bandura, a. (1977) ‘self-efficacy: toward a unifying theory of behavioural change’,...

Upload: shanelee102

Post on 02-Jun-2018

251 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    1/25

    Psychological

    Review

    1977, Vol. 84, No. 2,

    191-215

    Self-efficacy:

    Toward a

    Unifying Theory

    of

    Behavioral Change

    AlbertBandura

    Stanford

    U niversity

    The p resen t a r t i c le p resen ts an in teg ra t ive theore t i ca l f r a m e w o r k to explain

    and t o predict psychological changes achieved by dif ferent m o d e s of t r e a t m en t .

    This

    theory s ta tes tha t psychologica l p rocedur es , wha tever thei r f o r m , a l te r

    the

    level and s t r en g t h of

    self -e f f icacy .

    It is hypothesized t h a t expec ta t ions of per -

    sona l efficacy determ ine whether coping behavio r

    will

    be i n i t i a t ed , how m u c h

    e f f o r t will be expended , and how long it will be sus tained in the

    face

    of ob-

    stacles and aversive experiences. Persistence in activities t ha t are subjectively

    t h r e a t en ing

    but in f a c t re la t ively sa fe p roduces , th rough exper iences o f m a s t e r y ,

    f u r t h e r enhancemen t o f

    self-efficacy

    and co r r espond ing r ed uc t i ons in defensive

    behavior .

    In the

    p roposed model , expec ta t ions

    of

    persona l

    efficacy

    a re

    derived

    f r om

    f o u r principal sources

    o f

    i n f o r m a t i o n : p e r f o r m a n c e a c c o m p l is h m en t s,

    vicar ious

    experience, verbal persuasion,

    and

    physiological states.

    T he

    m o r e

    d e-

    pendable

    the

    experiential sources ,

    the

    grea ter

    a re the

    changes

    in

    perceived self-

    efficacy. A

    n u m b e r

    o f

    f a c t o r s

    a re

    identified

    as

    influencing

    the

    cog nitive processing

    of efficacy i n fo rm a t i on a r is ing f r om enactive, vicar ious , exhor ta t ive, and emot ive

    sources . The

    dif feren t ia l

    power of d iverse therapeu t i c p rocedures is analyzed in

    t e r m s of the pos tu la ted cogni t ive mechanism of operation. Findings a re r epo r t ed

    f r o m

    mi c roana ly ses o f enactive, vicar ious , a nd emo t ive modes o f t r e a t m en t t h a t

    suppo r t

    the

    hypothesized rela t ionship between perceived self-efficacy

    and be-

    havioral changes. Possible di rections

    for

    fu r ther resea rch

    a r e

    d i scussed .

    Current developments in the field of be-

    havioral change reflect

    two ma jo r

    divergent

    t rends . The

    difference

    is especially evident in

    the

    t r e a tmen t

    of dys func t i ona l

    inhibi t ions

    and defensive behavior . On the one hand, the

    mechanisms

    by

    which human behavior

    is

    acqui red an d regula ted are increasingly fo r-

    mula ted

    in

    t e rms

    of cognitive processes. On

    the o the r hand , it is pe r formance-based pro-

    cedures

    that

    a re

    proving

    to be

    mos t

    powerful

    fo r

    effecting psychological changes.

    A s a

    con-

    sequence, successful pe r form ance is replac ing

    symbolically based experiences as the pr in-

    ciple vehicle

    of

    change.

    The

    present article presents

    the view

    that

    changes achieved by di f feren t me thods de r ive

    f rom a common cogni t ive mechanism. The

    T he

    research

    by t he

    a u t h o r

    repo r t e d i n

    this art ic le

    wa s

    suppor ted

    by

    Resea rch Gran t M-5162

    f r om the

    Nat iona l In s t i t u t e s

    o f

    Health,

    Un i te d S t a t e s

    Publ ic

    Health Service.

    Requests fo r reprin ts should be sent to Albe r t

    B a n d u r a , D e p a r t m e n t o f Psychology, S t an fo rd Un i -

    versi ty , S ta nfo rd , Ca l i fo rn ia 94305.

    apparent divergence

    of

    theory

    and

    practice

    can

    be reconci led by postula t ing

    t h a t

    cogni-

    tive

    processes media te change but

    t h a t

    cog-

    ni t ive events are induced and al te red most

    readi ly by

    experience

    of

    mas te ry a r i s ing

    f rom

    effective p e r f o r ma n c e . The dis t inc t ion be-

    tween process and means is underscored, be-

    cause

    it is of ten

    assumed

    that

    a

    cognitive

    m ode of opera t ion requi res a sym bolic m eans

    o f

    induc t ion . Psychological changes can be

    produced th rough o the r means than pe r -

    f o rmance accomplishments. Therefore , the

    explanatory mechanism developed

    in

    this

    article is designed to account for changes in

    behavior

    resulting from diverse modes of

    t r e a tmen t .

    Cognit ive Locus of Operat ion

    Psychological t rea tm ents based on learning

    principles were originally co nceptualized to

    opera te th rough peripheral mechanisms.

    New

    behavior

    was

    presuma bly shaped au tom at -

    ically by i ts

    effects.

    Contingency learning

    through pai red s t imula t ion

    was

    const rued

    in

    191

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    2/25

    192

    A L B E R T B A N D U R A

    connec t ion i s t t e rms

    a s a

    process

    in

    which

    re-

    sponses were l inked d i rec t ly

    to

    s t imul i .

    A l-

    te r ing the ra te of preexis t ing behavior by re -

    i n fo r cemen t

    was

    p o r t r a y e d

    as a

    process

    wherein

    responses were regula ted by their

    immedia te consequences wi thou t requ i r ing

    any conscious

    involvement

    of the responders.

    Growing evidence f rom several lines of re-

    search a l te red theore t ica l perspec t ives

    on how

    behavior is

    acquired

    and

    regulated. Theo-

    re t ica l f o rm ula t ion s emphas iz ing pe r iphe ra l

    mechan i sms began

    to

    give

    way to

    c ogni t ive ly

    oriented theories that expla ined behavior

    in

    t e r m s

    of

    centra l processing

    of

    di rec t , vicar i -

    ous , and sym bol ic sou rces o f in fo r m a t ion .

    Detailed analysis

    of the

    empi r i ca l

    and

    con-

    ceptual i ssues (see Bandura , 1977) fa l l s

    beyond

    the

    scope

    of the

    present a r t ic le .

    To

    s u m m a r i z e

    br ief ly ,

    however, i t has now been

    amply

    documented that cognitive

    processes

    play

    a

    p rom ine n t r o l e

    in the

    acquisi tion

    an d

    re ten t ion

    of new

    behavio r pa t te rn s . T rans i to ry

    experiences leave lasting effects by being

    coded

    and

    retained

    in

    sym bo l s

    fo r

    m e m o r y

    representa t ion. Because acquisi t ion of re-

    sponse

    i n f o r ma t i on

    is a

    m a jo r

    aspec t of learn-

    i ng , m uch hum an be hav io r

    is

    developed

    th rough

    m od e l ing . F rom obse rving o the rs ,

    one

    f o r m s

    a conception of how new behavior pa t -

    terns are

    p e r f o r m e d ,

    and on

    later occasions

    the

    symbol i c cons t ruc t ion se rves

    as a

    gu i de

    for a c t i o n ( B a n d u r a , 1 9 7 1 ) .

    T he

    ini tial

    a p -

    proximations of

    response

    patterns

    learned

    obse rva t iona l l y a re fur ther re f ined t h rough

    self -cor rec t ive a d j u s tm e n t s ba se d on

    i n f o r ma -

    t ive feedback f rom p e r f o r m a n c e .

    Lea rn ing

    f rom

    response consequences is

    also conceived

    of

    largely

    as a

    cognitive

    process . Consequences se rve

    as an

    u n a r t i c u -

    lated way o f

    in fo rming pe r f o rm e r s wha t t hey

    m us t do t o g a i n beneficial ou t com e s and t o

    avoid punishing ones. By observing the d i f -

    ferential

    effects of their own actions, in -

    dividuals d iscern which responses

    a re ap-

    p rop r i a t e

    in

    which se t t ings

    and

    behave

    a c -

    co rd ing ly (Dulany , 1968) .

    Viewed

    f rom the

    cognit ive f r a m ework , l e a rn ing

    f rom

    di ffer-

    en t i a l ou tcomes becomes

    a

    special case

    of

    observa t ional learning.

    In

    t h i s m ode

    of

    con-

    vey ing re sponse in fo r m a t ion ,

    the

    conception

    of

    the

    appropria te behavior

    is

    g radua l l y con-

    s t r u c t e d

    f rom

    observing the

    effects

    of one's

    ac t ions r a the r than

    f rom the

    examples pro-

    vided

    by

    others .

    Changes

    in

    behavior produced

    by

    s t imul i

    that either signify events to c o m e or ind ica te

    probable response consequences also have

    been shown to rely

    heavily

    oncognitive rep-

    resenta t ions of cont ingencies . People a re not

    m u c h affected

    by pa i red s t imula t ion un le ss

    they recognize that

    the

    events

    are

    corre la ted

    (Dawson

    & F u r e d y ,

    1976 ; Gr i ng s , 1973 ) .

    S t imul i in f luence the l ikelihood of a be-

    havior 's being p e r f o r m e d

    by

    v i r t ue

    o f

    their

    p red ic t ive func t ion ,

    no t

    because

    the

    s t imul i

    are au toma t i ca l l y connec ted to responses by

    the i r having occu r red toge the r . Re in te rp re ta -

    tion ofantecedent determinants aspredictive

    cues, ra ther than as control l ing s t imul i , has

    shif ted

    the

    locus

    of the

    regula t ion

    of be-

    havior f rom

    the stimulus to the individual.

    The

    issue

    of the

    locus

    at

    which behaviora l

    d e te rm inan t s ope ra te app lie s

    to

    re inforce-

    m e n t influences

    a s

    well

    as to

    an teceden t

    en-

    v ironm e n t a l st im u l i. C on t r a r y to t he c o m m o n

    view that behavior

    is

    control led

    by its im-

    media te consequences, behavior

    is

    related

    to

    its ou t com e s at the level of aggrega te con-

    se que nce s r a t he r t h an m om e n t a r y effects

    ( B a u m ,

    1 9 7 3 ) . People process

    a nd

    synthesize

    feedback information f rom

    sequences of

    events over long in te rvals about

    the situa-

    t i ona l c i r cums tances and the

    patterns

    and

    ratesofac t ions thatare necessary to p r o d u c e

    given outcomes. Since consequences a f fec t be-

    havio r th rough

    the

    influence

    o f

    t hough t ,

    be-

    liefs about schedules

    of

    r e i n f o r c e m e n t

    can

    exer t grea te r inf luence

    on

    behavior

    than

    the

    r e in fo rcemen t it s elf (Ba ron , Ka u f m an , &

    S t aube r ,

    1969 ;

    K a u f m a n , B a r o n,

    &

    Kopp ,

    1966).

    Incidence

    of behavior that has been

    positively re inforced does not increase if in-

    dividuals believe, based

    on

    o the r i n f o rm a t i on ,

    that

    the

    same actions will

    not be rewarded

    on

    f u t u r e

    occasions

    (Estes, 1 9 7 2 ) ; and t h e

    same consequences

    can

    increase , reduce ,

    or

    have no effect on incidence of behavior d e-

    pend ing

    on

    whe the r ind iv idua l s

    are led to

    believe

    that the

    consequences signify co r rec t

    responses, incorrect responses,

    or

    occu r non-

    con t ingen t l y (Dulany , 1968).

    The discussion thus far has examined the

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    3/25

    SELF-EFFICACY THEORY

    193

    role

    of

    cognition

    in the

    acquisition

    and

    regulation

    of

    behavior. Motiva t ion , which

    is

    primari ly concerned with act ivat ion and per-

    sistence

    of

    behavior,

    is

    also

    pa r t l y

    rooted

    in

    cognitive activities.

    The capacity to

    repre-

    sent

    f u t u r e

    consequences

    in

    thought provides

    one

    cognitively based source

    of

    motivat ion.

    Thr ough cogni t ive representat ion

    o ff u t u r e

    out-

    comes

    individuals can generate current mo-

    t ivators of behavior . Seen

    f rom

    this per-

    spective,

    re in fo rcement opera t ions

    affect

    be-

    havior largely by creat ing expectat ions that

    behaving in a certain way willp rod uce an t i c i-

    pated benefits or aver t future diff icul ties

    (Bolles, 1972b) .

    In the enhancemen t of

    previously learned behavior ,

    reinforcement is

    conceived

    of

    ma in l y

    as a

    mot iva t ional

    d e-

    vice

    ra ther than as an au tomat ic response

    s t rengthener.

    A second cognitively based source of mo-

    t ivat ion operates through the in terven ing in-

    fluences of goal setting and self-evaluative

    reac t ions (Bandura ,

    1976b, 1 9 7 7 ) .

    Self-

    motivation involves s tandards agains t which

    to

    evaluate pe r f o rmance .

    By

    mak ing self-

    reward ing

    react ions condi t ional

    o n

    a t t a in ing

    a

    certain level of behavior , individuals

    create

    self - inducements

    to persist in their

    effor ts

    unt i l their perfo rm anc es m atch sel f -prescribed

    s t anda r ds .

    Perceived negative discrepancies

    between pe r fo rm ance and s t anda rds

    create

    dissatisfactions t h a t motivate correct ive

    changes in behavior. Both the ant icipated

    satisfactions of desi red accomplishments and

    the negative

    appraisals

    of insufficient per-

    f o rmance thus provide incentives

    for

    act ion.

    Having accomplished

    a

    given level

    of

    per-

    fo rmance , individual s of ten are no longer

    satisfied with

    i t and m a k e

    fu r the r self-reward

    cont ingent

    on h igher a t t a inments .

    The reconceptual izat ion of hum an learn-

    ing and mot iva t ion in t e rm s of cognitive

    processes has

    ma jo r

    implications for the

    mechanisms through which therapeutic pro-

    cedures al ter behavioral

    funct ioning.

    A l-

    though

    the advances in cognitive psychology

    are a subject of increasing interest in specu-

    lations about behavioral change

    processes,

    few new theories of psychotherapy have been

    proposed

    t h a t

    might prove

    useful

    in s t imu-

    lat ing research on explanatory mechanisms

    P R S O

    - BEHAVIOR

    I

    I

    r EFFICACY ~1

    I

    EXPECTATIONS

    I

    I

    OUTCOME I

    I EXPECTATIONS

    -

    OUTCOME

    Figure

    1.

    Diagrammat ic representa t ion

    of the

    differ-

    ence

    between

    efficacy

    expectations

    and

    ou t c ome

    ex -

    pectations.

    and in in tegrat ing the resul ts accompanying

    diverse

    modes

    of

    t r ea tmen t .

    The

    present

    article outlines a theoretical

    f r amework ,

    in

    which

    the

    concept

    of

    s e l f - e f f i c a c y

    is

    assigned

    a cen t ra l role, for analyzing changes achieved

    in fear ful and avoidant behavior . The ex-

    plana to ry value

    of

    this conceptual system

    is

    then evaluated

    by i ts

    abili ty

    to

    p redic t

    behavio ra l changes p roduced th rough di f -

    ferent m e th o d s

    of

    t r e a tm e n t .

    E f f i c a c y

    Expectat ions

    a s a

    M e c h a n i s m

    of

    Operat ion

    The

    present theory

    is

    based

    on the

    prin-

    cipal asssumption t h a t psychological pro-

    cedures , whatever thei r

    f o r m ,

    serve as means

    of

    c rea t ing

    and

    s t rengthening expectat ions

    of personal

    efficacy.

    Within this analysis,

    efficacy expectations

    are

    distinguished f rom

    response-outcome expectancies. The differ-

    ence is presented schematically in Figure 1.

    A n outcome expectancy is defined as a

    person 's es t imate that

    a

    given behavior will

    lead to cer ta in ou tcomes . A n

    efficacy

    ex -

    pec ta t ion

    is the

    conviction

    that one can

    successfully

    execute

    the

    behavior required

    to

    produce the ou tcomes . Outcome and

    efficacy

    expectat ions are dif ferent ia ted, because in-

    dividuals can believe

    that

    a par t icular course

    of act ion will p roduce cer ta in ou tcomes ,

    bu t

    i f they enter tain serious d oubts ab out whether

    they

    can pe r f o rm the

    necessary activities

    such

    information does

    not

    influence

    their

    behavior.

    In this c onceptual sys tem, expectat ions of

    personal mastery

    affect

    both ini t ia t ion and

    persistence of coping behavior. The strength

    of

    people's

    convictions

    in

    their

    own effective-

    ness

    is

    likely

    to

    affect whether they will even

    t ry to

    cope with given situations.

    A t

    this

    initial level, perceived

    self-efficacy

    influences

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    4/25

    194

    AL BE R T B A N D U R A

    choice of

    behavioral settings. People fear

    and

    tend

    to avoid th reatening si tua t ions they

    believe

    exceed their coping skills, whereas

    they get involved in activi ties and behave

    assuredly when they judge themselves cap-

    able of handling si tua t ions

    t h a t

    would o the r -

    wise

    be

    i n t im i da t i ng .

    Not

    only

    can perceived self-efficacy have

    di rec t ive influence on choice of activities

    and

    se t t ings, but , th rough expecta t ions

    of

    eventual success,

    it can affect

    coping effor ts

    once they a re ini t ia ted.

    Efficacy

    expecta t ions

    de t e rm ine

    how

    m u c h effor t people will

    ex -

    pend and how long they will persist in the

    face

    of obstacles and aversive experiences.

    The stronger the perceived

    self-efficacy,

    the

    m o r e

    active the effor ts .

    Those

    who persist in

    subjectively

    threatening ac t ivi t ies t h a t a r e

    in

    fa c t

    relatively safe will gain corrective

    experiences

    t h a t re inforce their sense of

    efficacy,

    thereby eventually e liminat ing thei r

    defensive

    behav ior .

    Thosewho

    cease their cop-

    in g

    efforts p rema tu re ly will re tain thei r self-

    debi l i t a t ing

    expecta t ions

    and

    fears

    for a

    long

    t ime.

    The preceding

    analysis

    of how perceived

    self-efficacy inf luences pe r for m anc e is not

    me a n t to imply t h a t expectation is the sole

    d e t e r m i n a n t of behavior . Expecta t ion alone

    will no t p roduce de s i r e d pe r fo rmance if the

    component capabi l i t ies are lacking. Moreover ,

    there are many th ings tha t people can do

    with

    ce r t a i n t y o f success that they do no t

    pe r f o rm because they have no incentives

    to do so. Given appropria te ski l ls and ade-

    quate incent ives, however ,

    efficacy

    expecta-

    t ions are a ma jo r de t e rminan t of people's

    choice of activi ties, how m u c h

    e f f o r t

    they will

    expend, and of how

    long they will sustain

    effor t

    in

    dealing wi th s t ressful si tua t ions.

    D i m e n s i o n s

    of

    E f f i c a c y Expectat ions

    Empi r i c a l tests of the relationship between

    expec tancy

    and p e r f o r m a n c e of

    threatening

    act ivi t ies have been hampered by inadequacy

    o f

    the

    expectancy analysis.

    In

    most s tudies

    the measu re s of expecta t ions a re mainly con-

    cerned

    with people's hopes

    for

    f avorab le out -

    comes rather than with their sense of per-

    sonal mastery. Moreover, expectations a re

    usually assessed globally only a t a single

    point in a change process as though they

    represent a

    s t a t ic ,

    unid imensiona l f a c to r .

    Par t ic ipan ts

    in

    experiments

    of

    this t ype

    a re

    simply asked to judge how much they expect

    to benefit

    f rom

    a

    given procedure. When

    asked to m a k e such

    estimates, participants

    assume, moreof ten than not ,

    t h a t

    the

    benefits

    will

    be

    produced

    by the

    external minis t ra-

    t ions ra ther than gained th rough the de-

    velopment of

    self-efficacy.

    Such global mea-

    sures reflect am ix t u r eo f , among other th ings,

    hope, wishful th ink ing,

    belief in the

    potency

    o f the procedures , and f a i th in the the rapis t .

    I t

    the re fo re

    comes

    as no

    surpr i se

    t h a t

    out-

    come expecta t ions of this type have

    little

    relation

    to

    ma g n i t u d e

    of

    behavioral change

    (Davison & W ilson, 1973 , Lick & Bootzin,

    1975).

    Efficacy expecta t ions vary on several d i-

    mensions t h a t have

    i mp o r t a n t p e r f o r ma n c e

    implica t ions. They d i f f e r

    in

    m agni t ude . Thus

    when t asks a re ordered in level of

    difficulty ,

    the efficacy expectations of different individ-

    uals may be l imi ted to the simpler

    tasks ,

    ex -

    tend to modera te ly difficult ones, or include

    even the

    mos t t axing pe r formances . Efficacy

    expecta t ions also d i f f e r

    in

    g e n e r a l i t y . Some

    experiences

    create ci rcumscribed mastery ex-

    pectations. Others insti ll

    a

    more generalized

    sense of efficacy t h a t extends well beyond the

    specific t rea tment s i tua t ion . In addi t ion,

    expectancies vary in s tr e n g th .Weak expecta-

    t ions are easily extinguishable by discon-

    f i rming

    experiences,

    whereas individuals who

    possess strong expectations

    of

    mastery will

    persevere

    in

    their coping

    efforts

    despite dis-

    confi rming experiences.

    A n adequate expectancy analysis , there-

    fore , requ i res de tai led assessment

    of the

    magni tude , generali ty ,

    and

    s t rength

    of ef-

    f i cacy

    expec ta t ions commensura te wi th

    the

    precision with which behavioral processes are

    measured . Both efficacy expecta t ions

    and

    per-

    f o rm ance should be assessed at

    significant

    j unc t u r e s

    in the change process to clarify

    their

    reciprocal effects

    on

    each other. Mas te r y

    expecta t ions influence p e r f o r ma n c e and are ,

    in

    turn, a l te red by the cumula t ive effects of

    one's effor ts .

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    5/25

    SELF-EFFI C AC Y

    THEORY

    95

    S o u r c e s

    o f

    E f f i c a c y

    Expecta t ions

    In this social learning analysis, expecta-

    tions of personal

    efficacy

    a re based on

    four

    ma jo r

    sources of i n f o rm a t i o n : p e r fo rm an ce

    accom plishm ents , vicarious experience, verbal

    persuasion, and physiological states. Figure 2

    presents the diverse influence p rocedures

    commonly used to reduce defensive behavior

    and

    presents

    the

    principal source through

    which each t reatment operates to create

    expecta t ions

    of mas te ry . A ny given method,

    depending on how it is appl ied, may o f

    course

    draw to a lesser extent on one or more

    other sources of

    efficacy

    i n f o rm a t i o n .For ex-

    ample, as we shal l see shor t ly , performance-

    based t rea tmen ts no t on ly p romote be-

    havioral accomplishments but also extinguish

    fear arousal , thus authenticat ing

    self-efficacy

    t h rough

    enactive and arousal sources of in-

    fo rma t ion . Other methods , however, provide

    fewer ways of acqu i r i ng in fo rma t ion abou t

    one's capabili ty

    for

    coping with threatening

    s i tua t ions . By pos tu la t ing a com m on m echa-

    nism of operation, this analysis provides a

    conceptual

    f ramework

    within which

    to

    s t u dy

    behavioral changes achieved

    by

    different

    modes

    of t r ea tmen t .

    Per formance

    a c c o m p l i s h m e n t s .

    This source

    of efficacy i n f o rm a t i o nis especially

    inf luent ia l

    because i t is based on personal mastery ex-

    periences. Successes raise mastery expecta-

    t ions; repeated fai lures lower them , par-

    t icular ly

    if the mishaps occur ear ly in the

    course

    of

    events.

    A f t e r

    s t rong

    efficacy

    ex -

    pectat ions are developed through repeated

    success, the negative impact of occasional

    failures is l ikely to be reduced. Indeed, oc-

    casional failures t h a t are la ter overcome by

    de te rmined effort

    can s t rengthen sel f -moti -

    vated persistence if one finds th rough ex -

    per ience t h a t even

    the

    mos t difficult

    ob-

    stacles

    can be

    mas te red

    by

    sustained effor t .

    The effects of fa i lure on personal efficacy

    therefore

    par t ly depend on the t im ing and

    the

    total

    pattern of experiences in which the

    fai lures occur .

    Once established, enhanced

    self-efficacy

    tends

    to

    generalize to other si tuations in which

    pe r fo rmance was sel f -debi l i ta ted by pre-

    occupat ion

    with personal inadequacies (Ban-

    d u r a , A d a m s , & Beyer, in p ress ; Band ura ,

    Jeffery, & Gajdos,

    1975 ) .

    A s a result, im -

    provements in behavioral

    funct ioning

    t r an s f e r

    no t only

    to

    similar si tuations

    but to

    a ctivities

    t h a t

    a re

    substantially

    different

    f rom those

    on

    which the t r ea tmen t was focused .

    Thus,

    for

    example,

    increased

    self-efficacy

    gained through

    rapid m astery of a specif ic anim al phobia can

    EFFICACY

    EXPECTATIONS

    SOURCE

    PERFORMANCE ACCOMPLISHMENTS

    MODE OFINDUCTION

    ..PARTICIPANT

    MODELING

    - -PERFORMANCE DESENSITIZATION

    --PERFORMANCE EXPOSURE

    SELF-INSTRU CTED PERFORMANCE

    V I C A R I O U S EXPERIENCE

    VERBAL

    PERSUASION

    EMOTIONAL AROUSAL

    ,,.,-LIVE MODELING

    ----SYMBOLIC MODELING

    ..SUGGESTION

    ^--EXHORTATION

    ----SELF-INSTRUCTION

    -INTERPRETIVE TREATMENTS

    .-ATTRIBUTION

    - -R E L A X A T I O N ,

    BIOFEEDBACK

    '-SYMBOLIC DESENSITIZATION

    ^SYMBOLIC EXPOSURE

    Fig u re 2 . Majo r sou rce s o f efficacy in fo rma t ion and t he principal sources th rough which different

    m o d e s

    of

    t r e a t m e n t

    opera te.

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    6/25

    196

    A L B E R T B A N D U R A

    increase coping

    ef fo r t s in

    social si tuat ions

    as well as r e duce f e a r s of o the r animals .

    However,

    the

    generalization

    effects

    occur

    mos t p red ic t ab ly to the act ivit ies that are

    m o s t

    similar to those in which

    self-efficacy

    wasrestored

    by

    treatment

    (Bandura , B lanch -

    a r d , &Ri t t e r ,

    1 9 69 ) .

    M e t h o d s

    of change that opera t e on the

    basis ofp e r f o rm anc e a ccom pl ishmen t s convey

    efficacy

    i n f o r ma t i on

    in m ore ways than s im ply

    t h r o u g h

    the

    evidence

    of

    p e r f o r m a n c e

    im -

    provement s . In t he course of t rea tment s em-

    ploying m ode ling wi th guided pe r fo rm an ce ,

    par t ic ipants acquire a general izable ski l l for

    deal ing

    successful ly with

    s t ress ful

    s i t ua t ions ,

    a skillthatthey use to overcomea variety of

    dys func t iona l f ea r s and inhibi t ionsin their ev-

    e r y d a y

    life

    ( B a n d u r a e t

    al.,

    in

    press;

    B a n d u r a

    et

    al.,

    1975) . Having a serviceable coping

    skillat one ' s d i sposa l undoub ted ly con t r ibu tes

    to one's sense of pe rsona l efficacy. Behav iora l

    capabi l i t ies can also be enhanced th rough

    mode l ing

    a lone (Ban d u ra ,

    1971 ;

    F lande rs ,

    1968) .

    However , pa r t i c ipan t m od e l ing pro-

    v ides add i t iona l oppor tuni t ies fo r t rans l a t ing

    behavioral

    conceptions

    to appropriate

    actions

    and f o r m ak ing co r r e c t iv e r e f i nem en t s t owa rd

    the pe r fec t ion of

    skills.

    Most of the treatment

    procedures

    de-

    veloped in recen t yea rs to e l im ina t e

    fea r fu l

    and d efensive behavior have been im ple-

    m en t ed e it h er t h rough pe r f o r m ance o r by

    symbol ic p rocedures . Regard les s of the

    methods involved , resu l t s o f c o m p a r a t i v e

    s tudies attest to the super io r i t y o f

    p e r f o r m -

    ance-based t rea tment s .

    In the

    desensi t iza t ion

    approach dev ised by Wolpe ( 1 9 7 4 ) , clients

    receive g r adua t ed exposu re

    to

    aversive events

    in conjunction with anxiety reducing

    ac-

    t ivit ies, usually

    in the

    fo rm

    of

    m u s c u l a r

    re l axa t ion .

    A

    n u m b e r

    of

    exper iment s have

    been reported in which relaxation is paired

    with scenes in which

    phobics

    visualize

    themselves engaging in progress ive ly more

    threatening act ivi t ies

    o r

    wi t h enac t men t

    o f

    the

    same h ie ra rchy

    o f

    act ivi t ies with

    the

    actual th reats . Findings based

    on d i f f e r en t

    types of phobias consis tent ly reveal that

    p e r fo rma n c e

    desensi t iza t ion produces sub-

    s tant ia l ly greater behavioral change than does

    symbol i c

    desensi t iza t ion (LoPicol lo, 1970;

    S h e r m a n , 1 9 7 2 ;Strahley, 1966) . Physiolog-

    ical

    measures yield s imilar resul t s . Symbolic

    desensitization

    reduces autonomic responses

    t o ima gined

    but not to

    ac tua l t h rea t s , whereas

    p e r f o r m a n c e

    d esensit iza t ion el im inates auto-

    nomic responses to both imagined andactual

    t h rea t s (Ba r low , Lei t enberg , Ag ras ,&W incze,

    1969 ) . The subs t an t i a l

    benefits

    of

    successful

    p e r f o r m a n c e

    are typically achieved in less

    t i m e tha n i s requ i red to ex t inguish a rousa l t o

    symbol ic represen t a t ions

    o f

    th rea t s .

    More recent ly, avoidance behavior

    has

    been

    t rea t ed by p roc edur es involving m assive

    exposureto aversive events.In th i s approach ,

    intense anxie ty

    is

    elicited

    by

    pro longed

    ex-

    posure to the most threatening situations

    and

    sus t a ined

    a t

    high levels, without

    relief ,

    un t i l em ot iona l reac t ions a re ex t inguished .

    Several inves t igators have compared the

    re la t ive

    success of pro longed exposure to

    aversive

    s i t ua t ions in imagery and ac tua l en-

    coun te r s

    with them

    in

    amel io ra t ing ch ronic

    agoraphobias . Real encounters wi th threats

    produce resul t s dec ide ly super io r t o imagined

    exposure ,

    which ha s weak, variable effects

    ( E m m e l k a m p

    &

    Wessels,

    1975;

    S tern

    &

    M a r k s , 1973 ; Wat son , Mul le t t ,

    &

    Pillay,

    1 9 7 3 ) . P ro longed encoun t e r sthat ensu re

    be-

    havioral improvements

    are

    more effect ive

    t h an d i s t r i bu t ed brief encoun t e r s that a r e

    likely to end be fo re succes sfu l pe r f o rm anc e

    of the act ivi ty is achieved (Rab avilas ,

    Boulougour i s ,

    &S t e f an i s, 19 76 ) .

    The p a r t i c ipan t m ode l ing app roach to the

    el imina t ion of defens ive

    behavior ut il izes

    successful

    pe r f o r m anc e a s t h e p r ima ry v eh ic le

    of

    psychological change. People displaying

    in t r a c t ab l e

    fea r s

    and inh ib i t ions a re no t

    about

    to do

    wha t they

    dread. In

    implement-

    in g

    pa r t i c ipan t mode l ing , t he rapis t s

    the re fo re

    s t r u c t u r e the env i ronment sothat clients can

    pe r fo rm successfully

    despite

    their

    incapaci-

    ties.

    This

    is achieved by enl is t ing a va rie ty of

    response

    i nduc t i on a id s , i nc l ud ing p rel im ina ry

    mode l ing

    o f

    th rea tening ac t iv i t ies , g radua ted

    t a s k s , enac t men t ov e r g r adua t ed t empora l

    in terva l s , jo in t pe r fo r m an ce wi th the the ra -

    pis t , protect ive aids t o . r e d u c e the likelihood

    of

    fea red consequences , and varia t ion in the

    severity

    of the threat

    i t se l f (Bandura ,

    Je f f e ry ,

    & Wr igh t , 1 9 7 4) . A s t rea tment p rogresses ,

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    7/25

    SELF EFFICACY

    THEORY

    97

    the

    supplemen ta ry a id s

    are

    w i t h d r a w n

    so

    that

    clients cope

    effect ively

    unassis ted .

    Self-

    direc ted mastery exper iences

    are

    then

    ar-

    ranged

    to

    r e in fo r c e

    a

    sense

    of

    personal

    ef-

    f i cacy .

    Through th is

    fo rm

    o f

    t r e a t m e n t

    in -

    capaci ta ted people rapid ly lose their fears ,

    they

    are

    able

    to

    engage

    in

    activit ies they

    f o rme r l y

    inh ib i t ed , an d they d isplay genera l-

    ized

    r educ t ion s

    of

    fea r s toward th rea t s

    beyond

    the

    specif ica l ly t rea ted condi t ions

    ( B a n d u r a ,

    1 9 7 6 a ) .

    Par t ic ipan t m od el ing has been com pa red

    with var ious symbolica l ly based treatments.

    These s tud ies co r robora te

    the

    super io r i t y

    of

    successful

    pe r fo rm ance f a c i l it a t ed

    by

    model -

    ing

    as

    com pa red to vicarious experience alone

    ( B a n d u r a e t al., 1969; Blancha rd , 1970b;

    Lewis, 1974; Rit ter ,

    1969 ;

    Rope r , Rachman ,

    &

    M a r k s ,

    1 9 7 5 ) ,

    to symbolic desensi t iza t ion

    ( B a n d u r a

    et

    al., 1969; Li tvak , 1969) ,

    and

    to imaginal modeling in which c l ients visual-

    ize themselves or others coping successfully

    wi th th rea t s (Thase

    &

    Moss, 19 76 ) . When

    pa r t ic ipan t m od el ing i s subsequen tly ad -

    minis t e red

    to those who benefi t only par -

    t ially

    f rom the

    symbol ic p rocedures , avoid -

    ance behavior

    is

    tho roughly e l im ina ted wi th in

    a b r ie f per iod .

    The findings

    summa r i zed above

    are

    con-

    sis tent with

    self-efficacy

    t heo r y ,

    bu t

    they

    d o

    no t

    shed much l igh t

    on the

    mechan i sm

    by

    which

    specific m as te ry exper iences p rod uce

    general ized

    and end u r ing changes inbehavio r .

    Verif ica t ion

    of the opera t ive mechanism re-

    qui res exper im en ta l evidence

    that

    experienced

    mas te ry does in

    f a c t

    alter the level and

    s t reng th

    of

    self-efficacy

    and that

    self-efficacy

    is , in turn , l inked to behavior . We shal l re-

    tu rn l a te r

    to

    research

    that

    addresses itself

    specifically

    to the l inkages between treat-

    m ent proc edures , perceived

    self-efficacy, and

    behavior .

    V i c a r i o u s e x p e r i e n c e . People do no t rely on

    experienced mas t e r y

    as the

    sole source

    of in-

    f o rm a t i on

    con cerning their level

    of self-

    efficacy.

    Many expec ta t ions

    a re

    der ived

    f rom

    vica r ious exper ience. Seeing others per form

    t h r e a t en ing

    ac t ivi t ies without adverse con-

    sequences can genera te expec ta t ions in ob-

    servers that they too wil l improve if they

    in tensi fy and

    persist

    in

    their

    efforts. They

    persuade themselves that ifothers can do i t ,

    they should

    be

    able

    to

    achieve

    at

    leas t some

    imp rovemen t

    in

    p e r f o r m a n c e ( B a n d u r a

    &

    Barab, 1973) . Vicar ious exper ience, relying

    as i t does on inferences

    f rom

    social com-

    par ison, is a less dependable source of in-

    f o rm a t i on

    a b o u t one's capab ilit ies th an

    is

    direc t evidence of per sona l accompl ishmen ts .

    Consequent ly , the

    efficacy

    expectations in-

    duced by m od eling a lone a re l ikely to be

    weaker and more vu lnerab le to change.

    A

    n u m b e r of m od el ing va r iab les that a re

    apt to af fec t

    expectations

    of

    personal

    efficacy

    have been shown

    to

    enhance

    the dis inhibi t ing

    inf luence

    of

    model ing p rocedures .

    Phobics

    benefit

    m o r e

    f rom

    seeing

    models

    overcome

    their

    difficulties

    b y

    de t e rm ined

    ef fo r t

    than

    f rom observing fac i le per form anc es

    by adept

    mode l s ( K a z d i n , 1973; Meichenbaum , 1971) .

    Showing the gains achieved by

    ef fo r t fu l

    coping behavior no t on ly min imizes for ob-

    servers

    the

    negat ive impac t

    o f

    t empo r a r y

    distress

    but

    demons t r a t e s

    that

    even

    the most

    anxious

    can

    eventual ly succeed through per-

    severance. Sim ila r i ty

    to the

    m o d e l

    in

    other

    characterist ics, which increases

    the

    personal

    re levance

    of

    v ica r ious ly der ived in fo rm a t ion ,

    can likewise enhance the

    effectiveness

    of

    symbol ic

    mode l ing (Kazd in , 1 9 7 4 b ) .

    Modeled behavio r wi th

    clear

    ou t comes

    conveys more efficacy i n f o r m a t i o n t h a n

    if the

    effects

    of the

    modeled ac t ions remain

    am -

    biguous . In

    invest iga t ions

    o f

    vicar ious pro-

    cesses, observing

    one

    per fo rm ac t iv i t ies

    that

    meet with success does, indeed, produce

    grea te r behavio ra l improvements

    than

    wit-

    ness ing the sam e per fo rm ances mod eled wi th -

    ou t any

    evident consequences (Ka zd in, 1974c,

    1 9 7 5 ) .

    Diversif ied modeling,

    in

    which

    the

    activit ies observers regard

    as

    haza rdous

    are

    repea tedly shown to be safe by a var iety of

    models , is superior to exposure to the s ame

    pe r fo rmances by a s ing le m odel (B an d ura &

    M enlove, 1968; Kaz d in, 1974a, 1975, 19 76 ) .

    If people of widely di f fe r ing charac ter is t ics

    can

    succeed, then observers have

    a

    reason-

    able basis for increasing

    their

    own sense of

    self-efficacy.

    T he pa t t e r n o f resul ts repor ted above o f-

    fers at least suggestive support for the view

    that exem plifications of success thr oug h sus-

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    8/25

    198

    ALBERT

    B A N D U R A

    tained effor t with

    substantiating

    compa ra t iv e

    in fo rma t i on can enhance

    observers'

    percep-

    t ions o f their own pe r f o rmance c apab i l i t i e s .

    Research wil l

    be

    presented below

    that

    bea rs

    more d i rec t ly

    on the

    propos i t ion

    that

    m o d e l -

    in g procedures a l t e r avoidance behav ior

    t h r o u g h

    the

    in te rvening

    inf luence of

    efficacy

    expecta t ions .

    V e r b a l

    p e r s u a s i o n . In a t t e m p t s to influence

    hum an behav ior , ve rba l pe rsuas ion i s wide ly

    used

    because of i t s ease and ready avai l -

    abi l i ty . People

    a re

    led , t h rough sugges t ion ,

    in to believing they

    can

    cope successful ly with

    w h a t has overwhelmed them in the past.

    Efficacy

    expec t a t ions induced in t h i s manne r

    arealso likely to beweaker

    than

    those arising

    f r o m

    one's

    own accompl i shment s because

    they do no t p rov i de a n au then t i c exper ien t i a l

    base fo r t h e m . In the face o f dis t ress ing

    th rea t s and a long h is to ry of

    f a i lu re

    in coping

    wi th them, wha tever mas te ry expec t a t ions

    a re

    i nduced

    by

    sugges t ion

    c an be

    read i ly

    extinguished

    by

    disconfi rming experiences .

    Results of several l ines of resea rch attest

    t o t he l imi t a t ion of p rocedures

    that

    a t t e m p t

    to ins t i l l outcome expecta t ions in people

    s imply by t e l l ing them wha t to expect . In

    l abora to ry s tud ies , placebo condi t ions

    de-

    signed suggestively

    to

    ra i se expec t a t ions

    of

    imp rov emen t p roduce

    little

    change in re-

    f r a c t o r y behavior (Lick & Boo t z in , 1975 ;

    M o o r e , 1965 ; Paul , 19 66 ) . W hethe r t h i s i s

    due to the low c red ib i l i t y of the sugges t ions

    or to the

    weakness

    of the

    induced expec t a -

    t i ons

    c anno t be de t e rm ined

    f r om

    these s tud-

    ies, because

    the

    expecta t ions were

    no t

    measu red .

    Numerous exper iment s have been con-

    ducted in which phobics receive desensi t iza-

    t ion

    t r e a t men t w i t hou t

    any

    expec t ancy

    in -

    f o r m a t i o n

    o r with sugges t ions that it is either

    highly

    efficacious

    o r ineffective. The

    d i f fer -

    ent ia l ou t come expec t a t ions

    a re

    verbal ly

    in -

    duced

    pr io r

    to ,

    d u r i n g ,

    o r

    immed i a t e l y a f t e r

    t r e a t m e n t in the various s tudies .The findings

    general ly show

    that

    desens i t i za t ion reduces

    phobic behav ior ,

    bu t the

    ou t come expec t ancy

    manipula t ions have e i the r no

    effect

    or weak ,

    inconsis tent ones

    (Hewlett

    & Nawas , 1971;

    M c G l y n n &M app , 197 0 ; M cG lynn , M ea lie a ,

    Nawas , 1969 ; M cG lynn, Reynolds ,

    Linde r , 1971) .

    As in the placebo

    studies,

    it

    is difficult to

    make conclusive

    interpretations

    because

    the

    out come expec t a t ions induced

    suggestively

    are not

    measu red p r i o r

    to the

    assessment of behavior changes, if at al l .

    S imply

    i n f o r m ing

    pa r t i c ipan t s

    that

    they will

    or

    will no t benefit

    f r o m

    t r e a t men t doe s no t

    m e a n

    that

    they necessarily believe what they

    are told, especially when it con t rad ic t s t he i r

    other personal experiences . Moreover, in the

    s tudies jus t c i ted the verbal influence is

    a imed ma in l y a t ra i s ing ou t come expec t a t ions

    r a t he r t h an at enhancing

    self-efficacy.

    It is

    changes on the l a t t e r d imens ion that are

    mos t re levan t

    to the

    t h eo ry unde r d i s -

    cussion.

    A l though social persuasion alone m ay have

    def ini te l imi t a t ions as a means of c rea t ing an

    e ndu r i ng sense of pe rsona l efficacy, i t can

    con t r i bu t e to the successes achieved through

    co r rec t ive p e r f o r m a n c e . That

    is ,

    people

    who

    are

    social ly persuaded that they possess the

    capabilities to mas t e r dif f icul t s i tuat ions and

    a r e provided with provis ional aids for e f -

    fec t ive act ion are l ikely to mobil ize greater

    effor t

    than those who receive only the per-

    f o r m a n c e aids . However,

    to

    raise

    by

    pe r -

    suas ion expecta t ions

    o f

    pe rsona l compe tence

    wi t hou t a r r ang ing cond i t i ons to f ac i l i t a t e ef -

    fective p e r f o r m a n c e will most l ikely lead to

    fa i lu res that d i s c r e d i t

    the

    pe rsuade rs

    a nd

    f u r t h e r

    u n d e r m i n e

    the

    recipients ' perceived

    self-efficacy.

    It is the re fo re the

    in te rac t ive ,

    as

    well as the i ndependen t , effects of social

    persuas ion

    on

    self-efficacy that mer i t exper i -

    menta l cons ide ra t ion .

    E m o t i o n a l a r o u s a l . S t ress fu l

    and

    t axing

    si tuat ions general ly e l ici t emot ional arousal

    that, depend ing on the c i rcums t ances , migh t

    have informa t ive va lue conce rning pe rsona l

    compe tency . There fore , emot iona l a rousa l i s

    ano the r cons t i t uen t source o f i n f o r m a t i o n

    that can a f fec t perceived self-efficacy in cop-

    in g with th reatening s i tuat ions . People re ly

    pa r t l y on the i r

    state

    of physiological arousal

    in

    jud ging the ir anx iety a nd vuln e rab i l it y t o

    s t ress . Because high arousal usual ly debi l i -

    tates

    pe r f o rm anc e , i nd iv idu a l s

    a r e

    more l ikely

    to

    expect success when they

    a re no t

    beset

    by

    avers ive arousal than

    if

    they

    a re

    tense

    and visceral ly agi ta ted. Fear react ions gen-

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    9/25

    SELF-EFFICACY THEORY

    199

    era te f u r t he r fear

    of

    impend ing

    st ressful

    si tua t ions th rough ant icipatory self-arousal .

    B y conjur ing up fea r -p rovok ing though ts

    about thei r inept i tude, individuals

    can

    rouse

    themselves to elevated levels of anxiety

    t h a t

    far

    exceed the

    fear

    experienced during the

    actual th reatening si tua t ion.

    A s

    will

    be

    recalled f rom

    the

    earlier dis-

    cussion, desensitization and massive exposure

    t rea tments a imed

    a t

    ext inguishing anx ie ty

    arousal produce some reduct ions in avoidance

    behavior . Anxiety arousal

    to

    t h r e a t s

    is

    like-

    wise diminished by modeling, and is even

    more thorough ly e l imina ted by exper ienced

    mastery achieved through par t ic ipant model-

    in g ( B a n d u r a & Ba r ab , 1973; B a n d u r a et

    al.,

    1969; Blanchard , 1970a) . Mode l ing

    ap-

    proaches have other advantages for enhancing

    self-efficacy

    and

    thereby removing

    dys func-

    t ional fears . In a dd i t ion to d im inish ing prone-

    ness to aversive arousal, such approaches also

    teach effective coping ski l ls by demonst ra t ing

    proficient

    ways

    of

    hand ling th r eatening situa-

    tions.

    The

    la t te r con t r ibu t ion

    is

    especially

    im -

    por tan t when fear arousal par t ly resul ts f rom

    behavioral defici ts. It is of ten the case

    that

    fears and defici ts a re interdepend ent. Avoid-

    ance

    of st ressful

    activities impedes develop-

    m e n t

    of

    coping skills,

    and the

    resul t ing lack

    o f compe tency prov ides a realistic basis for

    fea r .

    Acquir ing behavioral means

    for

    con-

    t rol l ing potent ia l th reats a t tenuates or e l im-

    ina tes

    fea r

    arousal (Averil l , 1973; Not te r -

    m a n , Schoenfeld, & Bersh , 1952; Szpiler &

    Epstein, 19 76 ). Behavioral cont rol n ot only

    allows

    one to

    manage

    the

    aversive aspects

    of

    an envi ronment . It also

    affects

    how the en-

    vironment is likely to be perceived. Poten-

    t ially

    st ressful si tua t ions

    t h a t

    can be con-

    trolled a re cons t rued as less threatening,

    and such cogni t ive

    appraisals f u r t he r

    re-

    duce ant icipatory emotional arousal (Averi l l ,

    1973) .

    Diminishing emotional arousal can reduce

    avoidance behavior, but

    di f feren t

    theories

    posit

    different

    explana tory mechanisms for

    the observed effects. In the theory f rom

    which

    the emotive

    treatments

    a re der ived,

    emotional arousal is conceived of as a dr ive

    t h a t

    ac tivates

    avoidance behavior.

    This

    view

    stresses

    the

    energizing

    funct ion of

    arousal

    and the

    re inforc ing func t ion

    of

    a rousa l

    re-

    duc t ion . Social learning theory,

    on the

    other

    hand , emphasizes the i n fo rma t ive func t ion

    of physiological arousal . Simply acknowledg-

    ing

    t h a t

    a rousa l i s bo th inform a t ive and m o-

    tivating by no means resolves the issue in

    dispute, because these

    are not

    necessari ly

    two

    separa te effects t h a t somehow join t ly p roduce

    behavior .

    Ra the r ,

    the cognitive

    appra isa l

    of

    arousal to a large extent de termines the level

    and

    di rec t ion

    of

    mot iva t iona l inducements

    to ac t ion. Cer tain cogni t ive

    appraisals

    of

    one's physiological s t a t e migh t be energizing,

    whereas other appraisals of the same state

    migh t no t (Weiner ,

    1 9 7 2) .

    Moreover , many

    f o rms

    of physiological arousal a re generated

    cognitively by arousing t ra ins of though t .

    When m ot iva t ion isconceptualized in t e rmso f

    cognitive

    processes (Ba nd ur a , 1977 ; Weiner,

    1972) ,

    the

    i n f o r ma t io n a l

    a nd

    motivat ional

    ef -

    fec ts of arousal are t rea ted a s interd ependent

    ra the r than as separa te events. W e shall re-

    turn to this issue later when we consider the

    di f feren t ia l pred ic t ions made f rom social learn-

    in g

    theory

    and f rom the

    dual-process theory

    o f

    avoidance behavior concerning

    the

    behavioral

    effects

    of extinguishing

    anxiety

    arousal .

    Researchers work ing wi thin the a t t r ibu-

    t ional

    f ramework

    have a t tem pted to m od i f y

    avoidance behavior by di rec t ly manipula t ing

    the cognitive labeling of emotional arousal

    (Valins &

    Nisbet t ,

    1 9 7 1 ) . The

    p re sump t i on

    is t h a t

    if

    phobics

    are led to

    bel ieve that

    the

    things they have previously feared no longer

    affect

    them in te rna l ly ,

    the

    cognitive reevalua-

    tion alone

    will

    reduce avoidance behavior . In

    t rea tment ana logues

    of

    this approach, pho-

    bics receive false physiological feedback sug-

    gesting t h a t they are no longer emotionally

    upset by threatening events. Results of this

    p rocedu re a re essentially negative.

    Ear ly

    cla ims that er roneous arousal feedback re-

    duces avoidance behavior (Valins & Ray,

    1967) are disputed by methodological ly

    super ior studies showing that false feedback

    o f

    physiological t ranqui l i ty in the presence

    of

    th r e a t s

    has

    ei ther

    no

    appreciable

    effect

    on

    subsequent

    fea r fu l

    behavior (Gaupp,

    Ste rn ,

    &

    Galb ra i th , 197 2;

    Hewlet t &

    Nawas,

    1971 ; Kent , Wilson, &Nelson, 19 72 ; Rosen,

    Rosen,

    &

    Reid , 1972; Sush insky

    &

    Bootz in ,

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    10/25

    20 0

    ALB ERT B A N DUR A

    1 9 7 0 ) ,

    o r p roduces minor changes unde r

    such l imi ted condi t ions as to be of little

    prac t ica l consequence (Borkovec, 1 9 7 3 ) .

    Misa t t r ibu t ion of emot iona l a rousa l is

    ano the r va r i an t o f the a t t r ibu t iona l ap -

    p roach

    to

    mod i f i ca t ion

    o f f e a r f u l

    behavior .

    T he s t ra tegy here is to lead

    f e a r f u l

    people

    into believing

    that

    thei r emotional a rousal

    is

    caused by a nonemotional source. To the

    ex ten t that they no longer label their

    agitated

    state as

    anxiety, they will behave

    more bo ld ly . It may be possible to r e d u c e

    mild fears

    by

    this means (Ross,

    Rodin, &

    Z i m b a r d o ,

    1 9 6 9 ) , but the

    highly anxious

    are

    not easily led into misattributing

    their

    anxiety

    to i r re levant sources (Nisbe t t & Schachte r ,

    1966). When evaluated systematically, mis-

    a t t r i bu t i on t r e a tm e n t s

    d o n o t

    produce sig-

    nificant

    changes

    in chronic

    anxiety

    condi-

    t ions (S inge rman , Borkovec , &B a r o n , 1 9 7 6 ) ,

    and

    some

    of the

    benefi ts reported with other

    dy s func t i on s

    canno t be rep li ca ted (Boo tz in ,

    H e r m a n , &

    Nicassio,

    1976; Kellogg &Ba ron ,

    1 9 7 S ) . There is also some suggestive evidence

    that

    in

    labo ra to ry s tud ies

    the

    a t tenua t ion

    of

    f e a r

    ma y be du e

    m o r e

    to the

    ve r id i ca l i ty

    of

    a rousa l in fo rma t ion than to misa t t r ibu t ion o f

    f e a r

    a rousa l to an innocuous sou rce (Ca lve r t -

    Boyanowsky

    &

    Leventhal , 1 9 7 5) .

    A ny

    reduc t ion

    in

    fear resul t ing

    f r om d e-

    ceptive feedback

    is apt to be

    short- l ived

    because i l lusory assurances

    are not an

    espe-

    cially reliable way o f c rea t ing du rab le self-

    expecta t ions. However , more veri table ex-

    periences that reduce the level of emotional

    a rousa l

    can set in

    mot ion

    a

    reciprocal process

    of change. In the social learning view, po-

    ten t i a l th rea t s ac t iva te fear la rge ly through

    cogni tive se l f - a r ou sa l (Bandu r a , 1969 , 1 9 7 7 ) .

    Perceived self-competence

    c an

    the re fo re

    a f-

    fec t susceptibil i ty to

    self -a rousa l .

    I n d i vi d u a l s

    who com e

    to

    believe that

    they arelessvulner-

    able than they previously assumed

    are

    less

    prone to generate f rightening thoughts in

    t h r e a t en i n g

    s i tua t ions .

    Those

    whose f ea r s

    are relatively weak may reduce their self-

    d o u b t s a nd debi l i t a t ing se l f - a rousa l to the

    point where they perfor m successfully . Per-

    f o r m a n c e successes, in tu rn , s t reng then self-

    efficacy.

    Such

    changes

    can, of course, be

    re l iably achieved wi thout resor t

    to

    ruses.

    Moreove r , m is l abe l ing a rousa l o r a t t r i bu t ing

    it to e r roneous sources is unlikely to be of

    m u c h

    help

    to the

    highly anxious. Severe

    ac rophobics ,

    fo r example , m a y b e t e m po r a r i l y

    misled into believing

    that

    they

    no

    longer

    f ea r high elevations,

    bu t

    they w ill reexp eri-

    ence

    unne rv ing in te rna l f eedback when con-

    f ron t ed wi th d readed he igh ts . I t should also

    be no ted that in a t t r ibu t iona l exp lana t ions

    of the success of behaviora l t re a tm en ts the

    heavy emphasis

    on

    physiological ar ousa l

    der ives

    m o r e f r om specu l a t ions abou t th e

    na tu r e

    o f

    emot ion (Schach te r , 1964) than

    f r om evidencethat a rousa l i s a ma jo r de te r -

    minant of defensive behavior.

    C o g n i t i v e P r o c e s s i n g

    oj

    Efficacy Informat ion

    T he

    discussion thus

    f a r h a s

    cente red pr i -

    ma r i l y on the many sou rces o f in fo rma t ion

    enact ive , vicar ious, exhor ta t ive ,

    and

    emot ive

    that people use to judge their level of

    self-efficacy.

    A t th i s po in t a d i s t inc t ion mus t

    be d rawn be tween i n f o r ma t i on conta ined in

    envi ronmen ta l even ts and

    i n f o r ma t i on

    a s

    processed and t r a n s f o r m e d by the ind iv idua l .

    The

    i m p a c t

    of i n f o rm a t i o n on

    efficacy

    ex -

    pecta t ions

    will

    depend

    on how it is

    cog nitively

    appraised . A n u m b e r of con tex tua l f ac to r s ,

    inc luding the social , si tuational , and t e m po r a l

    c i r cum s tances und e r which even ts occu r , en ter

    in to such appraisa ls .

    Fo r

    this reason, even

    success experiences

    do no t

    necessari ly create

    s t rong genera l ized expecta t ions of pe rsona l

    efficacy. Expectations

    that have

    served self-

    protec t ive

    func t ions

    fo r yea rs a re no t qu ick ly

    discarded.

    When

    experience

    contradicts

    firmly

    establ ished expecta t ions of self-efficacy, they

    may unde rgo little change if the

    conditions

    of p e r f o r m a n c e are such as to lead one to

    discount

    the

    impor t

    of the

    experience.

    The co r r ec t ive va lue of i n f o rm a t i on de r ived

    f rom successful performance

    can be at-

    t enua ted

    in

    several ways.

    The f i rs t

    involves

    discr imina t ion processes. The consequences

    i nd iv idua ls an t i c ipa te we re they to

    pe r f o r m

    feared activities

    differ

    in circum stances which

    va ry i n s a f e gua r d s . A s a r e su l t , t hey m ay

    behave boldly in situations signi fying safe ty ,

    bu t r e t a i n unchang e d t he ir s e lf - doub t s und e r

    less secure conditions. Such mitigative dis-

    c r imina t ions

    can

    extend

    to the

    t r e a tm e n t s

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    11/25

    SELF-EFFICACY THEORY

    201

    themselves, as well as to the si tuational cir-

    cumstances in which behavioral a t tainments

    occur . This

    is

    especially true

    of

    t r ea tmen t s

    relying

    solely

    on

    symbolic

    and

    vicarious

    ex -

    perience. Achieving reductions

    in

    fear

    to

    threat s presented sym bolical ly is unl ikely to

    enhance perceived self-efficacy to any great

    extent in people who believe t h a t success in

    imagery does not por tend accomplishments

    in rea l i ty . In fo rmat ion conveyed by facilely

    modeled performances might l ikewise be

    minimized

    by

    anxious observers

    on the

    g rounds that

    the models possess special

    expertise enabling them

    to

    prevent injurious

    consequences

    that

    might o therwise befal l the

    unski l led. Because such discriminations, even

    though object ively mis taken, impede change

    in

    self-efficacy,

    observers will be re luc tan t to

    a t t e m p t feared activities

    and will be

    easily

    dissuaded

    by

    nega tive experience.

    Cognitive appraisals of the causes of one's

    behavior , which have been examined exten-

    sively

    in

    investigations

    of

    self-a t t r ibut ional

    processes (Bern, 1 9 7 2) , can s imilar ly del imi t

    gains in self-efficacy from behavioral a t tain-

    men t s . I t was previously shown that a t-

    t r ibu t ions of

    affect

    and actions to illusory

    competence have li t t le,

    if

    any , effect

    on

    r e f r a c t o ry behavior . This does not , of course,

    mean

    t h a t causal appraisals

    are of

    l imi ted

    impor t ance in theprocess ofbehavior change.

    Quite the contrary, per fo rmance a t t a inmen t

    is a p rominen t source of efficacyi n f o rm a t i o n,

    but i t is by no

    means unambiguous .

    As a l -

    ready men t ioned

    briefly,

    people

    can

    gain com-

    petence th rough au then t ic means bu t , be-

    cause of faul ty appraisals of the c i rcum-

    s tances und er which they im prove, wil l cr edi t

    thei r achievements to external factors ra ther

    than to thei r own capabi l i t ies . Here the

    problem is one of inaccurate ascrip t ion of

    personal competency

    to

    si tuat ional factors .

    Successes are m or e likely to enha nce self-

    efficacy

    if pe r fo rmances a re perceived as

    resul t ing

    f rom

    skil l than

    f rom

    f o r t u i t o u s

    or

    special external aids. Conversely,

    failures

    would be expected to p roduce g rea te r reduc-

    t ions

    in

    self-efficacy

    when a t t r ibu ted to

    abi l i ty ra ther than to unusual s i tuat ional

    c i rcumstances .

    The more extensive the si tua-

    t ional aids for p e r f o rm an ce , the greater a re

    the chances

    that

    behavior will be ascribed

    to

    external factors (Bern, 1972; Weiner ,

    1972) .

    Even under condi t ions of perceived self-

    de t e rmina t i on of

    outcomes,

    the

    impac t

    of

    pe r fo rmance a t t a inments on self-efficacy will

    vary depending on whether one's accomplish-

    m e n t s

    a re

    ascribed mainly

    to

    abili ty

    or to

    effort .

    Success with minimal

    effort

    fosters

    abi l i ty ascrip t ions t h a t re in fo rce

    a

    strong

    sense

    of

    self-efficacy.

    By

    contrast, analogous

    successes achieved through high expenditure

    of effor t connote

    a

    lesser ability

    and a r e

    thus

    likely

    to have a weaker effect on perceived

    self-efficacy.

    Co gnitive appra isals

    of the

    d i f -

    ficulty level of thetaskswillfu r the r

    affect

    the

    impac t

    of

    per fo r m anc e accom plishments

    on

    perceived

    self-efficacy.

    To succeed a t easy

    tasks provides

    no new

    in fo rma t ion

    for alter-

    ing one's sense of

    self-efficacy,

    whereas mas-

    tery

    of

    challenging

    tasks

    conveys salient evi-

    dence of enhanced competence. The

    rate

    and

    pa t t e rn

    of a t t a i nmen t s furnish addi t ional in -

    format ion forjud ging personal

    efficacy. Thus,

    people who

    experience setbacks

    bu t

    detect

    relative progress will raise their perceived

    efficacy

    more than those

    who

    succeed

    but see

    thei r performances level ing off compared to

    their prior r a t e

    of

    improvement .

    Extrapolat ions from theories about a t t r ibu-

    tion

    and

    self-perception

    to the field of be-

    havioral change often imply that

    people must

    l abor unaided

    or

    under inconspicuously

    a r -

    ranged

    inf luences

    if

    they

    are to

    convince

    themselves of their personal competence

    (Kopel &

    Arkowi tz ,

    1975) .

    Such

    prescrip-

    t ions are open to question on both con-

    ceptual and empirical grounds. Cognitive

    misappraisals

    that

    a t tenua te

    the

    impac t

    of

    disconfi rming experiences

    can be

    minimized

    without sacrificing the substantial benefits

    of powerful induction procedures . This is

    achieved

    by

    providing opportunities

    for

    self-

    di rec ted accomplishments after the desired

    behavior has been established. Any lingering

    doubts people might have, either about their

    capabili ties or about probable response con-

    sequences

    under unprotected condi t ions ,

    a re

    dispelled easily in this manner (Bandura

    et al., 197S ) . The more varied the c i rcum-

    stances

    in

    which th rea t s

    a re

    mas te red

    in -

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    12/25

    202

    A L B E R T B A N D U R A

    dependent ly ,

    the

    more likely

    a re

    success

    ex -

    periences to authenticate personal efficacy

    and to impede fo rma t i on of d i sc r imina t ions

    t h a t

    insulate self-perceptions f rom disconf i rm-

    in g evidence.

    Results of recent s tudies suppor t the

    thesis t h a t generalized, lasting changes in

    self-efficacy and

    behavior

    can

    best

    be

    achieved

    by

    par t ic ipant methods using power-

    f u l induct ion procedures ini t ia l ly

    to

    develop

    capabili t ies, then removing external aids

    to

    ver i fy personal efficacy, then

    finally

    us ing

    self-directed

    mas te ry to s t rengthen and gen-

    eralize expectations of personal

    efficacy

    ( B a n d u r a et

    al.,

    1 9 7 5 ) .

    Independent pe r -

    fo rmance can enhance

    efficacy

    expectations

    in

    several ways: (a) I t creates addi t ional

    exposure

    to fo rme r threa ts , which provides

    pa r t i c ipan ts wi th f u r t he r evidence t h a t they

    are no

    longer aversively aroused

    by

    wha t

    they previously fea red . Reduced emot iona l

    a rousa l conf i rms increased copingcapabili ties,

    ( b ) Se l f -d irec ted mas te ry prov ides oppor tuni -

    ties to perfect coping skills, which lessen

    personal vulnerability

    to

    stress,

    (c )

    Indepen-

    den t pe r fo rmance ,

    if

    well executed, produces

    success experiences, which f u r t he r re inforce

    expectations of

    self -competency.

    Extensive self-di rec ted performance of

    f o rm e r l y threatening ac t ivi t ies under pro-

    gressively challenging conditions a t a t ime

    when t rea tments a re usua l ly te rmina ted could

    also

    serve

    to reduce susceptibility to relearn-

    ing of defensive pat te rns of behavior . A few

    negat ive encounters among many successful

    experiences

    t h a t

    have

    instilled

    a

    strong sense

    o f self-efficacy

    will,

    a t

    most, establish dis-

    c r imina t ive

    avoidance

    of

    realistic

    threa ts , an

    effect t h a t has adapt ive value. In cont rast , i f

    people have limited contact with previously

    feared

    objects

    a f te r

    t rea tment , wha tever

    expectations of self-efficacy were insta ted

    would

    be

    weaker

    and

    more vulnerable

    to

    change. Consequent ly,

    a few

    unfavo r ab le

    experiences

    are likely to reestablish defensive

    behavior

    t h a t

    generalizes inappropriately.

    We have already examined how cogni t ive

    processing of information conveyed by mode l -

    ing might influence the extent to which

    vicarious experience effects changes

    in

    self-

    efficacy. A m o n g

    the

    especial ly informative

    elements are the models' characteristics (e.g.,

    adeptness, perseverance, age, exper tness) ,

    the

    similar i ty between models and observers, the

    difficulty

    of the

    p e r f o r ma n c e

    t asks , the

    si tua-

    t ional a r r angements unde r which the mode led

    achievements occur , and the diversi ty of

    mode led a t t a inments .

    Ju s t

    as the value of

    efficacy

    i n fo rma t i on

    generated enactively and vicariously depends

    on cogni t ive appraisal , so does the informa-

    tion arising f rom exhor ta t ive

    and

    emotive

    sources. The impac t of verbal persuasion on

    self-efficacy

    may va ry subs tan t i a l ly depend-

    ing on the perceived credibi l i ty of the per-

    suaders, thei r prest ige, t rustwor thiness, ex-

    pertise,

    and assuredness. The

    more

    believable

    the source of the i n f o r ma t i o n , the m o r e

    l ikely are efficacy expecta t ions to change.

    The influence of credibi l i ty on a t t i t ud i na l

    change has,

    of

    course, received intensive

    s tudy .

    But its

    effects

    on

    perceived self-efficacy

    r ema in

    to be investigated.

    People judge their physiological arousal

    largely on the basis of their appra isa l of the

    instigating conditions. Thus , visceral a rousa l

    occur r ing

    in si tuations perceived to be

    threatening is interpre ted as fea r , arousal in

    thwar t ing si tua t ions is experienced as anger ,

    and t h a t resul t ing f rom irretrievable loss of

    valued objects as sorrow (Hunt , Cole , &

    Reis, 1958). Even the same source of phys-

    iologicalarousal m a y b ein te rpre ted

    di f feren t ly

    in ambiguous situations depending on the

    emotional react ions of others in the same

    se t t ing (Mandle r ,

    1975;

    Schachter & Singer,

    1 9 6 2 ) .

    When t asks a r e pe r f o rm e d in ambiguous or

    complex s i tua t ions in which there is a variety

    o f evocat ive s t imuli ,

    the

    informa t iona l va lue

    of the resul tant a rousal wil l depend on the

    meaning imposed upon

    it .

    People

    who

    per-

    ceive their arousal

    as

    s t emming f rom personal

    inadequacies are more likely to lower their

    efficacy expecta t ions than those who a t t r i b u t e

    thei r a rousal

    to

    cer ta in si tua t ional fac tors .

    Given

    a

    proneness

    to

    ascr ibe arousal

    to

    per-

    sonal deficiencies, the heightened a t tent ion to

    internal events

    c an

    result

    in

    reciprocally

    escalating arousal. Indeed,

    as

    Sa rason (1976)

    has

    amply

    documented, individuals who are

    especially susceptible to anxie ty arousal

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    13/25

    SELF-EFFICACY THEORY

    203

    readily become self-preoccupied with their

    perceived inadequacies

    in the

    face

    of

    d i f -

    ficulties ra ther than wi th the task a t h a n d .

    Differing

    P e rs p e c t i v e s

    on

    S e l f - e f f i c a c y

    The phenomena encompassed by the con-

    struct of

    self-efficacy

    have been the

    subject

    o f interest

    in

    other theories

    of

    h u m a n

    be-

    havior. The theoretical perspectives

    differ,

    however,

    in how they view the nature and

    origins of

    personal

    efficacy and the interven-

    ing

    processes

    by

    which perceived self-efficacy

    affects behavior . In seeking a motivat ional

    explanat ion of explora to ry and manipula t ive

    behavior, White

    (1959)

    postulated an effec-

    t ance

    motive,

    which is conceptualized as an

    int r insic

    dr ive

    fo r

    t ransact ions wi th

    the en-

    vi ronment .

    Unlike inst igators ar ising f rom

    t issue defici ts, effectance motivation is be-

    lieved to be

    aroused

    by

    novel s t imula t ion

    and i s

    sus ta ined when

    the

    resul tant inquisi -

    tive

    and

    explora to ry ac t ions produce

    f u r t he r

    elements of

    novelty

    in the

    s t imu lus

    field.

    The

    effectance

    m otive presum ably develops

    th rough

    cumula t ive acquis i t ion

    of

    knowledge

    and ski l ls in dealing wi th the envi ronment .

    However, the process by which an effectance

    mot ive emerges

    from effective

    t r ansac t ions

    with

    the envi ron m ent is not spelled out in

    White's theory. Nor is the existence of the

    motive easy to veri fy , because effectance

    mot iva t ion

    is infe r red

    f rom

    the

    explora to ry

    behavior it supposedly causes. Without an

    independent measure

    of

    motive

    st rength one

    can not te ll whether people explore and m anip-

    ula te th ings because of a competence motive

    to do so, or for any

    n u m b e r

    of

    other reasons.

    Athough the

    theory

    of effectance motivation

    has not been fo rmula t ed in sufficient detail

    to permi t extensive theore t ical comparisons,

    there are several issues on which the social

    learning and effectance theories clearly differ.

    In the social learning analysis, choice be-

    havior

    and

    effort

    expend i t u r e

    a re

    governed

    in p a r t by percepts of

    self-efficacy

    ra the r

    than

    by a drive condition. Because efficacy

    expecta t ions

    a re

    defined

    and measured in-

    dependent ly of pe r fo rm ance , t h ey p rov ide

    an

    explici t basis

    for

    pred ic t ing

    the

    occurrence,

    generali ty, and persistence of coping be-

    havior, whereas an om nibus m otive does

    not. People will approach, explore ,

    and t r y

    to

    deal wi th si tua t ions wi thin thei r

    self-

    perceived capabili t ies, but they will avoid

    t ransact ions wi th s t ressful aspects of their

    envi ronment

    they perceive as exceeding their

    abi l i ty .

    The alternative views also d i f f e r on the

    origins

    of efficacy. Wi th in the f r a me wo r kof

    effectance theory , the effectance dr ive d e-

    velops gradually th rough prolonged t rans-

    actions with one's sur roundings .

    This

    theory

    thus focuses almost exclusively

    on the effects

    produced by

    one's

    own

    ac t ions.

    In the

    social

    learning theory ,

    self-efficacy

    isc onceptualized

    as arising from d iverse sources of info rm a-

    tion conveyed by d i r e c t and me d i a t e d ex-

    perience.

    These

    differences in theore t ical ap-

    proach have

    significant

    implica t ions

    for how

    one goes

    abou t

    s tudying the role of perceived

    self-efficacy

    in

    mo t iv a t i ona l

    and

    behavioral

    processes. Expecta t ions

    of

    personal efficacy

    d o no t

    opera te

    as

    d i sposi t iona l de te rminants

    independently of contex tua l fac tors . Some

    si tuat ions requi re greater sk i l l and m o r e

    a r d u o u s p e r f o r m a n c e sand car ry higher r i sk

    of negat ive consequences than d o o the rs .

    Expecta t ions will vary accordingly.Thus, for

    example, the level and s t rength of perceived

    self-efficacy

    in

    public speaking will

    differ

    depend ing

    on t he sub je c t ma t t e r , t h e fo rma t

    of the presenta t ion, and the types of audi -

    ences t h a t

    will

    be

    add re s sed .

    The

    social

    lea rning

    approach i s

    the re fore

    based on a

    microanalysis of

    perceived coping capabili t ies

    ra the r than on g loba l pe rsona l i ty t r a i t s o r

    motives of

    effectance. From this perspect ive,

    it is no

    mo r e i n f o r ma t i v e

    to

    speak

    of

    self-

    efficacy

    in

    genera l te rms than

    to

    speak

    of

    nonspecific approach behavior .

    To

    e luc ida te

    how

    perceived self-efficacy affects behavior

    requi res a mic roana lys i s of bo th fac to r s .

    Discrepancies

    between

    efficacy

    expecta-

    t ions

    and pe r fo rm anc e a r e m os t lik e ly t o

    a r i se unde r condi t ions

    in

    which si tua t ional

    and

    task

    f ac to r s a re am biguous . When pe r -

    fo rmance

    requi rements a re ill-defined,

    people

    who

    u n d e r e s t i ma t e

    the

    s i tua t iona l demands

    will display posi tive discrepancies between

    self-efficacy

    and

    p e r f o r ma n c e a t t a i n me n t s ;

    those

    who

    overes t ima te

    the

    demands wi l l

  • 8/10/2019 Bandura, A. (1977) Self-efficacy: toward a unifying theory of behavioural change, Psychological Review 84, 191

    14/25

    204

    AL BE R T BANDUR A

    exhibi t

    negative discrepancies. Therefore, in

    test ing predic t ions f rom the conceptual

    scheme presented here it is i mp o r t a n t t h a t

    sub jec t s unde rs t and wha t k ind

    of

    behavior

    will

    be requi red and the c i r cums t ance s in

    which they will be asked to pe r f o rm t hem.

    Moreover , pe r formances

    and the

    cor respond-

    in g efficacy expectations should be analyzed

    into sepa ra te activities, and p re fe r ab ly

    ordered by level of difficulty . In this type

    of microanalysis both

    the

    efficacy

    expecta-

    t ions

    and the corresponding behaviors are

    measured in t e r ms of explici t types of pe r -

    f o rmances

    ra the r than on the basis of global

    indices.

    The social learning determinants of self-

    efficacy

    can be var ied systematical ly and

    their

    effects measured. Hence, proposi t ions

    concerning the or igins of

    self-efficacy

    a re

    verif iable with some precision. A slowly d e-

    veloping motive, however, does

    not

    easily

    lend itself to being tested experimental ly.

    Ano the r

    dimension on which the al te rnat ive

    theories might be judged is their power to

    p r o d u c e the phenomena they purpor t to

    explain.

    As we shall see la ter , there are more

    diverse, expeditious, and powerful ways of

    c rea t ing

    self-efficacy

    t h a n by relying solely

    on novel s t imula t ion ar ising f rom explora to ry

    act ions.

    Wi th the ascendency of cognitive viewsof

    behavior, the concept of expectancy is as-

    suming an increasingly prominent place in

    contem porar y psychological thought (Bolles,

    1972b ; Heneman , & Schwab, 1972; I rwin,

    1 9 7 1 ) .

    However , vi r tual ly all of the theoriz-

    in g and

    exper imenta t ion

    has

    focused

    on

    ac t i on -ou t come expecta t ions.

    T he

    ideas

    ad -

    vanced in some of the theories nevertheless

    bear some likeness

    to the

    not ion

    of

    self-

    efficacy. According

    to the theory of

    person-

    al i ty

    proposed by Rot ter

    (1966 ) ,

    behavior

    varies as a

    func t ion

    of

    generalized expect-

    ancies t h a t outcomes a re de te rmined by one's

    ac t ions or by external forces beyond one's

    con t ro l . Such expecta t ions about the inst ru-

    menta l i ty

    of

    behavior

    a re

    considered

    to be

    largely a produc t ofone'shis toryof reinforce-

    ment . Much of the research wi thin th is t radi -

    t ion is concerned wi th the behavioral cor-

    re la tes

    of

    ind iv idua l differences

    in the

    tend-

    ency

    to

    perceive events

    as

    being ei ther per-

    sonally o r ex te rna l ly de te rmined .

    The not ion of locus of control is

    o f ten

    t rea ted in the l i te ra ture as analogous to self-

    efficacy.

    However , Rot t er 's (19 66 ) conceptual

    scheme is pr imari ly concerned wi th causal

    beliefs a b o u t

    a c t ion -ou tcome

    cont ingencies

    ra ther than wi th personal efficacy. Perceived

    self-efficacy

    and beliefs abou t the locus of

    causal i ty must be dis t inguished, because con-

    victions t h a t outcomes a re de te rmined by

    one's own ac t ions can have any n u mb e r of

    effects

    on

    self-efficacy

    and behavior. People

    who rega rd ou tcomes

    as

    personally de term ined

    but who lack the r equisi te skills wou ld ex-

    perience

    lo w

    self-efficacy

    and

    view activities

    with

    a

    sense

    of

    fu t i l i t y . Thus ,

    for

    example,

    a chi ld who

    fails

    to g rasp a r i thmet i c concepts