinforcement and control groups (inferred from t= 3.31 for 39 degrees of freedom)

“Knowledge of results does not appear to be a sufficient incentive to significantly improve test performance among below-average I.Q. subjects…Immediate rewards or response cost may be more effective with below-average I.Q. subjects while other conditions may be more effective with average or above-average subjects.” (p. 83)

Many other studies (see ADHK).Heckman Hard Evidence on Soft Skills


Responsiveness to Incentives on IQ and Achievement TestsDepends on Personality

Segal (2006) shows that introducing performance-based cashincentives in a low-stakes administration of a measure of IQincreases performance substantially among roughly one-third ofparticipants.

Less conscientious men are particularly affected by incentives.

Borghans, Meijers and ter Weel (2006) show that adults spendsubstantially more time answering IQ questions when rewardsare higher, but subjects high in emotional stability andconscientiousness are much less affected by these incentives.



Personality traits affect IQ scores indirectly through theknowledge acquired by individuals who are more open toexperience, more curious and more perseverant.

There is a correlation between cognitive and non-cognitivefactors.

Hansen, Heckman and Mullen (2004), Heckman, Stixrud andUrzua (2006), and Urzua (2007) show how schooling and otheracquired traits substantially causally affect measured cognitiveand non-cognitive test scores.

IQ test performance reflects not only pure intelligence, but alsointrinsic motivation, anxiety, knowledge, and reactions toextrinsic incentives to perform well.



Task

Performance

IQ (“gf”)

Knowledge

(“gc”)

Personality

Effort



Task

Performance

IQ (“gf”)

Knowledge

(“gc”)

Personality

Effort

?



Task

Performance

IQ (“gf”)

Knowledge

(“gc”)

Personality

Effort

?

?



Causal Evidence on the Power of Personality

Present causal evidence on the power of soft skills in producinglife outcomes.

Evidence that soft skills can be fostered.

Evidence that soft skills are a mechanism of policy for reducingpoverty, promoting productivity and enhancing social justice.

Sources of Evidence

i The GEDii Perry Interventioniii Causal Effect of Schooling on Test Scores







Sources of Evidencei The GED

ii Perry Interventioniii Causal Effect of Schooling on Test Scores







Sources of Evidencei The GEDii Perry Intervention

iii Causal Effect of Schooling on Test Scores







Sources of Evidencei The GEDii Perry Interventioniii Causal Effect of Schooling on Test Scores



The GED as a case study of the power of soft skills andcosts of neglecting them

Draw in part on a forthcoming book.

Studies Of The GED Testing Program,University of Chicago Press, 2012.

GED is an achievement test that secondary dropouts can taketo certify that they are the equivalents of ordinary secondaryschool graduates.

GED is a group of 5 achievement tests normed against nationalsamples of high school graduates (70% can pass).

14% of all secondary school certificates issued in the U.S. aregiven to GEDs.



Use multiple data sets on outcomes, backgrounds, and abilitiesfor all major economic and social groups in the U.S overmultiple periods.

GEDs are as smart as secondary school graduates who do notgo on to college.

GEDs who go on to college and succeed are indistinguishablefrom other college graduates in terms of annual wage income(True for AA and BA students).



However, terminal GEDs perform at a level closer to that ofdropouts.

Identical to dropouts if we control for their greater cognitiveability.

We examine what essential life skills GEDs lack.

Comparing GEDs to Dropouts standardizes ability (as measuredby achievement tests) and demonstrates the importance ofpersonality traits in predicting life outcomes.



Figure 11: Cognitive ability by educational status (no college sample, allethnic groups)

Source: Heckman, Humphries, Urzua, and Veramendi (2010)



Figure 12: Cognitive ability by educational status (no college sample, allethnic groups)

Source: Heckman, Humphries, Urzua, and Veramendi (2010)



GEDs lack noncognitive — personality — traits measured inmany ways: behaviors and personality test scores.



Figure 13: Mean Personality Measures across Education Groups -(All Races, All Post-Secondary Education)

(a) Males (b) Females

−.8

−.6

−.4

−.2

0.2

Self−Esteem (N

LSY79)

Locus of Control (N

LSY79)

Locus of Control (N

ELS)

Self Concept (N

ELS)

Dropout GED 5% Sig (GED vs.HSG)

High School S.E. 5% Sig (GED/HSG vs.Drop)

−.8

−.6

−.4

−.2

0.2

Self−Esteem (N

LSY79)

Locus of Control (N

LSY79)

Locus of Control (N

ELS)

Self Concept (N

ELS)



Sources: National Longitudinal Survey of Youth 1979, National Educational LongitudinalSurvey. Notes Rosenberg is a ten measure self-confidence scale administered in 1980. Rotter

is a 4 (two part) measure of locus of control. Locus of control is a measure of how muchcontrol an individual believes they have over their life. The Self Concept measure included in

NELS measures evaluates the respondents sense of self-worth or self-confidence.



Figure 14: Adolescent Smoking and Drinking across Education Groups -(All Races, All Post-Secondary Education)


0.2

.4.6

Smokes by 15 (N

LSY79)D

rinks by 15 (NLSY79)

Smoke by 14 (N

LSY97)D

rink by 14 (NLSY97)

Smokes 8th G

r. (NELS)

Binge Drink 10th G

r. (NELS)



0.2

.4.6

Smokes by 15 (N

LSY79)D

rinks by 15 (NLSY79)

Smoke by 14 (N

LSY97)D

rink by 14 (NLSY97)

Smokes 8th G

r. (NELS)

Binge Drink 10th G

r. (NELS)



Sources: National Longitudinal Survey of Youth 1979, National Longitudinal Survey of Youth1997, National Educational Longitudinal Survey.



Figure 15: Sex and Violent Behavior across Education Groups -(All Races, All Post-Secondary Education)


0.1

.2.3

.4.5

Sex at 15 (NLSY79)

Fight by 14 (NLSY97)

Gang by 14 (N

LSY97)

School Fight 8th Gr. (N

ELS)



0.1

.2.3

.4.5

Sex at 15 (NLSY79)

Fight by 14 (NLSY97)

Gang by 14 (N

LSY97)

School Fight 8th Gr. (N

ELS)






Figure 16: Adolescent Criminal Behavior across Education Groups -(All Races, All Post-Secondary Education)


0.2

.4.6

.8

Minor C

rime (N

LSY79)M

ajor Crim

e (NLSY79)

Violent Crim

e (NLSY79)

Arrested by 14 (NLSY97)

Prop Crim

e by 14 (NLSY97)

Theft by 14 (NLSY97)



0.2

.4.6

.8

Minor C

rime (N

LSY79)M

ajor Crim

e (NLSY79)

Violent Crim

e (NLSY79)

Arrested by 14 (NLSY97)

Prop Crim

e by 14 (NLSY97)

Theft by 14 (NLSY97)



Sources: National Longitudinal Survey of Youth 1979 and the National Longitudinal Survey ofYouth 1997. Notes: Minor crime includes vandalism, shoplifting, petty theft, fraud, holding or

selling stolen goods. Major crime includes auto theft, breaking/entering private property,grand theft. Violent crime includes fighting, assault, aggravated assault.



Figure 17: Adolescent GPA across Education Groups -(All Races, All Post-Secondary Education )


02

46

9th Gr. G

PA

9th Gr. C

redits


High School S.E. 5% Sig (GED/HSG vs.Drop)0

24

6

9th Gr. G

PA

9th Gr. C

redits



Sources: National Longitudinal Survey of Youth 1979.



Figure 18: Highest Grade Completed in High School across EducationGroups - (All Races, All Post-Secondary Education)


05

1015

HG

C high school (N

LSY79)

HG

C high school (N

LSY97)

LGE high school (N

ELS)



05

1015

HG

C high school (N

LSY79)

HG

C high school (N

LSY97)

LGE high school (N

ELS)



Sources: National Longitudinal Survey of Youth 1979, National Longitudinal Survey of Youth1997, National Educational Longitudinal Survey. Notes: HGC high school is the total number

of years of school attended through high school. It does not include any post-secondaryeducation.



Figure 19: Distribution of Non-Cognitive Skills by Education Group

Source: Reproduced from Heckman et al. (2011). National Longitudinal Study of Youth 1979.



Gaps in achievement and personality deficits emerge early.



Figure 20: PIAT Scores across Ages and Education Groups acrossEducation Groups - (All Races, All Post-Secondary Education)


−.8

−.6

−.4

−.2

0.2

Age 6 PIAT

Age 8 PIAT

Age 10 PIAT

Age 12 PIAT

Age 14 PIAT



−.8

−.6

−.4

−.2

0.2

Age 6 PIAT

Age 8 PIAT

Age 10 PIAT

Age 12 PIAT

Age 14 PIAT



Source: Children of the National Longitudinal Survey of Youth 1979. Notes: The PeabodyIndividual Achievement Test (PIAT) is a widely childhood achievement test.



Figure 21: Behavioral Problems Index (BPI) across Education Groups -(All Races , All Post-Secondary Education )


−.5

0.5

1

Age 6

Age 8

Age 10

Age 12

Age 14



−.5

0.5

1

Age 6

Age 8

Age 10

Age 12

Age 14



Sources: Children of the National Longitudinal Survey of Youth 1979. Notes: The BehavioralProblems Index (BPI) is based on a 28 question survey given to parents about their child. The

BPI is normalized to have mean 0 and variance 1.



The early childhood environments of both dropouts and GEDsare worse than those of H.S. Grads.(Moon, 2011; Cunha et al., 2010)



Figure 22: Broken Home Rates across Education Groups -(All Races, All Post-Secondary Education)


0.2

.4.6

Broken at 14 (NLSY79)



Broken 8th Gr. (N

ELS)



0.2

.4.6




Broken 8th Gr. (N

ELS)






Figure 23: Childhood Investment across Education Groups -(All Races, All Post-Secondary Education)


−.8

−.6

−.4

−.2

0.2

Age 1−3 Material

Age 4−7 Material

Age 1−3 Cognitive

Age 4−7 Cognitive

Age 1−3 Emotional

Age 4−7 Emotional



−.8

−.6

−.4

−.2

0.2

Age 1−3 Material

Age 4−7 Material

Age 1−3 Cognitive

Age 4−7 Cognitive

Age 1−3 Emotional

Age 4−7 Emotional



Sources: Children of the National Longitudinal Survey of Youth 1979. Variable Definitions:Material resources includes the child’s access to books, toys, CD or tape player, musicalinstruments, and whether the home contains books or magazines. Cognitive stimulation

investments include how often the children are read to, taught lessons, brought to culturalevents, and characteristics of the home environment. Emotional support investments includeparents verbal and physical interactions with child, disciplinary behavior, and responsibility ofchild for household chores. Moon (2010) provides a detailed description of these measures.



Traits Predict Many Outcomes

These traits are highly predictive of who graduates fromsecondary school and who does not.

Noncognitive traits do not predict GED certification.



Figure 24: Probability of Graduating from Secondary School (bycognitive and non-cognitive skill decile)

Source: Reproduced from Heckman et al. (2011).



Figure 25: Probability of GED Certification (conditional on dropping out,by cognitive and non-cognitive decile)

Source: Reproduced from Heckman et al. (2011).



GED certificate holders attempt postsecondary education.

Few succeed.

This is a recurrent pattern.

GEDs tend to drop out of everything they start (school,marriage, jobs, military).



Figure 26: Final Educational Attainment by High School Exit Status

Males

45%

50%

33.3% 32.9% 32.3%

30%

35%

40%

ent

20%

25%

30%

Perce

9.0%

4 3%

13.9%

6.4%4.7%

7.7%

5%

10%

15%

4.3%

0.0% 0.0%1.7%

0.0% 0.2%

2.9%

0%

5%

Dropout GED High School Graduate

Some Post Secondary Education Certificate AA BA MA

Source: National Educational Longitudinal Survey 1988.



Figure 27: Final Educational Attainment by High School Exit Status

Females

10.4%

36.1%

29.0%

6 5%

14.9%

8.8% 8.4%

35.6%

10%

15%

20%

25%

30%

35%

40%

45%

50%

Perc

ent

6.5%

0.0%

3.6%

0.0%1.7%

0.0% 0.0%

4.5%

0%

5%


Some Post Secondary Education Certificate AA BA MA

Source: National Educational Longitudinal Survey 1988.



If a GED gets a BA or a MA, he/she earns at that level.

However, there is usually delay and this has substantial costs.

The present value of earnings for such people is substantially(20–30%) lower than if they had not dropped out of school.



But GEDs who do not complete a further certificate or degreeearn at the rate of dropouts for people with comparableschooling attained (at the dropout stage) and ability.

No benefit for males.

This is robust across a variety of specifications.



Figure 28: Annual Income Differences - By Age - NLSY79 -(Males, All Races)

(a) All Post-Secondary Education (b) No Post-Secondary Education

010

000

2000

030

000

Age 20 to 24 Age 25 to 29 Age 30 to 34 Age 35 to 39Raw Abil BG Raw Abil BG Raw Abil BG Raw Abil BG

GED p<0.05(GED vs.HSG) S.E.

High School p<0.05(GED/HSG vs.Dropout)

010

000

2000

030

000




Source: National Longitudinal Survey of Youth 1979. Note: Estimates for hours worked andhourly wages exclude non-workers. Controls: “Raw” – age and region or state of residence;“Abil” – AFQT adjusted for schooling at time of test; “BG” – mother’s highest gradecompleted, urban status at age 14, family income in 1979, broken home status in 1979, southat age 14 and AFQT. Regressions exclude those reporting earning more than $300,000 orworking more than 4,000 hours.



Figure 29: Hourly Wage Differences - By Age - NLSY79 -(Males, All Races)


05

10




05

10







Figure 30: Employment Differences - By Age - NLSY79 -(Males, All Races)


−.0

50

.05

.1.1

5




−.0

50

.05

.1.1

5







Figure 31: Hours Worked Differences - By Age - NLSY79 -(Males, All Races)


−20

0−

100

010

020

030

0




−20

0−

100

010

020

030

0







Figure 32: Distribution of the Effect of the GED Certificate and HighSchool Graduation on Annual Income Across Models (Males)

(a) GED Recipients vs. HighSchool Dropouts (Males)

(b) High School Graduates vs.High School Dropouts

Sources: National Longitudinal Survey of Youth 1979. Notes: This graph plots the estimated coefficients from a series oflinear regressions. All models control for region, age, year, and AFQT score. The models differ in other controls andsub-populations of the data. The set of models includes all combinations of mother’s highest grade completed, urbanresidence at age 14, family income, lives in the south at age 14, smoked at 15, has had sex by 15, has committed a majorcrime, and 9th grade GPA. The sub-populations are all combinations of race, post-secondary education (everyone, has somepost-secondary education, no post-secondary education), and age (measured in 5-year categories from 20 to 39) for males andfemales.



Figure 33: Distribution of p-values from Tests Comparing Annual Incomeof GED Recipients and High School Graduates to High School Dropouts(Males)

(a) GED Recipients vs. HighSchool Dropouts (Males)

(b) High School Graduatesvs. High School Dropouts(Males)




Figure 33: Distribution of p-values from Tests Comparing Annual Incomeof GED Recipients and High School Graduates to High School Dropouts(Males)

(c) GED Recipients vs. High School Graduates (Males)




Women

The early literature focused on results for males.

Simplifies the analysis: avoids statistical problems associatedwith selection into the labor force as an empirical issue. (Manywomen not working, and wages are missing for them.)

But misses an important empirical phenomenon.

There are GED effects for certain groups of females.

But only for employment and earnings, not hourly wage rates,or for hours of work of the employed.



Figure 34: Distribution of the Effect of the GED Certificate and HighSchool Graduation on Annual Income Across Models (Females)

(a) GED Recipients vs. HighSchool Dropouts (Females)

(b) High School Graduates vs.High School Dropouts




Figure 35: Distribution of p-values from Tests Comparing Annual Incomeof GED Recipients and High School Graduates to High School Dropouts(Females)

(a) GED Recipients vs. HighSchool Dropouts (Females)

(b) High School Graduates vs.High School Dropouts (Fe-males)




Figure 35: Distribution of p-values from Tests Comparing Annual Incomeof GED Recipients and High School Graduates to High School Dropouts(Females)

(c) GED Recipients vs. High School Graduates (Females)




Why Do Some Women Benefit?

Two groups of women benefit.

Bright girls who drop out of school early (pregnancy) but weresuccessful in school compared to other dropouts—they GEDcertify late after their children enroll in primary school.

They work hard, do not invest much on the job, have littlewage growth.

A second group is bright girls who were screw-ups in highschool, dropped out and go to and graduate from college.



This evidence points to change—either in preferences and/orconstraints that cause some to turn around.

“Desistance” in the criminology literature. (e.g., Sampson andLaub)

The females have better non-cognitive traits than males andsort on those traits.



Figure 36: Average Occupational Factor Scores by Final Education -Males

0.8

1

0.2

0.4

0.6

-0.2

0

0.2

0 8

-0.6

-0.4

-1

-0.8


Some College AA BAGraduate

Cognitive Traits Social Traits Physical Traits

Source: The American Community Survey 2009 and O-Net. Notes: All educational categories are final education at time ofinterview. Each factor is based on the following O-Net occupational importance scores: Cognitive - active learning, analyticalthinking, complex problem solving, critical thinking, deductive reasoning, inductive reasoning, interpretation of meaning, mathreasoning, mathematics, processing information, reading comprehension, creative thinking, updating knowledge andvisualization. Social - communicate to outside organizations, concern for others, customer or personal service, establishrelationships, leadership, oral expression, persuasion, social perceptiveness, speaking, writing, written expression, activelistening, and cooperation. Physical Traits - arm and hand steadiness, control and precision, coordination, depth perception,explosive strength, finger dexterity, gross body coordination, gross body equilibrium, manual dexterity, multi-limb coordination,reaction time, spatial orientation, stamina, static strength, stress tolerance, trunk strength, and wrist and finger speed.



Figure 37: Average Occupational Factor Scores by Final Education -Females

0.4

0.6

0

0.2

-0.4

-0.2

1

-0.8

-0.6

-1.2

-1


Some College AA BAGraduate

Cognitive Traits Social Traits Physical Traits

Source: The American Community Survey 2009 and O-Net. Notes: All educational categories are final education at time ofinterview. Each factor is based on the following O-Net occupational importance scores: Cognitive - active learning, analyticalthinking, complex problem solving, critical thinking, deductive reasoning, inductive reasoning, interpretation of meaning, mathreasoning, mathematics, processing information, reading comprehension, creative thinking, updating knowledge andvisualization. Social - communicate to outside organizations, concern for others, customer or personal service, establishrelationships, leadership, oral expression, persuasion, social perceptiveness, speaking, writing, written expression, activelistening, and cooperation. Physical Traits - arm and hand steadiness, control and precision, coordination, depth perception,explosive strength, finger dexterity, gross body coordination, gross body equilibrium, manual dexterity, multi-limb coordination,reaction time, spatial orientation, stamina, static strength, stress tolerance, trunk strength, and wrist and finger speed.



We have no direct measures of personality post-pregnancy, butwe have evidence on their behaviors.

They are persistent in the workplace after they re-enter.



Longitudinal analysis supports the cross section analysis.

Using a variety of procedures, employment gains for women.

No hourly wage gains — no evidence of investment.



Can People Change?

Stability of Traits

Are traits stable? Can people change?

The evidence for women suggests this might be a possibility.

But as a group, post-GED turnover behavior is quite stable.

High quit rates for all who start (school; marriage; job;military).



Figure 38: Probability of Leaving a Given State Within Five Years

Source: National Longitudinal Survey of Youth of 1979 (NLSY79) nationally-representativecross-sectional subsample and minority oversamples. Notes: The ”Loses Employment” and”Becomes Divorced” are conditional on being in an employed or married state. Individuals

who have ever been jailed are excluded from the calculation.



Figure 39: Probability of Leaving a Given State Within Five Years

Source: National Longitudinal Survey of Youth of 1979 (NLSY79) nationally-representativecross-sectional subsample and minority oversamples. Notes: The ”Loses Employment” and”Becomes Divorced” are conditional on being in an employed or married state. Individuals

who have ever been jailed are excluded from the calculation.



Are Personality Traits Ephemeral?

Claim: People adapt fully to situations.

No such thing as a stable trait.



Situational Specificity Hypothesis

The modern origins of the situational specificity hypothe-sis is based on the work of the social psychologist Walter Mischel:

“. . . with the possible exception of intelligence, highlygeneralized behavioral consistencies have not beendemonstrated, and the concept of personality traits asbroad dispositions is thus untenable”

-Mischel (1968, p. 146)



Many behavioral economists hold a similar view and appeal toMischel as a guiding influence.

“The great contribution to psychology by WalterMischel [. . . ] is to show that there is no such thing asa stable personality trait.”-Thaler (2008)



The stability of traits found for GEDs challenges the situationalspecificity hypothesis current in behavioral economics.

The evidence from the GED and much other evidence speaksstrongly against the claims of Mischel and the behavioraleconomists. (See Almlund et al., 2011.)

There is irony in Mischel’s claim for he is most famous for hisresearch on the marshmallow test.

His work shows stability of traits over 20 years.



Stability and Change in Personality Traits and Preferences

Traits change over the life cycle.



Figure 40: Cumulative Mean-Level Changes in Personality Across theLife Cycle

Note: Social vitality and social dominance are aspects of Big Five Extraversion. Cumulative d values represent total lifetimechange in units of standard deviations (“effect sizes”).Source: Figure taken from Roberts, Walton and Viechtbauer [2006] and Roberts and Mroczek [2008]. Reprinted withpermission of the authors.



Figure 41: Cumulative Mean-Level Changes in Personality Across theLife Cycle

Note: Social vitality and social dominance are aspects of Big Five Extraversion. Cumulative d values represent total lifetimechange in units of standard deviations (“effect sizes”).Source: Figure taken from Roberts, Walton and Viechtbauer [2006] and Roberts and Mroczek [2008]. Reprinted withpermission of the authors.



Figure 42: Longitudinal Analysis of Cognitive Skills

Notes: T-scores on the y-axis are standardized scores with a mean of 50 and a standard deviation of ten.Source: Figures taken from Schaie [1994]. Used with permission of the publisher.



Three Processes of Development Discussed in the Literature onPersonality and Cognition

Ontogeny (programmed developmental processes common toall persons) and sociogeny (shared socialization processes).

Personality changes through external forces above and beyondcommon ontogenic and sociogenic processes that operatethrough alterations in normal biology, such as brain lesions andchemical interventions. (Phineas Gage is the most celebratedexample.)

Investment: educational interventions and parental investmentcan affect personality throughout the lifecycle.



Ontogenic Processes



Figure 43: Proportion of Individuals in Each Age Group Scoring at orAbove the Mean for 26- to 30-Year-Olds on Indices of Intellectual andPsychosocial Maturity.

his conjecture about how basic intellectual ability and psychosocial maturity (related to,e.g., impulsivity, risk perception, sensation seeking, future orientation) evolve over thelife cycle.221 He argues that intellectual ability matures more rapidly than psychosocialmaturity. In his model, increases in adolescent risk taking are due to a restructuring ofthe brain’s dopaminergic system (responsible for the brain’s reward processing) in sucha way that immediate or novel experiences yield higher rewards, especially in the pre-sence of peers. He attributes declines in risk taking to the development of the brain’scognitive control system, specifically improvements in the prefrontal cortex that promoteaspects of executive function such as response inhibition, planning ahead, weighingrisks and rewards, and the simultaneous consideration of multiple information sources.Interestingly, even in his model, sensation seeking partially depends on the presence ofpeers, which corresponds to aspects of the situation (h in the framework of Section 3).This example highlights the difficulty in disentangling situational and biological changesin personality.

p1600 What factors other than preprogrammed genetic influences might account for mean-level changes in personality? Personality change in adulthood may be precipitated by majorshifts in social roles (e.g., getting a job for the first time or becoming a parent). If social rolechanges are experienced by most people in a population at the same time, we will observethe effects as mean-level changes in measured personality. If, on the other hand, these socialroles are not assumed synchronously, we will observe rank-order changes.

p1605 One difficulty with many of the studies that address this question is the problem ofreverse causality. Changes in personality may drive social role changes rather than theother way around.

5

15

25

35

45

55

% s

corin

g at

mea

n ad

ult le

vel

Age

10 − 11 12 − 13 14 − 15 16 − 17 18 − 21 22 − 25 26 − 30

Intellectual ability

Psychosocial maturity

f0140 Figure 1.27 Proportion of Individuals in Each Age Group Scoring at or Above the Mean for 26- to 30-Year Olds on Indices of Intellectual and Psychosocial Maturity.

Source: From Steinberg, Cauffman, Woolard, Graham, and Banich (2009) submitted for publication.

fn1110221 Spear (2000a,b) also finds that sensation seeking reaches its peak in adolescence.

Hanushek_2011 978-0-444-53444-6 00001

Personality Psychology and Economics 125

Source: From Steinberg, Cauffman, Woolard et al. (2009), submitted for publication.



These developmental processes suggest that giving secondaryschool students the option to drop out and take a GED willlead to decisions they may later regret.



Investment



Schooling boosts cognitive and noncognitive traits.

The effect is causal.

This is in agreement with common sense and all of the claimsby leading educators over the years.



Figure 44: Causal Effect of Schooling on Measures on Cognition (fromASVAB)

Source: Heckman et al. (2006).



Figure 45: Causal Effect of Schooling on Measures on Cognition (fromASVAB)




Figure 46: Causal Effect of Schooling on Two Measures ofSocioemotional Skills




Figure 47: Causal Effect of Schooling on Two Measures ofSocioemotional Skills




Additional Causal Evidence on the Power of NoncognitiveTraits and Evidence that They Can Be Boosted byInvestment

Perry Preschool Study

Early childhood intervention

Perry had lasting effects on life outcomes for both boys andgirls.

With a 7–10% annual rate of return for both(Heckman et al., 2010)

Did not boost IQ



Figure 48: Cognitive Evolution Through Time, Perry Males: MaleCognitive Dynamics



Boosted Achievement Test Scores



Figure 49: Perry Age 14 Total CAT Scores, by Treatment Group

CAT = California Achievement TestTreatment: N = 49; Control: N = 46Statistically Significant Effect for Males and Females (p-values 0.009, 0.021 respectively)Source: Heckman, Malofeeva, Pinto, and Savelyev (2008).



But we have already seen that a substantial component of thevariability in achievement test scores is due to variability inpersonality traits.

Perry Boosted Measured Noncognitive Traits



Figure 50: Personal Behavior Index, by Treatment Group



Figure 51: Socio-Emotional Index by Treatment Group

Control Treatment



Figure 52: Decomposition of Treatment Effects

Polarization

Argument

Skills

Evidence


Environment

Intuitive

Estimates

Illustration

Summary

Figure 1: Decompositions of Treatment Effects, Males

90 100%0 10 3020 40 50 60 70 80

Personal Behavior

Cognitive Factors

Other Factors

Socio-Emotional State

CAT total*, age 14(+)

Employed, age 19 (+)

Monthly Income, age 27 (+)

No tobacco use, age 27 (+)

# of adult arrests, age 27 (-)

Jobless for more than 2 years, age 40 (-)

Ever on welfare (-)

Total charges of viol.crimes with victim costs, age 40, (-)

Total charges of all crimes, age 40 (-)

Total # of lifetime arrests, age 40 (-)

Total # of adult arrests, age 40 (-)

Total # of misdemeanor arrests, age 40 (-)

Total charges of all crimes with victim costs, age 40 (-)

Any charges of a crime with victim cost, age 40 (-)

Source: Heckman, Malofeeva, et al. (2008, revised 2011).



The Dynamics of Life Cycle Investment



Gaps in cognitive and noncognitive skills open up early acrosssocial and economic groups.

For both cognitive and socioemotional traits, ability gaps acrosssocioeconomic groups open up at early ages and persist.



Figure 53: Trend in mean cognitive score by maternal education

0.5

1M

ean

co

gn

itiv

e sc

ore

3 5 8 18Age (years)

College grad

Each score standardized within observed sample. Using all observations and assuming datamissing at random. Source: Brooks-Gunn et al. (2006).




0.5

1M

ean

co

gn

itiv

e sc

ore

3 5 8 18Age (years)

College grad Some college





0.5

1M

ean

co

gn

itiv

e sc

ore

3 5 8 18Age (years)

College grad Some college HS Grad





0.5

1M

ean

co

gn

itiv

e sc

ore

3 5 8 18Age (years)

College grad Some college HS Grad Less than HS




Figure 54: Average percentile rank on anti-social behavior score, byincome quartile

(The higher the score, the worse are behavioral problems)Polarization

Argument

Skills

Evidence


Environment

Intuitive

Estimates

Illustration

Summary

Average Percentile Rank on Anti-Social Behavior Score, by Income Quartile

55

30

35

45

40

50

25

20

Lowest Income Quartile

4 Yrs 6 Yrs 12 Yrs

Scor

e Pe

rcen

tile

8 Yrs 10 Yrs

Age





Argument

Skills

Evidence


Environment

Intuitive

Estimates

Illustration

Summary


55

30

35

45

40

50

25

20

Second Income Quartile Lowest Income Quartile

4 Yrs 6 Yrs 12 Yrs

Scor

e Pe

rcen

tile

8 Yrs 10 Yrs

Age





Argument

Skills

Evidence


Environment

Intuitive

Estimates

Illustration

Summary


Third Income Quartile

55

30

35

45

40

50

25

20


4 Yrs 6 Yrs 12 Yrs

Scor

e Pe

rcen

tile

8 Yrs 10 Yrs

Age





Argument

Skills

Evidence


Environment

Intuitive

Estimates

Illustration

Summary


Third Income Quartile

55

30

35

45

40

50

25

20


Highest Income Quartile

4 Yrs 6 Yrs 12 Yrs

Scor

e Pe

rcen

tile

8 Yrs 10 Yrs

Age



The early origins of gaps in cognitive and noncognitive skillsmay suggest a genetic basis.

Cognitive and noncognitive traits are not determined solely bygenetics but genetics plays a role.

50% of variance in the traits is heritable.

Family investment and early childhood programs promote bothcognitive and noncognitive skills.



IQ can be fostered in the very early years (0-3).

IQ becomes rank stable by the early teenage years.

Achievement (crystallized intelligence or knowledge) can beacquired throughout ones lifetime but not raw “fluid”intelligence.

Personality skills are more malleable until later ages.

Schools and early family environments (parenting practices)serve to shape these skills.



Figure 55: A Life Cycle Framework for Organizing Studies andIntegrating Evidence: The Technology of Skill Formation

PRENATAL

BIRTH

EARLY

CHILDHOOD 0-3

LATER

CHILDHOOD 3-6

ADULTHOOD

Traits

Traits

Childhood traits

(personality, cognition,

and health)

Adult Traits

Fetal TraitsParental Prenatal

Investment

Prenatal Parental

Environments

Investment:

Parenting and

Preschool

Perinatal Parental

Environments

Parental

Environments

Parenting and

School

Parental

Environments

Parenting and

Preschool



Figure 56: Synergies among Investment and Traits

Cognition (early)

Personality (early)

Investment (early)

Cognition (later)

Personality (later)

Investment (later)



Cunha, Heckman and Schennach [2010] estimate theserelationships using longitudinal data on the development ofchildren with rich measures of parental investment and childtraits.

Persistence of traits becomes stronger as children becomeolder, for both cognitive and noncognitive capabilities.

It is more difficult to compensate for the effects of adverseenvironments on cognitive endowments at later ages than it isat earlier ages.



Explains a large body of evidence on ineffective cognitiveremediation strategies for disadvantaged adolescents.

Personality traits foster the development of cognition but notvice versa.

It is estimated to be equally easy to substitute at both stagesfor personality skills over the life cycle (Cunha, Heckman andSchennach [2010]).

Overall, 16% of the variation in educational attainment isexplained by factors extracted from cognitive traits, 12% is dueto factors extracted from personality traits, and 15% is due tofactors extracted from measured parental investments.



Prioritizing Investment

Given the fluidity and malleability of traits while young.

And the difficulty in changing traits the older the person.

We should invest relatively more in the early-preschool years.

Builds the skill base that makes later-age investment moreproductive.

Avoids costly and usually ineffective remediation.



Figure 57: Returns to a unit cruzeiro invested

AGE



Yet society invests relatively more in the later years.

Brasil used to be terrible in this regard, especially fordisadvantaged children.

Has shifted focus towards primary education.

Should shift further toward preschool years.



Costs of Neglecting Soft Skills: Lessons from the GEDTesting Program in the U.S.

Induces Dropping Out

Deceptive Statistics



The GED Induces Students to Drop Out

Loss of skills acquired in high school.

Delay in attaining a BA even for those who get it.



Figure 58: Effect of Eliminating the GED

Source: Heckman et al. (2008).aNote: The numbers in columns (1) and (2) are computed as fractions of the overallpopulation.



Figure 59: Graduation Rate Before and After Implementing the GEDProgram, California vs. All other States

75.5%

73.3%

77.8%77.9%

76.2%

79.8%

73.7%

71.3%

76.1%

73.0%

70.6%

75.5%

66%

68%

70%

72%

74%

76%

78%

80%

82%

84%

Overall Males Females

Gra

du

ati

on

Rate

Figure 11. Graduation Rate Before and After Implementing the GED Program, California vs. All other States

U.S. (Excl. CA) Pre-74 California Pre-74 U.S. (Excl. CA) Post-74 California Post-74

DiD Estimate -3.1%

Notes: Authors' calculations based on NCES data. The graduation rate is the number of regular public and private high school diplomas issued over the 14 year old population four years previous. Population totals for the U.S. were obtained from the U.S. Census Bureau. California population estimates were obtained from the California Demographic Research Unit. Huber-White robust standard errors in parentheses. State 15 year old population are used as weights. Pre-period is defined as 1971-1973 and Post-period as 1975-1977.

DiD Estimate-3.6%

DiD Estimate-2.6%




Immature decision makers tempted by bad choices (Steinberg,2009).

(a) Slowly developing prefrontal cortex.(b) Youth need guidance.(c) Giving them a GED option distorts lifetime choices.





(a) Slowly developing prefrontal cortex.

(b) Youth need guidance.(c) Giving them a GED option distorts lifetime choices.





(a) Slowly developing prefrontal cortex.(b) Youth need guidance.

(c) Giving them a GED option distorts lifetime choices.





(a) Slowly developing prefrontal cortex.(b) Youth need guidance.(c) Giving them a GED option distorts lifetime choices.



Figure 39: Proportion of Individuals in Each Age Group Scoring at orAbove the Mean for 26- to 30-Year-Olds on Indices of Intellectual andPsychosocial Maturity.

his conjecture about how basic intellectual ability and psychosocial maturity (related to,e.g., impulsivity, risk perception, sensation seeking, future orientation) evolve over thelife cycle.221 He argues that intellectual ability matures more rapidly than psychosocialmaturity. In his model, increases in adolescent risk taking are due to a restructuring ofthe brain’s dopaminergic system (responsible for the brain’s reward processing) in sucha way that immediate or novel experiences yield higher rewards, especially in the pre-sence of peers. He attributes declines in risk taking to the development of the brain’scognitive control system, specifically improvements in the prefrontal cortex that promoteaspects of executive function such as response inhibition, planning ahead, weighingrisks and rewards, and the simultaneous consideration of multiple information sources.Interestingly, even in his model, sensation seeking partially depends on the presence ofpeers, which corresponds to aspects of the situation (h in the framework of Section 3).This example highlights the difficulty in disentangling situational and biological changesin personality.

p1600 What factors other than preprogrammed genetic influences might account for mean-level changes in personality? Personality change in adulthood may be precipitated by majorshifts in social roles (e.g., getting a job for the first time or becoming a parent). If social rolechanges are experienced by most people in a population at the same time, we will observethe effects as mean-level changes in measured personality. If, on the other hand, these socialroles are not assumed synchronously, we will observe rank-order changes.

p1605 One difficulty with many of the studies that address this question is the problem ofreverse causality. Changes in personality may drive social role changes rather than theother way around.

5

15

25

35

45

55

% s

corin

g at

mea

n ad

ult le

vel

Age

10 − 11 12 − 13 14 − 15 16 − 17 18 − 21 22 − 25 26 − 30

Intellectual ability

Psychosocial maturity

f0140 Figure 1.27 Proportion of Individuals in Each Age Group Scoring at or Above the Mean for 26- to 30-Year Olds on Indices of Intellectual and Psychosocial Maturity.

Source: From Steinberg, Cauffman, Woolard, Graham, and Banich (2009) submitted for publication.

fn1110221 Spear (2000a,b) also finds that sensation seeking reaches its peak in adolescence.

Hanushek_2011 978-0-444-53444-6 00001

Personality Psychology and Economics 125

Source: From Steinberg, Cauffman, Woolard et al. (2009), submitted for publication.



Costs of dropout and delay:

(a) Delay in attainment: 10% loss in present value for each year ofdelay; delay is 2–3 years for those who attain the degree.

(b) Loss of skills acquired in high school from seat time.



Distorts Statistics

On Secondary School Dropout Rate



Inflates the High School Graduation RateEffects of the GED on the Measured High School Graduation Rate

Reduction inGraduation Rate If

GEDs ExcludedPanel A. OverallAll Races -7.4%Whites -7.5%Blacks -9.5%Hispanics -5.7%Panel B. MalesAll Races -8.1%Whites -8.7%Blacks -10.3%Hispanics -5.0%Panel C. FemalesAll Races -6.6%Whites -6.3%Blacks -8.7%Hispanics -6.5%

Notes: Authors’ calculations based on Census 2000 data (IPUMS). All estimates are weighted and race categories aremutually exclusive. Calculations are for the 20-24 year old population. Total GED recipients are estimated from GED testingservice data. The recent immigrant category contains only those who are in the civilian non-institutional population and whoemigrated to the U.S. after 1990. Those still enrolled in high school are excluded from calculations. The percentage of GEDswho are recent immigrants is estimated from CPS October data. Estimates of GEDs who are incarcerated or in the militaryare obtained from BJS and DOD data, respectively. The bias calculations are computed sequentially so that those belongingto multiple groups are only counted once. The order of the categories excluded matches the column order in each table.Source: Heckman and LaFontaine (2010).



Overall U.S. Graduation Rate by Race, Census IPUMS 1970-2000

Year of Birth Range 1946-1950 1951-1955 1956-1960 1961-1965 1966-1970 1971-1975 1976-1980Panel A. Males and FemalesAll Races 80.8% 81.0% 78.6% 77.9% 79.4% 79.2% 77.1%Whites 83.8% 84.2% 81.7% 81.6% 82.6% 82.8% 81.7%Blacks 63.7% 69.1% 68.0% 63.9% 69.2% 68.9% 66.4%Hispanics 58.6% 64.6% 65.2% 62.6% 64.9% 62.3% 62.9%Hispanics Inc. Immigrants 56.6% 54.5% 56.8% 53.3% 54.7% 52.0% 53.9%Panel B. MalesAll Races 80.8% 80.6% 76.8% 76.2% 77.1% 77.3% 74.1%Whites 84.0% 83.8% 80.2% 80.1% 80.8% 81.5% 79.1%Blacks 60.9% 66.1% 62.8% 62.3% 64.6% 65.8% 61.0%Hispanics 60.0% 64.0% 62.8% 61.0% 62.9% 61.4% 59.5%Hispanics Inc. Immigrants 58.1% 54.0% 54.6% 49.6% 51.1% 48.2% 50.1%Panel C. FemalesAll Races 80.7% 81.4% 80.5% 79.7% 80.3% 81.0% 80.3%Whites 83.6% 84.5% 83.2% 83.3% 83.3% 84.1% 84.3%Blacks 66.0% 71.5% 72.6% 66.2% 71.3% 74.5% 71.5%Hispanics 57.5% 64.4% 66.6% 71.0% 68.1% 67.3% 66.7%Hispanics Inc. Immigrants 55.4% 55.1% 59.0% 57.5% 58.6% 56.3% 58.5%

Table III. Overall U.S. Graduation Rate by Race, Census IPUMS 1970-2000

Note: Authors' calculations from Census 1970, 1980, 1990 and 2000 data. Census graduation rates are ages 20-24 or 25-29 depending on cohort and do not include recent immigrants. Recent immigrants are those who entered the U.S. within the last ten years for 20-24 year olds and within the last fifteen years for 25-29 year olds. GED recipients are estimated for each cohort using GEDTS data and are deductedfrom the Census high school completer totals. 1981-1985 estimates from 2004 ACS data. Those who report never having enrolled in school are excluded. All races calculations include Asians, Native Americans and other race groups not shown separately.

Source: Heckman and LaFontaine (2010).



Distorts Statistics on Wage Differentials



The Role of the GED in Explaining Rising Educational Wage Gaps, Malesand Females, Ages 25-29, NLSY79

Annual Earnings Weekly Wage

Growth in College-HS Gap 18.40 12.82

(8.26) (6.04)

Growth in College-Dropout Gap 23.67 13.60

(27.30) (23.23)

Source: Heckman and LaFontaine (2010).



Summary

1 Personality traits can be measured and have predictive poweron par with cognitive traits.

2 There is a growing body of evidence that shows the causalstatus of these traits—not just a correlation.

3 Both cognitive and personality traits can be enhanced by policy.

4 Cognitive traits less malleable after ages 10–12.

5 Personality traits are more malleable until later ages and are anavenue of policy through the young adult years.



Summary

6 There is a dynamic to capability formation: capabilities crossfoster each other and are self-productive.

7 It is important to lay down the foundations early when childrenare malleable.

8 Early intervention more cost effective than later liferemediation.

9 There are real costs to society of ignoring soft skills—

a produces perverse incentives for children and teachers;b produces distorted statistics.



Summary




9 There are real costs to society of ignoring soft skills—a produces perverse incentives for children and teachers;

b produces distorted statistics.



Summary




9 There are real costs to society of ignoring soft skills—a produces perverse incentives for children and teachers;b produces distorted statistics.



Appendix



Figure 56: Annual Income Differences - By Age - NLSY79 -(Females, All Races)


050

0010

000

1500

020

000




050

0010

000

1500

020

000




Source: National Longitudinal Survey of Youth 1979. Controls: “Raw” – age and region orstate of residence; “Abil” – AFQT adjusted for schooling at time of test; “BG” – mother’shighest grade completed, urban status at age 14, family income in 1979, broken home statusin 1979, south at age 14 and AFQT. Regressions exclude those reporting earning more than$300,000 or working more than 4,000 hours.



Figure 57: Hourly Wage Differences - By Age - NLSY79 -(Females, All Races)


−2

02

46

8




−2

02

46

8







Figure 58: Employment Differences - By Age - NLSY79 -(Females, All Races)


0.1

.2.3

.4




0.1

.2.3

.4







Figure 59: Hours Worked Differences - By Age - NLSY79 -(Females, All Races)


−20

00

200

400

600




−20

00

200

400

600






Glossary of Psychology Terms

Psychology Term

Definition (American Psychological Association 2007 Dictionary definition in quotes) Example Measures

Affect

“Any experience of feeling or emotion, ranging from suffering to elation, from the simplest to the most complex sensations of feeing, and from the most normal to the most pathological emotional reactions.”

Positive and Negative Affect Scale (PANAS; (Watson, Clark, & Tellegen, 1988))

Big Five Agreeableness

“The tendency to act in a cooperative, unselfish manner, construed as one end of a dimension of individual differences (agreeableness vs. disagreeableness) in the Big Five personality model.”

Agreeableness domain of any Big Five questionnaire.

Big Five Conscientiousness

“The tendency to be organized, responsible, and hardworking, construed as one end of a dimension of individual differences (conscientiousness vs. lack of direction) in the Big Five personality model.”

Conscientiousness domain of any Big Five questionnaire

Big Five Extraversion

“An orientation of one’s interests and energies toward the outer world of people and things rather than the inner world of subjective experience. Extraversion is a broad personality trait and, like introversion, exists on a continuum of attitudes and behaviors. Extroverts are relatively more outgoing, gregarious, sociable, and openly expressive.”

Extraversion domain of any Big Five questionnaire.

Big Five Neuroticism (or Emotional Stability)

“One of the dimensions of the…Big Five personality model characterized by a chronic level of emotional instability and proneness to psychological distress.”

Neuroticism domain of any Big Five questionnaire.

Big Five Openness to Experience (or Intellect)

“A dimension of the Big Five personality model that refers to individual differences in the tendency to be open to new aesthetic, cultural, or intellectual experiences.”

Openness domain of any Big Five questionnaire; Typical Intellectual Engagement (Goff & Ackerman, 1992)

Big Five personality model

“A model of the primary dimensions of individual differences in personality. The dimensions are usually labeled extraversion, neuroticism, agreeableness, conscientiousness, and openness to experience, though the labels vary somewhat among researchers.”

NEO-PI-R (Costa & McCrae, 1992); Big Five Inventory (John & Srivastava, 1999)

Cognitive reflection A specific mental ability. The tendency to reflect before taking an intuitive answer as correct. Cognitive Reflection Test (Frederick, 2005)

Creativity “Ability to produce original work, theories, techniques or thoughts […] Related with imagination, expressiveness, originality.”

Creative Personality Scale (Gough, 1979)

Delay of gratification “Forgoing immediate reward in order to obtain a larger or more desirable reward in the future”

Preschool Delay of Gratification Task (Mischel & Metzner, 1962); Choice Delay task (Duckworth & Seligman, 2005)

Glossary of Psychology Terms (Continued)

Emotional Intelligence

“Ability to process emotional information and use it in reasoning and other cognitive activities. According to Mayer and Salovey 1997 model it comprises four abilities: to perceive and appraise emotions accurately, to access and evoke emotions when they facilitate cognition, to comprehend emotional language and make use of emotional information, and to regulate one’s own and others’ emotions to promote growth and well-being”

MSCEIT (Mayer, Salovey, & Caruso, 2002)

Executive Function

“Higher level cognitive processes that organize and order behavior, including logic and reasoning, abstract thinking, problem solving, planning and carrying out and terminating goal-directed behavior” Executive function generally refers to the broad array of functions attributed to the frontal cortex; some psychologists include general intelligence in this array whereas others do not.

Innumerable neuropsychology tasks (e.g., go/no-go, Stroop, Continuous Performance Task); BRIEF rating scale (Donders, 2002)

Goals and Motives

Goal: “The end state toward which a human is striving” Motive: “Physiological or psychological state of arousal that directs an organism’s energies toward a goal”

Thematic Apperception Test (McClelland, Atkinson, Clark, & Lowell, 1976; McClelland & Koestner, 1992); Jackson Personality Research Form (Jackson, 1974)

Intelligence, g, IQ

Intelligence: “The ability to derive information, learn from experience, adapt to the environment, understand and correctly utilize thought and reason. There are many different definitions of intelligence, including an operational one, proposed by Edwin Boring, that intelligence is what is tested by intelligence tests. There is currently much debate, as there has been in the past, over the exact nature of intelligence.” Intelligence is used by most psychologists synonymously with cognitive ability and mental ability. g or general factor: “A hypothetical source of individual differences in general ability, which represents individuals’ abilities to perceive relationships and to derive conclusions from them. The general factor is said to be a basic ability that underlies the performance of different varieties of intellectual tasks, in contrast to specific abilities, which are alleged each to be unique to a single task. Even theorists who posit multiple mental abilities have often suggested that a general factor may underlie these (correlated) mental abilities” IQ or intelligence quotient: “A standard measure of an individual’s intelligence level based on psychological tests.”

Weschler Intelligence Scale for Children (WISC), Weschler Adult Intelligence Scale (WAIS), Raven Progressive Matrices


Interests “Attitude characterized by a need to give selective attention to something that is significant to the individual”.

Self-Directed Search (http://www.self-directed-search.com/Holland.html), Strong Interest Inventory (http://www.cpp.com/products/strong/index.asp)

Locus of control “Perception of how much control individuals have over conditions of their lives”. Internal-External Locus of Control Scale (Rotter, 1966)

Myers-Briggs Type Indicator (MBTI)

“A personality test designed to classify individuals according to their expressed choices between contrasting alternatives in certain categories of traits. The categories, based on Jungian typology, are extraversion-Introversion, Sensing-Intuition, Thinking-Feeling, and Judging-Perceiving…The test has little credibility among research psychologists but is widely used in educational counseling and human resource management…”

MBTI (http://www.cpp.com/products/mbti/index.asp)

Psychopathology

“Patterns of behavior or thought processes that are abnormal or maladaptive”. A broad category comprising dysfunctional patterns of thought, feeling, or behavior. Most disorders are included in the Diagnostic and Statistical Manual of Mental Disorders (DSM) manual published by the American Psychiatric Association. Axis I disorders (e.g., depression) are more intense and episodic/discreet, whereas Axis II disorders (i.e., personality disorders) are more tonic and enduring.

Beck Depression Inventory; Beck Anxiety Inventory; Minnesota Multiphasic Personality Inventory (omnibus measure of multiple disorders); Child Behavior Checklist

Self-efficacy “An individual’s capacity to act effectively to bring about desired results, especially as perceived by the individual.”

Generalized self-efficacy scales1. Children's Perceived Self-Efficacy Scales (Pastorelli et al., 2001).

Self-esteem “The degree to which the qualities and characteristics contained in one’s self concept are perceived to be positive”

Rosenberg Self-Esteem Scale (Rosenberg, 1989)

Sensation Seeking

“The tendency to search out and engage in thrilling activities as a method of increasing stimulation and arousal. It takes the form of engaging in highly stimulating activities accompanied by a perception of danger.” A facet of either Big Five Conscientiousness or Extraversion

Sensation-Seeking Scale (M. Zuckerman, 1974; Marvin Zuckerman, 2006)

Social dominance A dimension of Big Five Extraversion that includes facets such as dominance, independent, and self-confidence, especially in social settings.

NEO-PI-R Assertiveness scale; 16PF Dominance Scale

Specific mental abilities

“Abilities as measured by tests of an individual in areas of spatial visualization, perceptual need, number facility, verbal comprehension, word fluency, memory, inductive reasoning and so forth”

Subtest scores on IQ tests

1 Bandura, originator of the concept "self-efficacy" emphasizes the importance of measuring domain-specific measures (Bandura, 1999).


Social vitality A dimension of Big Five Extraversion that includes traits such as sociability, positive effect, and gregariousness.

California Personality Inventory (CPI)Sociability Scale; NEO-PI-R Gregariousness and Activity Scales

Temperament (childhood)

“Basic foundation of personality, usually assumed to be biologically determined and present early in life[…] Includes characteristics such as energy level, emotional responsiveness, response tempo and willingness to explore”

Children's Behavior Questionnaire (http://www.bowdoin.edu/~sputnam/rothbart-temperament-questionnaires/)

Type A/Type B personality

Type A personality is “a personality pattern characterized by chronic competitiveness, high levels of achievement motivation, and hostility.” Type B personality is “a personality pattern characterized by low levels of competitiveness and frustration and a relaxed, easy going approach.”

Jenkins Activity Survey (Jenkins, Zyzanski, & Rosenman, 1971)

Values “A moral, social or aesthetic principle accepted by an individual (or society) as a guide to what is good, desirable or important.”

Values in Action Inventory of Strengths (Peterson & Seligman, 2004)

Ackerman, P. L. and E. D. Heggestad (1997). Intelligence,personality, and interests: Evidence for overlapping traits.Psychological Bulletin 121, 219–245.

Almlund, M., A. Duckworth, J. J. Heckman, and T. Kautz (2011).Personality psychology and economics. In E. A. Hanushek,S. Machin, and L. Woßmann (Eds.), Handbook of the Economicsof Education, Volume 4. Amsterdam: Elsevier. Forthcoming.

Boesel, D., N. Alsalam, and T. M. Smith (1998). Educational andLabor Market Performance of GED Recipients. Washington, DC:U.S. Dept. of Education, Office of Educational Research andImprovement, National Library of Education.

Carroll, J. B. (1993). Human Cognitive Abilities: A Survey ofFactor-Analytic Studies. New York: Cambridge University Press.

Feingold, A. (1982). The validity of the information and vocabularysubtests of the WAIS. Journal of Clinical Psychology 38(1),169–174.

Greene, A. C., G. L. Sapp, and B. Chissom (1990). Validation of theStanford-Binet Intelligence Scale: Fourth edition with exceptionalblack male students. Psychology in the Schools 27(1), 35–41.

Hartlage, L. C. and C. T. Steele (1977). WISC and WISC-Rcorrelates of academic achievement. Psychology in theSchools 14(1), 15–18.

Heckman, J. J., J. E. Humphries, S. Urzua, and G. Veramendi(2011). The effects of educational choices on labor market,health, and social outcomes. Unpublished manuscript, Universityof Chicago, Department of Economics.

Heckman, J. J., S. H. Moon, R. Pinto, P. A. Savelyev, and A. Q.Yavitz (2010, February). The rate of return to the HighScopePerry Preschool Program. Journal of Public Economics 94(1-2),114–128.

Heckman, J. J., J. Stixrud, and S. Urzua (2006, July). The effectsof cognitive and noncognitive abilities on labor market outcomesand social behavior. Journal of Labor Economics 24(3), 411–482.

Hunter, J. E. (1986). Cognitive ability, cognitive aptitudes, jobknowledge, and job performance. Journal of VocationalBehavior 29, 340–362.

John, O. P. and S. Srivastava (1999). The big five trait taxonomy:History, measurement and theoretical perspectives. In L. A. Pervinand O. P. John (Eds.), Handbook of Personality: Theory andResearch, Chapter 4, pp. 102–138. New York: The Guilford Press.

Kobrin, J. L., B. F. Patterson, E. J. Shaw, K. D. Mattern, andS. M. Barbuti (2008). Validity of the SAT for predicting first-yearcollege grade point average. The College Board 5, 1–10.

Michalko, K. T. and D. H. Saklofske (1996). A psychometricinvestigation of the Wechsler Individual Achievement Test with asample of Saskatchewan schoolchildren. Canadian Journal ofSchool Psychology 12(1), 44–54.

Mischel, W. (1968). Personality and Assessment. New York: Wiley.Moon, S. H. (2010). Decomposing racial skill gaps in the U.S.

Unpublished manuscript, University of Chicago, Department ofEconomics.

O’Leary, U.-M., K. M. Rusch, and S. J. Guastello (1991).Estimating age-stratified WAIS-R IQs from scores on the Raven’sStandard Progressive Matrices. Journal of ClinicalPsychology 47(2), 277–284.

Omizo, M. M. (1980). The differential aptitude tests as predictorsof success in a high school for engineering program. Educationaland Psychological Measurement 40(1), 197–203.

Rothlisberg, B. (1987). Comparing the Stanford-Binet, fourthedition to the WISC-R: A concurrent validity study. Journal ofSchool Psychology 25(2), 193–196.

Schmidt, F. L., J. E. Hunter, and M. Larson (1988, August).General cognitive ability versus general and specific aptitudes inthe prediction of training performance: Some preliminary findings.Paper presented at the 96th Annual Convention of the AmericanPsychological Association, Atlanta, GA.

Sticht, T. G., L. R. Hooke, and J. S. Caylor (1982). Literacy, oracy,and vocational aptitude as predictors of attrition and promotion inthe armed services. Professional Paper 2-82, Human ResourcesResearch Organization.

Thaler, R. H. (2008, July). A short course in behavioral economics.Edge Master Class, Sonoma, CA, July 25-27, 2008.