Testing for Competence Rather Than for Intelligence(word版)

Testing for Competence RatherThan for

The

testing

movement

in

the

United

States

hasbeen

a

success,

if

one

judges

success

by the usualAmerican criteria of size, influence, and igence

and

aptitude

tests

are

usednearly

everywhere

by

schools,

colleges,

and

is

a

sign

of

backwardness

not

to

havetest

scores

in

the

school

records

of

children.

TheEducational

Testing

Service

alone

employs

about2,000

people,

annually

administers

Scholastic

AptitudeTests

to

thousands

of

aspirants

to

college,

andmakes

enough

money

to

support

a

large

basic

researchoperation.

Its

tests

have

tremendous

powerover

the

lives

of

young

people

by

stamping

some

ofthem

and others

protest

of

some),

the

tests

have

served

asa

very

efficient

device

for

screening

out

black,Spanish-speaking,

and

other

minority

applicants

tocolleges.

Admissions

officers

have

protested

thatthey

take

other

qualities

besides

test

achievementsinto

account in granting admission, but carefulstudies by Wing and Wallach (1971) and

othershave shown that this is true only to a very limiteddegree.

Why should intelligence or aptitude tests haveall this power? What justifies the

use of such testsin selecting applicants for college entrance or jobs?On what

assumptions is the success of the movementbased? They deserve careful examination

beforewe

go

on

rather

blindly

promoting

the

use

oftests

as

instruments

of

power

over

the lives of manyAmericans.

The key issue is obviously the validity of socalledintelligence tests. Their use

could not bejustified unless they were valid, and it is myconviction that the

evidence for their validity is byno means so overwhelming as most of us, rather

unthinkingly,had come to think it was. In point offact, most of us just believed

the results that thetesters gave us, without subjecting them to thekind of fierce

skepticism that greets, for example,the latest attempt to show that ESP exists.

Myobjectives are to review skeptically the main lines ofevidence for the validity

of intelligence and aptitudetests and to draw some inferences from this reviewas

to new lines that testing might take in the future.

Let us grant at the outset that brain-damagedor retarded people do less well on

intelligence

teststhan

other

people.

Wechsler

(19S8)

initially

usedthis

criterion

to

validate

his

instrument,

although

ithas

an

obvious

weakness:

brain-damaged

people

doless

well

on

almost

any

test

so

that

it

is

hard

toargue

that

something

unique

called

of

intelligence

responsible

for

the

deficiency

in

testscores.

The

multimethod, multitrait criterion hasnot been applied here.

Tests Predict Grades in School

The games people are required to play on aptitudetests are similar to the games

teachers

requirein

the

classroom.

In

fact,

many

of

Binet's

originaltests

were

taken

from exercises that teachers usedin French schools. So it is scarcely surprising

thataptitude test scores are correlated highly withgrades in school. The whole

Scholastic

AptitudeTesting

movement

rests

its

case

largely

on

thissingle

undeniable

fact.

Defenders

of

intelligencetesting,

like

McNemar

(1964),

often

seem

to

besuggesting

that

this

is

the

only

kind

of

r

remarked

that

manualof

the

Differential

Aptitude

Test

of

the

PsychologicalCorporation

contains a staggering total of 4,096,yes I counted 'em, validity coefficients.

Whatmore could you ask for, ladies and gentlemen? Itwas not until I looked at the

manual

myself

(Mc-Nemar

certainly

did

not

enlighten

me)

that

I

confirmedmy

suspicion

that almost every one of those

incourses

—

in other words, performing on similar typesof tests.

So

what

about

grades?

How

valid

are

they

aspredictors?

Researchers

have

in

fact

had

greatdifficulty

demonstrating

that

grades

in

school

arerelated

to

any

other

behaviors of importance

—

other than doing well on aptitude tests. Yet thegeneral

public

—

including many psychologists andmost college officials

—

simply has been

unable

tobelieve

or

accept

this

fact.

It

seems

so

self-evidentto

educators

that

those

who

do

well

in

their

classesmust

go

on

to

do

better

in

life

that

they

systematicallyhave

disregarded

evidence

to

the

contrary

thathas

been

accumulating

for

some

time.

In

theearly

1950s,

a

committee

of

the

Social

Science

ResearchCouncil

of

which

I

was

chairman

looked

intothe

matter

and

concluded

that

while

grade

level

attainedseemed related to future measures of successin life, performance within

grade was related onlyslightly. In other words, being a high school orcollege

graduate gave one a credential that openedup certain higher level jobs, but the

poorer studentsin high school or college did as well in life as thetop students.

As a college teacher, I found thishard to believe until I made a simple check. I

tookthe

top

eight

students

in

a

class

in

the

late

1940sat

Wesleyan

University

where

I was teaching

—

allstraight A students

—

and contrasted what theywere doing in the

early

1960s

with

what

eightreally

poor

students

were

doing

—

all

of

whom

weregetting

barely

passing

averages

in

college

(C

—

orbelow).

To

my

great

surprise,

I

could

not

distinguishthe two lists of men 15-18 years were lawyers, doctors,

research scientists, andcollege and high school teachers in both only

difference

I

noted

was

that

those

withbetter

grades

got

into

better

law

or

medical

schools,but

even

with

this

supposed

advantage

they

did

nothave

notably

more

successful

careers

as

comparedwith

the

poorer

students

who

had

had

to

be

satisfiedwith

law

and

medical

schools

at

theoutset.

Doubtless

the

C

—

students

could

not

getinto

even

second-rate

law

and

medical

schools

underthe

stricter admissions testing standards of that an advantage for society?

Such outcomes have been documented carefullyby many researchers (cf. Hoyt, 1965)

both inBritain (Hudson, 1960) and in

the United (1970), in a book

suggestively titled Educationand Jobs: The Great Training Robbery, hassummarized

studies showing that neither amountof education nor grades in school are related

to vocationalsuccess as a factory worker, bank teller, orair traffic controller.

Even

for

highly

intellectualjobs

like

scientific

researcher,

Taylor,

Smith,

andGhiselin (1963) have shown that superior on-thejobperformance is related in no

way to bettergrades in college. The average college grade forthe top third in

research

success

was

2.73

(aboutB

—

)

,

and

for

the

bottom

third,

2.69

(also

B-).Such

facts

have

been

known

for

some

time.

Theymake

it

abundantly

clear

that

the

testing

movementis

in

grave

danger

of

perpetuating

a

mythologicalmeritocracy

in

which

none

of

the

measures

of

meritbears

significant

demonstrable

validity

with

respectto

any

measures outside of the charmed logists used to say as a kind of an

jokethat

intelligence

is

what

the

intelligence

tests

seems

to

be

uncomfortably

near

thewhole

truth

and

nothing

but

the

truth.

But

what'sfunny

about

it, when the public took us moreseriously than we did ourselves and used the tests

toscreen

people

out

of

opportunities

for

educationand

high-status

jobs?

And

why

call

excellence atthese test games intelligence?

Even further, why keep the best education forthose who are already doing well at

the

games?This

in

effect

is

what

the

colleges

are

doing

whenthey

select

from

their

applicants those with thehighest Scholastic Aptitude Test scores. Isn't thislike

saying

that

we

will

coach

especially

those

whoalready

can

play

tennis

well?

One

would

thinkthat

the

purpose

of

education

is

precisely

to

improvethe

performance

of

those

who

are

not

doingvery

well.

So

when

psychologists

predict

on

thebasis

of

the

Scholastic

Aptitude

Test

who

is

mostlikely

to

do

well

in

college,

they

are

suggestingimplicitly

that

these

are

the

bets

to

in

another

sense,

if

the

colleges

were

interestedin

proving

that

they

could

educate

people,

highscoringstudents might be poor bets because theywould be less likely to show

improvement in be sure, the teachers want studentswho will do well

in their courses, but should societyallow the teachers to determine who deserves

to beeducated, particularly when the performance of interestto teachers bears so

little relation to anyother type of life performance?

Do Intelligence Tests Tap Abilities ThatAre Responsible for Job Success?

Most

psychologists

think

so;

certainly

the

generalpublic

thinks

so

(Cronbach,

1970,

p. 300), butthe evidence is a whole lot less satisfactory thanone would think it

ought to be to justify suchconfidence.

Thorndike

and

Hagen

(19S9),

for

instance,

obtained12,000

correlations

between

aptitude

testscores

and

various

measures

of

later

occupationalsuccess

on

over

10,000

respondents

and

concludedthat

the

number

of

significant

correlations

did

notexceed

what

would

be

expected

by

chance.

Inother

words,

the

tests

were

invalid.

Yet

psychologistsgo on using them, trusting that the poorvalidities must be due to

restriction in range dueto the fact that occupations do not admit individualswith

lower

scores.

But

even

here

it

is

notclear

whether

the

characteristics

required

for

entryare,

in

fact,

essential

to

success

in

the

field.

Onemight

suppose

that

finger

dexterity

is

essential

tobeing

a

dentist,

and

require

a

minimum

test

scorefor

entry.

Yet,

it

was

found

by

Thorndike

andHagen

(1959)

to

be

related

negatively

to

incomeas

a dentist! Holland and Richards (1965) andElton and Shevel (1969) have shown that

no

consistentrelationships

exist

between

scholastic

aptitudescores

in

college

students

and

their

actualaccomplishments

in

social

leadership,

the

arts,

science,music, writing, and speech and drama.

Yet what are we to make of Ghiselli's (1966, p.121) conclusions, based on a review

of

50

years

ofresearch,

that

general

intelligence

tests

correlate

.42with

trainability

and .23

with

proficiency

across

alltypes

of

jobs?

Each

of

these

correlations is basedon over 10,000 cases. It is small wonder thatpsychologists

believe intelligence tests are validpredictors of job success. Unfortunately, it

is

impossibleto

evaluate

Ghiselli's

conclusion,

as

he

doesnot

cite

his

sources

and

he

does

not

state

exactlyhow

job

proficiency

was

measured

for

each

of

hiscorrelations.

We can draw some conclusions fromhis results, however, and we can make a good

guessthat job proficiency often was measured by supervisors'ratings or by such

indirect

indicators

ofsupervisors'

opinions

as

turnover,

promotion,

salaryincreases, and the like.

What

is

interesting

to

observe

is

that

intelligencetest

correlations

with

proficiency

in

higher

statusjobs

are

regularly

higher

than

with

proficiency

inlower

status jobs (Ghiselli, 1966, pp. 34, 78).Consider the fact that intelligence test

scores correlate

—

.08 with proficiency as a canvasser orsolicitor and .45 with

proficiency as a stock andbond salesman. This should be a strong clue as towhat

intelligence tests are getting at, but most observershave overlooked it or simply

assumed thatit takes more general ability to be a stock andbond salesman than a

canvasser.

But

these

twojobs

differ

also

in

social

status,

in

the

language,accent,

clothing, manner, and connections by educationand family necessary for success in

the

basic

problem

with

many

job

proficiencymeasures

for

validating

ability

tests is that theydepend heavily on the credentials the man brings tothe job

—

the

habits, values, accent, interests, etc.

—

that mean he is acceptable to management

and

toclients.

Since

we

also

know

that

social

class

backgroundis

related

to

getting

higher ability test scores(Nuttall&Fozard, 1970), as well as to having theright

personal

credentials

for

success,

the

correlationbetween

intelligence

test

scores

and

job

successoften

may

be

an

artifact,

the

product

of

theirjoint

association

with

class

status.

Employers

mayhave

a

right

to

select

bond

salesmen

who

have

goneto

the

right

schools

because

they

do

better,

butpsychologists

do

not

have

a

right

to

argue

that itis their intelligence that makes them more proficientin their jobs.

We know that correlation does not equal causation,but we keep forgetting it. Far

too

many

psychologistsstill

report

average-ability

test

scores

forhigh-

and

low- prestige occupations, inferring incorrectlythat this evidence shows it takes

more

of

thistype

of

brains

to

perform

a

high-level

than

a

lowleveljob.

For

instance,

Jensen (1972) wroterecently:

He

certainly

leaves

the

impression

that

it

is

as

we

ordinarily

think

of it

prestige. But the association canbe interpreted as meaning, just as reasonably,

thatit takes more -pull, more opportunity, to get thevocabulary and other habits

required

by

those

inpower

from

incumbents

of

high-status

l

studies

that try to separate the credentialfactor from the ability factor in job success

havebeen very few in number.

Ghiselli (1966) simply did not deal with theproblem of what the criteria of job

proficiency

maymean

for

validating

the

tests.

For

example,

he

reporteda

correlation

of .27

between

intelligence

testscores

and

proficiency

as

a

policeman

or

a

detective(p. 83), with no attention given to the very importantissues involved in

how a policeman's performanceis to be evaluated. Will supervisors' ratingsdo? If

so,

it

discriminates

against

blackpolicemen

(Baehr,

Furcon,

&Froemel,

1968)

becausewhite

supervisors

regard

them

as

what

about

the

public?

Shouldn't

their

opinionas

to

how

they

are

served

by

the

police

be

partof

the

criterion?

The

most

recent

careful

review(Kent

&

Eisenberg,

1972)

of

the

evidence

relatingability

test

scores

to

police

performance

concludedthat

there

is

no

stable,

significant

is

concrete

evidence

that

one

must

view

withconsiderable skepticism the assumed relation of intelligencetest scores to

success on the job.

One other illustration may serve to warn theunwary about accepting uncritically

simple statementsabout the role of ability, as measured byintelligence tests, in

life

outcomes.

It

is

statedwidely

that

intelligence

promotes

general

adjustmentand

results in lower neuroticism. For example,Anderson (1960) reported a significant

correlationbetween intelligence test scores obtainedfrom boys in 1950, age 14-17,

and

follow-up

ratingsof

general

adjustment

made

five

years

later.

Canwe

assume

that

intelligence

promotes

better

adjustmentto

life

as

has

been

often

claimed?

Itsounds

reasonable

until

we

reflect

that

the

is

a

test

of

ability

to

do

well

in

school(to

take

academic

type

tests),

that

many

of

Anderson'ssample

were

still

in school or getting startedon careers, and that those who are not doing well

in

school

or

getting

a

good

first

job

because

of

itare

likely

to

be

considered

poorly

adjusted

by

themselvesand

others.

Here

the

test

has

become

partof

the

criterion

and

has

introduced

the

correlationartificially.

In

case

this

sounds

like

special

reasoning,consider

the

fact,

not

commented

on

particularlyby

Anderson,

that

the

same

correlation

between

test

scores

and

adjustment

ingirls

was

an

insignificant

.06.

Are

we

to

concludethat

intelligence

does

not

promote

adjustmentin

girls? It would seem more reasonable to arguethat the particular ability tested,

here associatedwith scholastic success, is more important to success(and hence

adjustment)

for

boys

than

for

this

is

a

far

cry

from

the

careless

inference

thatintelligence

tests

tap

a

general

ability

to

adapt

successfullyto

life's

problems

because high-IQ children(read

To

make

the

point

even

more

vividly,

supposeyou

are

a

ghetto

resident

in

the

Roxbury

section

ofBoston.

To

qualify

for

being

a

policeman

you

haveto

take

a

three-hour-long

general intelligence testin which you must know the meaning of words like

playanalogy

games

with

them,

you

do

not

qualify andmust

be

satisfied

with

some

such

job as being ajanitor for which an

Massachusetts

Civil

Service

,

not

unreasonably,

feel

angry,

upset,and

unsuccessful. Because you do not know thosewords, you are considered to have low

intelligence,and

since

you

consequently

have

to

take

a

lowstatusjob

and

are

unhappy,

you

contribute

to

thecelebrated

correlations

of

low

intelligence

with

lowoccupational

status

and

poor

adjustment.

Psychologistsshould

be

ashamed

of

themselves for promotinga view of general intelligence that has encouragedsuch a

testing program, particularly whenthere is no solid evidence that significantly

relatesperformance

on

this

type

of

intelligence

test

withperformance

as

a

policeman.

The Role oj Power in Controlling Life-Outcome Criteria

Psychologists

have

been,

until

recently,

incrediblynaive

about

the

role

of

powerful

interests in controllingthe criteria against which psychologistshave validated

their

tests.

Terman

felt

that

hisstudies

had

proved

conclusively

that

success. By and large, psychologistshave agreed with him. Kohlberg, LaCrosse,and

Ricks

(1970),

for

instance,

in

a

recent

summarystatement concluded

that

Terman

and

Oden's(1947)

study

the

gifted

were

more

successfuloccupationally,

maritally, and socially thanthe average group, and were lower in 'morallydeviant'

forms

of

psychopathology

(e.g.,

Jensen

(1972)

agreed:

One

of

the

most

convincing

demonstrations

that

I.Q.

isrelated

to

life

indicators of ability was provided ina classic study by Terman and his associates

at StanfordUniversity. . . . Terman found that for the most partthese high-I.Q.

children

in

later

adulthood

markedly

excelledthe

general

population

on

every

indicator

of

achievementthat

was

examined:

a

higher

level

of

education

completed;more scholastic honors and awards; higher occupationalstatus; higher

income; production of more articles,books, patents and other signs of creativity;

more

entriesin

Who's

Who;

a

lower

mortality

rate;

better

physical

andmental

health;

and

a

lower

divorce

rate.

.

.

.

Findings

suchas

these

establish

beyond

a

doubt

that

I.Q.

tests

measurecharacteristics

that

are

obviously

of

considerable

importancein

our present technological society. To say that the kindof ability measured by

intelligence tests is irrelevant orunimportant would be tantamount to repudiating

civilizationas we know it [p. 9],

I

do

not

want

to

repudiate

civilization

as

we

knowit,

or

even

to

dismiss

intelligence

tests

as

irrelevantor

unimportant,

but

I

do

want

to

state,

as

emphaticallyas

possible,

that

Terman's

studies

do

notdemonstrate

unequivocally

that

it

is

the

kind

ofability

measured

by

the

intelligence

tests

that

isresponsible

for

(i.e.,

causes)

the

greater

success ofthe high- IQ children. Terman's studies may showonly that the rich and

powerful

have

more

opportunities,and

therefore

do

better

in

life.

And

if

thatis

even

possibly

true,

it

is

socially

irresponsible

tostate

that

psychologists

have

established

adoubt

that

the

kind

of

ability

measured

by

intelligencetests

is

essential

for

high-level

performancein

our

society.

For,

by

current

methodologicalstandards, Terman's studies (and others like them)were naive. No

attempt

was

made

to

equate

foropportunity

to

be

successful

occupationally

andsocially.

His

gifted

people

clearly

came

fromsuperior

socioeconomic

backgrounds

to those hecompared them with (at one point all men in California,including day

laborers). He had no unequivocalevidence that it was

his

test

scores)

that

was

responsible

forthe

superior

performance

of

his

group.

It

would beas legitimate (though also not proven) to concludethat sons of the rich,

powerful,

and

educated

wereapt

to

be

more

successful

occupationally,

maritally,and

socially

because

they

had

more

material

make

the

point

in

another

way,

considerthe

data

in

Table

1,

which

are

fairly

representativeof

findings

in

this

area.

They were obtainedby Havighurst, Bowman, Liddle, Matthews,and Pierce (1962) from

a

typical

town

in

MiddleAmerica.

One

observes

the

usual

strong

relationshipbetween

social

class

and

IQ

and

between

IQand

college- going

—

which

leads

on

to

occupationalsuccess. The traditional interpretation of suchfindings is that more

stupid

children

come

from

thelower

classes

because

their

parents

are

also

stupidwhich

explains

why

they

are

lower

class.

A

higherproportion

of

children

with

high IQ go to collegebecause they are more intelligent and more suitedto college

study. This is as it should be becauseIQ predicts academic success. The fact that

moreintelligent

people

going

to

college

come

more

oftenfrom

the

upper

class

follows

naturally

because

theupper

classes

contain

more

intelligent

people.

Sothe

traditional argument has gone for years. Itseemed all very simple and obvious to

Terman andhis followers.

However, a closer look at Table 1 suggests anotherinterpretation that is equally

plausible,

thoughnot

more

required

by

the

data

than

the

one

justgiven.

Compare

the

percentages going to collegein the

—

high socioeconomic statusand

low IQ versus high IQ and low socioeconomicstatus. It appears to be no more likely

for thebright children (high IQ) from the lower classes togo to college (despite

their high aptitude for it)than for the

Why

is

this?

An

obvious

possibility

isthat

the

bright

but

poor

children

do

not

have

themoney

to

go

to

college,

or

they

do

not

want

to

go,preferring

to

work

or

do

other

things. In the currentlingo, they are

not

had

access

to

the

other

factors(values,

aspirations,

money)

that

promote

collegegoingin upper-class children. But now we havean alternative explanation of

college- going

—

namely,socioeconomic

status

which

seems

to

be

as

good

apredictor

of

this

type

of

success

as

ability.

Howcan

we

claim

that

ability

as

measured

by

thesetests

is

the

critical

factor

in

college-going?

Veryfew

children,

even

with

good

test-taking ability, goto college if they are from poor families. One couldargue

that

they

are

victims

of

oppression:

theydo

not

have

the

opportunity

or

the

values

thatpermit or encourage going to college. Isn't it likelythat the same oppressive

forces may have preventedeven more of them from learning to playschool games well

at all?

Belonging

to

the

power

elite

(high

socioeconomicstatus)

not

only

helps

a

young

man

go to collegeand get jobs through contacts his family has, italso gives him easy

access

as

a

child

to

the

credentialsthat

permit

him

to

get

into

certain

ys,

those

credentials

include

thewords

and

word-game

skills

used

in

Scholastic

AptitudeTests.

In

the

Middle

Ages

they

requiredknowledge

of

Latin

for

the

learned

professions

oflaw,

medicine,

and

theology.

Only

those

youngmen

who

could

read and write Latin could get intothose occupations, and if tests had been given

inLatin, I am sure they would have shown that professionalsscored higher in Latin

than

men

in

general,that

sons

who

grew

up

in

families

where

Latinwas

used

would

have

an

advantage

in

those

testscompared

to

those

in

poor

families

where

Latin

wasunknown,

and

that

these

men

were

more

likely

toget

into

the

professions.

But

would

we

concludewe were dealing with a general ability factor?Many a ghetto resident must

or should feel that heis in a similar position with regard to the kind ofEnglish

he must learn in order to do well on tests,in school, and in occupations today in

America. Iwas recently in Jamaica where all around me poorpeople were speaking an

English that was almostentirely incomprehensible to me. If I insisted, theywould

speak

patiently

in

a

way

that

I

could

understand,but

I

felt

like

a

slow-witted

child.

I havewondered how well I would do in Jamaican societyif this kind of English were

standard among therich and powerful (which, by the way, it is not),and therefore

required

by

them

for

admission

intotheir

better

schools

and

occupations

(as

determinedby a test administered perhaps by the JamaicanTesting Service). I would

feel

oppressed,

not

lessintelligent,

as

the

test

would

doubtless

decide

I

wasbecause

I was so slow of comprehension and soignorant of ordinary vocabulary.

When

Cronbach

(1970)

concluded

that

such

atest

giving

realistic

information

on

the presenceof a handicap,

recognize

that

it

is

those

in

powerin

a

society

who

often

decide

what

is

a

should be a lot more cautious about acceptingas ultimate criteria of ability the

standards imposedby whatever group happens to be in power.

Does this mean that intelligence tests are invalid?As so often when you examine a

questioncarefully

in

psychology,

the

answer

depends

onwhat

you

mean.

Valid

for

what?

Certainly theyare valid for predicting who will get ahead in anumber of prestige

jobs

where

credentials

are

is

white

skin:

it

too

is

a

valid

predictorof

job success in prestige jobs. But no one wouldargue that white skin per se is an

ability of the celebrated correlations between so- calledintelligence

test

scores

and

success

can

lay

nogreater

claim

to

representing

an

ability

factor.

Valid

for

predicting

success

in

school?Certainly,because

school

success

depends

on

taking

similartypes

of

tests.

Yet,

neither

the

tests

nor

schoolgrades

seem

to

have

much

power

to

predict

realcompetence

in

many

life

outcomes,

aside

from

theadvantages

that credentials convey on the individualsconcerned.

Are

there

no

studies

which

show

that

general

intelligencetest

scores

predict

competence

with

all

ofthese

other

factors

controlled?

I

can

only

assertthat

I

have

had

a

very

hard

time

finding

a

goodcarefully

controlled

study

of

the

problem

becausetesters simply have not worked very hard on it:they have believed so much

that

they

were

measuringtrue

competence

that

they

have

not

botheredto

try

to

prove

that

they

were.

Studies

do

exist,of

course,

which

show

significant

positive

correlationsbetween

special

test

scores

and

job-relatedskills.

For

example,

perceptual

speed

scores

arerelated

to

clerical

proficiency.

So

are

tests

ofvocabulary, immediate memory, substitution, andarithmetic. Motor ability test

scores

are

relatedto

proficiency

as

a

vehicle

operator

(Ghiselli,

1966).And

so

on.

Here we are on the safe and uncontroversialground of using tests as criterion

this

is

a

far

cry

from

inferring

that

there

is

ageneral

ability

factor

that

enables

a

person

to

bemore

competent

in

anything

he

tries.

The

evidencefor

this

general ability factor turns out to be contaminatedheavily by the power of those

at the topof the social hierarchy to insist that the skills theyhave are the ones

that indicate superior adaptivecapacity.

Where Do We Go from Here?

Criticisms

of

the

testing

movement

are

not

Social

Science

Research

Council

Committee

onEarly

Identification

of

Talent

made

some

of

thesesame

points

nearly

IS

years

ago

(McClelland,Baldwin,

Bronfenbrenner,

&Strodtbeck,

19S8).But

the

beliefs

on

which

the

movement

is

basedare

held

so

firmly

that

such

theoretical

or

empiricalobjections

have

had

little

impact

up

to

now.

Thetesting

movement

continues

to grow and extend intoevery corner of our society. It is unlikely that itcan be

simply stopped, although minority groupsmay have the political power to stop it.

For thetests are clearly discriminatory against those whohave not been exposed to

the

culture,

entrance

towhich

is

guarded

by

the

tests.

What

hopefully

canhappen

is

that

testers

will

recognize

what

is

goingon

and

attempt

to

redirect

their

energies

in

asounder

direction.

The

report

of

the

special

committeeon

testing

to

the

College

Entrance ExaminationBoard (1970) is an important sign that changesin thinking are

occurring

—

if only they can be implementedat a practical level. The report's gist

isthat

a

wider

array

of

talents

should

be

assessed

forcollege

entrance

and

reported

as

a

profile

to

thecolleges.

This

is

a

step

in

the

right

direction

ifeveryone

keeps

firmly

in

mind

that

the

criteria

forestablishing

the

of

these

new

measuresreally

ought

to

be

not

grades

in

school,

but

life

in

the

broadest

theoretical and practicalsense.

But

now

I

am

on

the

spot.

Having

criticizedwhat

the

testing

movement

has

been

doing,

I

feelsome

obligation

to

suggest

alternatives.

How

wouldI

do

things

differently

or

better? I do not mindmaking suggestions, but I am well aware that someof them are

as

open

to

criticism

on

other

groundsas

the

procedures

I

have

been

criticizing.

So

Imust offer them in a spirit of considerable humility,as approaches that at least

some

people

might

beinterested

in

pursuing

who

are

discouraged

withwhat

we

have

been

doing.

My

goal

is

to

brainstorma

bit

on

how

things

might

be

different,

notto

present

hard

evidence

that

my

proposals

arebetter

than

what

has

been

done

to

date.

Howwould

one

test

for

competence,

if

I

may

use

thatword

as

a

symbol

for

an

alternative

approach

totraditional intelligence testing?

1. The best testing is criterion sampling. Thepoint is so obvious that it would

scarcely be worthmentioning, if it had not been obscured so oftenby psychologists

like McNemar and Jensen whotout a general intelligence factor. If you want toknow

how well a person can drive a car (the criterion),sample his ability to do so by

giving hima driver's test. Do not give him a paper-and-penciltest for following

directions,

a

general

intelligencetest,

etc.

As

noted

above,

there

is

ample

evidencethat tests which sample job skills will predict proficiencyon the job.

Academic

skill

tests

are

successful

precisely

becausethey

involve

criterion

sampling

for

the

mostpart.

As

already

pointed

out,

the

Scholastic

AptitudeTest

taps

skills

that

the

teacher

is

lookingfor

and

will

give

high

grades

for.

No

one

couldobject

if

it

had

been

recognized

widely

that

thiswas

all

that

was

going

on

when

aptitude

tests

wereused

to

predict

who

would

do

well

in

e

started

only

when people assumed thatthese skills had some more general validity, as impliedin

the

use

of

words

like

intelligence.

Yet,even

a

little

criterion

analysis

would

show

thatthere are almost no occupations or life situationsthat require a person to do

word

analogies,

choosethe

most

correct

of

four

alternative

meanings

ofa

word,

etc.

Criterion

sampling

means

that

testers

have

got

toget

out

of

their

offices

where

they

play endless wordand paper-and-pencil games and into the field wherethey actually

analyze

performance

into

its

you

want

to

test

who

will

be

a

goodpoliceman,

go

find

out

what

a

policeman

does.

Followhim

around,

make

a

list

of

his activities, andsample from that list in screening applicants. Someof the job

sampling

will

have

to

be

based

on

theoryas

well

as

practice.

If

policemen

generally

discriminateagainst blacks, that is clearly not part ofthe criterion because the

law says that they mustnot. So include a test which shows the applicantdoes not