Teacher Efficacy, Supervision, and the Special Education Resource-Room Teacher Author(s): Theodore Coladarci and William A.

Breton Source: The Journal of Educational Research, Vol. 90, No. 4 (Mar. - Apr., 1997), pp. 230-239 Published by: Heldref Publications Stable URL: http://www.jstor.org/stable/27542097 Accessed: 11/03/2010 14:57
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=held. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Heldref Publications is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Educational Research.

http://www.jstor.org

Teacher and the

Efficacy, Special

Supervision, Education Teacher

WILLIAM A. BRETON
Kids Peace National Centers for Kids in Crisis

Resource-Room
THEODORE COLADARCI
University of Maine

beliefs

regarding

two

classes estimate and

of

ABSTRACT
Scale room resulted relation was modified A in a factor between context.

The Gibson
for use factor

and Dembo

Teacher Efficacy
resource instrument on reg The

expectation, lead to certain

"a person's outcomes," that one can

expectations: that a given an

an

outcome will

behavior

in the

ular-education

analysis structure comparable as reported teachers, instructional

education special of the modified to one

"conviction

based

successfully

the expectation, efficacy execute the behavior

was these teachers cy among resource-room and job satisfaction tenure, not frequency?of perceived utility?but nificantly findings context are related for both to teacher research

in prior research. and teacher effica supervision also examined. With sex, age, held constant, was of the sig these

required to produce the outcome" (Bandura, 1977, p. 193). Within the context of teaching, an outcome expectation is illustrated by the teacher who believes, for example, that skillful instruction can offset the effects of an impoverished
home self environment. but, rather, for Here, an efficacy abstract is expressed collective of not for one's teachers?the

supervision

The efficacy. implications in the special and practice

education

considered.

"normative

efficacy expectation, teacher's confidence

such instruction?that

teacher" (Denham & Michael, 1981, p. 41). An in contrast, would be reflected by the that he or she personally is capable of
one possesses personal agency with

the construct of teacher efficacy has enjoyed a considerable amount of empirical scrutiny in the Although past 15 years, few researchers have explored the import of this construct for the special education context. We had two general objectives in conducting this study: (a) to establish the validity of the Gibson and Dembo (1984) teacher effi
cacy scale when revised teachers and for (b) use with special the education association resource-room to examine

respect to the task of pedagogy. Teacher efficacy researchers traditionally have labeled the two sets of beliefs teaching efficacy and personal teach ing efficacy, respectively (Ashton & Webb, 1986; Gibson & Dembo, 1984). This language invites confusion, however, given the superordinate construct teacher efficacy. Although for somewhat different reasons, Hoy and Woolfolk (1990)
labeled these constructs general teaching efficacy and per

between teacher efficacy and the frequency and utility of that resource teachers reported instructional supervision an overview of the teacher effica receiving. We began with
cy construct and associated research.

sonal teaching efficacy, a distinction we simplify further to general efficacy and personal efficacy. is critical "because However labeled, this distinction individuals can believe that a particular course of action will
produce doubts certain about outcomes, they but can if they entertain the necessary serious activ whether perform

Research

on Teacher

Efficacy

ities such information (Bandura,

abilities uncertainties of the

The Teacher Efficacy Construct to a flurry of 15 years have borne witness research activity devoted to the study of teacher efficacy, or, as Dembo and Gibson (1985) defined the construct, "the extent to which teachers believe they can affect student generally credit Bandura learning" (p. 173). Researchers (1977) for providing the theoretical framework for studying The past
this construct. In his theory of self-efficacy, Bandura argued

does not influence their behavior" 1977, p. 193). Thus, one may be confident in the
normative his teacher, or her own yet harbor considerable prowess. about instructional

on teacher efficacy Research typically either the two items from the seminal Rand

has

employed study (Berman

that human

behavior

is influenced

by

the individual's 230

to Theodore Address Coladarci, correspondence 5766 Shibles Hall, Maine, College of Education, 04469-5766.

University Orono,

of ME

March/April

1997 [Vol. 90(No. 4)]

231 on personal efficacy. Smylie also found that interactions with one's colleagues about instructional matters carried a indirect effect on personal efficacy through the positive intervening variable "certainty of practice." And in their study of teacher efficacy and school climate, Hoy and Woolfolk (1993) found that school-level measures of acad emic emphasis, institutional integrity, and principal's influ ence each correlated with either personal or general efficacy. Both personal
to be higher

& McLaughlin, 1977) or some combination of the 30-item Teacher Efficacy Scale, later developed by Gibson and Dembo (1984). The popularity of these two approaches probably is related to the fact that both sets of measures
bear come some and semblance efficacy to Bandura's expectations distinction and, therefore, between out ostensibly

permit the delineation of general and personal efficacy. For is indicated if a example, general efficacy presumably teacher disagrees with the first Rand item, "When it comes right down to it, a teacher really can't do much because
most of a student's motivation and performance depends on

and general efficacy

elementary-level

also have been found

teachers when com

among

his or her home

Dembo item, "The

environment,"
amount that

or with
a student

the Gibson
can learn

and
is pri

marily related to family background." In contrast, personal efficacy is suggested if one agrees with the second Rand item, "If I really try hard, I can get through to even the most difficult or unmotivated students," or with the Gibson and Dembo item, "When the grades of my students improve, it I found more is usually because effective teaching
approaches." Teacher efficacy researchers can choose from other

pared with high school teachers (Fink, 1988; Parkay, Ole it is not clear whether this jnik, & Proller, 1986). However, difference can be attributed to a school effect or, rather, those who between reflects existing differences merely
select elementaryversus secondary-level teaching. Evans

and Tribble
ence between teachers.

(1986), for example,

elementaryand

found an analogous
secondary-level

differ

preservice

Consequences

of Teacher Efficacy

such as the Responsibility for Student instruments, Achievement Questionnaire 1981, (Guskey, Guskey, 1987), the Teacher Locus of Control Scale (Rose & Medway, 1981), theWebb Efficacy Scale (Ashton & Webb, 1986), or the Efficacy Vignettes (Ashton & Webb, 1986). Or, follow ing Midgley, Feldlaufer, and Eccles (1989), one can create a hybrid measure from existing instruments. These options
notwithstanding, most researchers have used either the

There is some evidence that teacher efficacy is related to academic achievement and teacher behaviors known to fos ter academic achievement (Ashton & Webb, 1986; Berman & McLaughlin, 1977; Gibson & Dembo, 1984; Greene, Anderson, & Loewen, 1993; Soar 1988; Hoy & Woolfolk, & Soar, 1982; also see Ashton, 1984, Dembo & Gibson, 1985), as well as with important student cognitions such as and appraisals (Midgley et al., performance expectancies 1989) and efficacy for achievement (Greene et al., 1988).
More peers, efficacious also show teachers, a preference relative for to their less-efficacious work rela collaborative

Rand items or the Teacher Efficacy Scale. (See Fink [1988] and Coladarci and Fink [1995] for an extended discussion
of the extant teacher efficacy instruments, the research asso

ciated with each, and results regarding their convergent discriminant validity.) Antecedents
Several education

and

of Teacher Efficacy
researchers on have examined of prospective the effects teachers' of teacher sense of

tionships (Morrison, Walker, Wakefield, & Solberg, 1994) and are more likely to adopt change proposals associated with formal innovations and staff development programs Kreutzer, Fritz, Miller-Heyl, (Berman & McLaughlin; MacPhee, 1995; Guskey, 1988; Poole, Okeafor, & Sloan, 1989; Rose & Medway, 1981; Smylie, 1988). Reporting a related finding, Coladarci (1992) found that teacher effica cy, when compared with such factors as income and school
climate, was the strongest predictor of a teacher's commit

the formation

efficacy. Spector ( 1990) found that personal efficacy among undergraduate students increased linearly during the 4-year
undergraduate program, culminating in student teaching.

ment

Perhaps consistent with this finding, Hoy and Woolfolk (1993) observed that personal efficacy was higher among
practicing teachers who had taken extra graduate courses in

to the teaching profession. Teacher efficacy has been linked to parent involvement
activities. Hoover-Dempsey, Bassler, and Brissie

in

school

(1987) found that teacher efficacy,

level, five with was the strongest of or among dimensions this parent

aggregated
the strongest Perhaps

at the school
predictors consistent rela of

education.
for general,

Spector also found a significant

but not personal, efficacy. That

quadratic
is, general

trend
effi

involvement.

linearly for the first 3 years of the under graduate experience but, unlike personal efficacy, declined after student teaching. A similar decline in general efficacy was reported by Hoy andWoolfolk (1990); also see Dembo and Gibson (1985, p. 178). cacy Other researchers have examined the effects of school context variables on teacher efficacy. Using path analysis, Smylie (1988) reported that the proportion of low-achieving
students in a teacher's classroom had a negative direct effect

increased

is the finding

that more-efficacious

teachers,

tive to their less-efficacious

regard teacher-parent relations

colleagues,
as a source

are less likely to

of stress (Parkay

et al., 1986). The Special Education

What we know?and

Context
do not know?about teacher effica

cy is limited largely to regular-education

settings. Notwith

232 Standing the unique pedagogical

teachers and, further, the arguable

The Journal of Educational Research

of mastery,

demands
importance

facing
of

resource
a strong

provides

for

successive

approximation

which,

too, is an important determinant

1977). It seems reasonable to

sense of teacher efficacy (e.g., DiBella-McCarthy McDaniel, 1989), there research involving special
ies that do exist, however,

of self-efficacy
conjecture,

(Bandura,
that

within this instructional context & McDaniel, 1989; Miller & is a paucity of teacher efficacy education settings. The few stud
are provocative.

therefore,

instructional
teacher efficacy.

supervision

would

have a salutary effect on

Allinder (1994), using the Teacher Efficacy Scale (Gib son & Dembo, 1984), found that resource teachers high in personal
planfulness,

there is little direct empirical support for Unfortunately, the posited relation between supervision and teacher effi cacy. In their Handbook of Research on Teacher Education for example, Glickman and Bey chapter on supervision, (1990) adduced two dissertations and one study of a single teacher in support of this association. In none of these studies was it clear that teacher efficacy was equivalent to the prevailing view of teacher efficacy, where the focus is on the teacher's sense of personal agency for effecting
change in one's students or classroom (e.g., Ashton &

efficacy
fairness,

tended

to exhibit
and

greater
clarity

organization,
in instruction.

enthusiasm,

These
toward to try

high-efficacy
instructional a variety

teachers

also were
and

more

inclined

experimentation?that of materials approaches

is, "willingness to teaching,

desire to find better ways of teaching, and implementation of progressive and innovative techniques" (p. 89). As noted above, this latter result is consonant with research involving regular education teachers (e.g., Guskey, 1988; Smylie, 1988). Although sparse, there also is evidence that teacher effi
cacy among may be related to special education referrals, at least regular-education teachers. Low-efficacy teachers

Webb, 1986; Dembo & Gibson, studies described by Glickman

teachers' school relation ever room sense practice between conceptualized?can context. of and their policy. competence

1985). In contrast, the few and Bey (1990) examined

or case, teacher within their no influence on on the research efficacy?how the resource

In any and be found

supervision

perhaps are more likely to refer students with academic teachers (Podell & Soodak, problems than are high-efficacy 1993). Similarly, teachers low in personal efficacy, unlike their high-efficacy counterparts, may tend to question the
appropriateness of a regular-education placement for stu

Method
Participants

difficulties 1993). (Soodak & Podell, experiencing Like Allinder (1994), those researchers used the Teacher Efficacy Scale. dents The Present Study We had two general objectives in this study. First, we set out to establish the validity of the Teacher Efficacy Scale when revised for use with resource teachers. The validity of the modified scale was assessed by comparing its factor structure to that obtained when administered to regular-edu cation teachers, as reported by Gibson and Dembo (1984) and Coladarci (1986). The former study was selected because it was the original factor analysis attending the publication of the Teacher Efficacy Scale; the latter study
was derived chosen from because, a representative like the current of sample investigation, a known population it

resource teachers a survey, We mailed the 865 Maine a to participate in our study. with letter them along inviting
Five days after the specified return date, we sent nonre

spondents a follow-up letter and an additional copy of the the two mailings combined, 580 resource survey. With teachers (67%) agreed to participate. Among these respon
dents the modal teacher was a woman with a baccalaureate

degree who had been teaching in the resource room for 6 to 10 years and was between 30 and 39 years of age. We did not hear from all resource teachers in Maine.
Nonetheless, an examination of state department documents

indicated guishable
in teaching

that these 580 teachers were from the population of Maine

experience, educational

virtually resource

indistin teachers
age, and

attainment,

sex. In short, the 580 teachers who chose to participate in our study appeared to be equivalent to the known popula
tion of resource teachers in Maine on general teacher char acteristics.

(and involving
Second, resource we teachers.

the same state).

pursued correlates of we teacher examined efficacy the among relation In particular,

between
supervision

teacher efficacy
that resource

and the frequency

teachers reported

and utility of
received.

having

A causal plausible
1990, prises p.

link between supervision and teacher efficacy is and has been proposed by others (e.g., Glickman,
22). Instructional monitoring, supervision, observing, insofar and as it com dialogue"

"assistance,

(Glickman & Bey, 1990, p. 549), arguably entails verbal persuasion, which is an important determinant of self-effi cacy for the task at hand (Bandura, 1977). And through its provision of constructive feedback, such supervision also

As will be seen below, however, this sample was reduced with the introduction of list-wise deletion of miss ing cases, particularly when applied to the multiple regres sion analyses (N = 378). The resulting sample for the lat ter analyses still constituted a large number of teachers (65% of the initial respondents, 44% of the population). the restricted sample, like the initial sample, Moreover, to the population of resource proved to be equivalent teachers, at least with respect to data available
state department age, (i.e., and sex). teaching experience, attainment,

through the
educational

March/April
Instruments

1997 [Vol. 90(No. 4)]

233 item, which lowing personal-efficacy items from the Rand study: is one of the two

Teacher efficacy. All items rested on a 6-point scale rang ing from strongly disagree to strongly agree.1 For most the Gibson and Dembo instrument simply items, modifying
entailed teacher, corrected changing or, similarly, the term teacher to resource-room room. (e.g., We he/she), also as classroom to resource

"If I really try hard, I can get through to even the most
difficult or unmotivated students."

several

semantic

awkwardnesses

Here, 1 in every 4 (26%) resource teachers disagreed with this statement.2 Even greater variability was found on the two items below: "The time spent in my resource room program has little influence on students compared to the influence of their
home "Even ties may environment." a resource-room not reach many teacher with good teaching abili

as substituted the two efficacy items from the Rand 1977) for two items on the study (Berman & McLaughlin, instrument judged to be equivalent Gibson and Dembo (their No. 15 and No. 16). The latter change allowed us to well
directly seminal examine and the resource Rand teachers' items. Because responses the two to sets the of often-used

students."

items were equivalent, this substitution seemingly did not attenuate the reliability or validity of the Gibson and
Dembo instrument. Resource teachers also were asked to rate Supervision.

Maximum

variability was observed the teachers agreed; half disagreed.

for both items: Half of

the frequency Two domains

observation,

and utility of the supervision they received. of supervision were specified: (a) formal
in which classroom observations are scheduled

Not all items demonstrated such variability, however. For example, 81% of teachers disagreed with the following gen eral efficacy statement, which was derived from the second Rand item:
"When really tion it comes can't and right down because depends to it, a resource-room most on of a student's teacher motiva

at a predetermined time for identifying strengths and weaknesses and (b) performance
which between represents a resource informal, teacher often and spontaneous, a supervisor

instructional consultation,
exchanges about instruc

do much

performance

the home

environment."

tional practices. We asked for separate ratings of each of the following possible supervisors: building principal, special
education assistant For was director, superintendent, each assessed domain through and a curriculum and coordinator, any other superintendent, supervisor. supervision scale ranging frequency from 1

we these item-level observations, Moving beyond items and then reversed the scales of negatively worded
determined the teacher's mean response across all 30 items,

supervisor, Likert-type

yielding a summary indicator of teacher efficacy ranging from 1 to 6. A mean of 3.5 represented the demarcation between low and high efficacy. We obtained a mean of 4.25
(SD = .45) on this summary measure, with a range of 2.33 to

to 7 (weekly)', the utility scale ranged from 1 (not helpful at all) to 5 (extremely helpful). We formed a super vision-frequency composite by taking the mean of a (never)
teacher's frequency ratings across domains and supervisors;

5.50. It is difficult

to appraise this figure in the absence of an

resource to more a teacher items clearly than not.

but the average standard, accepted was an sentiment expressing efficacy Factor analysis. We conducted

a supervision-utility composite was formed in a parallel fashion. By considering the frequency and utility of both domains of supervision, we hoped to get at the "assistance,
monitoring, observing, and dialogue" function of supervi

principal-components

sion (Glickman & Bey, Results

First, we report the

1990, p. 549).

factor analysis of the modified teacher efficacy scale and, consistent with the prevailing conceptualization of teacher to two forced the We solution factors. efficacy, orthogonal used list-wise deletion of missing cases for this analysis, resulting inN= 520 (89.7% of the initial sample). As noted
above, analyses we compared reported in our two results studies with those of the factor involving regular-educa

results

from

the

teacher

efficacy

analyses. We begin by considering descriptive analyses of the modified teacher efficacy scale and then proceed to the factor structure of the instrument. This is followed by the
results bearing on the relation between resource teachers'

tion teachers (Coladarci, 1986; Gibson & Dembo, 1984). Roughly 28% of the total item variance was explained by these two factors, a finding comparable to the 29% and 27% obtained (1986),
the first

and Dembo (1984) and Coladarci by Gibson for 17% of item variance, respectively. Accounting
factor represented a resource teacher's sense of per

sense of efficacy
vision they

and the frequency

and utility of the super

received.

sonal efficacy (see Table 1). For example, below carried the highest factor loadings:

the three items

Teacher Efficacy Descriptive analyses. A simple examination of item dis tributions revealed considerable variability among these resource teachers in their efficacy beliefs. Consider the fol

"When any of my students show improvement, it is because I found better ways of teaching them." "If my supervisor suggested that I change some of my class curriculum, I would feel confident that I have the
necessary skills to implement the change."

234

The

Journal

of Educational

Research

Table

1.?Teacher

Efficacy:

Factor

Loadings

Factor Item loadinga Personal If one of my When

efficacy .60 them. .58

students couldn't do a class assignment, I would be able to special education the assignment was at the correct level of difficulty. accurately assess whether any of my students show improvement, it is because I found better ways I would of teaching feel confident

If my supervisor suggests that I change some of my class curriculum, that I have the necessary skills to implement the change. If one of my necessary When students mastered steps in teaching of my students a new concept that concept. improve, is having quickly, it probably because

.58 be because I knew the .57

would

the grades approaches.

it is usually with

I found more

effective

teaching .56 able to adjust .53 the .52

When

a resource-room it to the student's

student level.

difficulty

an assignment,

I am usually

Between

my necessary

program and my own teacher-training skills to be an effective resource-room

teaching experience, teacher.

I have obtained

If one of my students did not remember information I gave in the previous how to increase the student's retention in the next lesson. If students inmy class become to redirect them quickly. If I really disruptive and noisy, the most I feel assured

lesson,

I would

know .52

that I know

some

techniques .50 .47 .46

try hard, I can get through training

to even

difficult

or unmotivated resource

students. room.

I have enough

to deal with most

learning problems

in my

If parents comment to me that their child behaves much better inmy resource room program that at home, it would probably be because I have some specific techniques of managing the child's behavior which they may lack. When one of my extra effort. students does better than expected, many times it is because I exerted a little

.45

.43

General When

efficacy because most of .67 .61

teacher really can't do much it comes right down to it, a resource-room a student's motivation and performance depends on the home environment. amount that a special education student will learn is primarily on students related

The The

to family background. to the influence

time spent in my resource of their home environment.

room has little influence

compared

.53 I am frustrated likely to accept teachers inmy attempts to help inmy students. resource .51 are not a very powerful influence on .49 .52

Because

of lack of support from the community, at home, they aren't

If students aren't disciplined room program. When

any discipline

resource-room all factors are considered, resource-room student achievement.

teacher judge how much to expect from a student by Parent conferences help a resource-room and so on. giving the teacher an idea of the parents' values toward education, discipline, If parents would do more with their children, I could do more inmy resource room. assignment,

.49 .46

students If one of my new resource-room there is little that I can do to increase Even The a resource-room influences teaching. aFactor loadings greater or equal teacher with education good

cannot remain on task for a particular that student's attention. teaching abilities may experience not reach many can be overcome

.45 students. by good -.41 .44

of a special

student's

home

to .40 are presented;

list-wise N = 520.

"If one of my special education students couldn't do a class assignment, Iwould be able to accurately assess whether the assignment was at the correct level of difficulty." Twelve of the 13 items that loaded on this factor also loaded on the comparable factor in at least one of the two studies; eight of these items loaded on both. comparison

The remaining
common to

item that loaded on this factor, although not

comparison study, nonetheless reflects

either

the concept of personal "If parents comment

better in my resource

efficacy: tome
room

that their child behaves much

program than at home, it

would

probably

be because

I have

some specific

tech

March/April niques
may

1997 [Vol. 90(No. 4)] of managing

235 annually or more. The modal teacher found observations to be "somewhat" helpful; special education directors received slightly higher ratings than principals did. A similar picture emerged regarding the informal consul tation that these resource teachers received on instructional issues. For example, 43% of the teachers reported that either the principal or special education director provided such consultation; 33% received consultation from both supervi sors; and 18% received no such consultation. Ratings of the frequency and utility of informal consultations tended to be generally higher than those for formal classroom observations.
Regression analyses. We used ordinary least squares

their child's

behavior

which

they

lack."

% of total item variance, the second for 11 Accounting not uni factor comprised general efficacy items?although formly so. The three highest loading items were as follows:
"When really tion ment." it comes can't and right down because to it, a resource-room most on of a student's the home teacher motiva environ do much performance

depends

"The amount that a special education student will learn is primarily related to family background." "The time spent in my resource room program has little influence on students compared
home environment."

to the influence of their

the extent to which the frequency and regression of supervision predicted teacher efficacy. The depen utility to examine
teacher dent variable, efficacy, was constructed in two ways:

For this factor, 9 of the 11 items also loaded on the gen factor in at least one of the two comparison eral-efficacy studies; 5 of these items loaded in both. The two remaining items loading on this factor were not common to either comparison study. One of these items appears to represent personal efficacy (in this case, its absence):
"If one of my new resource room students cannot remain

(a) total efficacy, obtained by summing a teacher's respons es across all 30 items on the modified teacher efficacy scale = (Cronbach's a .77) and (b) personal efficacy, the sum of 13 items loading on our first factor (Cronbach's a = .75). The full 6-point scale of each item was used in creating these
composites; negatively worded items were recoded.

We considered six independent variables. the two supervision composites?frequency

several ables: teacher sex, age, characteristics resource-room were included and tenure,

In addition to and utility?

as control job vari satisfaction

on task for a particular

can do to increase that

assignment,
student's

there is little that I

attention."

our factor analysis This errant item notwithstanding, the modified teacher efficacy scale, when administered
resource teachers, produced a factor structure comparable

of to

(a composite derived from questions regarding the teacher's to satisfaction with the current position and commitment
special education). Means, standard deviations, and inter

correlations
In separate ures was

appear in Table 2.
equations, regressed on each the of six the teacher independent efficacy meas variables.

to that reported by Gibson and Dembo (1984) and Coladar ci (1986) in their studies of regular-education teachers.
However, the errant item underscores an emerging question

Because

in the teacher efficacy literature:What does the general effi cacy factor really represent? We return to this question in
our cern final about discussion. the meaning But of because general our data sustain (e.g., the con efficacy Coladarci,

(a) not all teachers responded to all items on the survey and (b) we used list-wise deletion of missing cases, these analyses were based on 378 teachers. As indicated
however, the 378 resource teachers' general charac

above,

1992; Guskey & Passaro, 1994;Woolfolk & Hoy, did not include this factor in the analyses below. Teacher Efficacy and Supervision

1990), we

First, we briefly provided descriptive information bearing on the frequency and utility of supervision that the resource teachers reported to have received; Breton and Donaldson (1991) described these and related data in greater detail. these descriptive data are the results of the Following
regression dependent analyses, variable. in which teacher efficacy served as the

teristics were similar to both the initial sample and the pop ulation. Further, the results of the list-wise analyses did not differ appreciably from those based on pair-wise deletion of missing cases, which made use of the full sample. The regression similar results, equations produced whether the dependent variable was total efficacy or per sonal efficacy (see Table 3). First, both multiple correlations were modest, if statistically significant (a = .05): .36 for total efficacy and .31 for personal efficacy. Thus, between 10% and 13% of the variance in teacher efficacy was explained by the linear combination of the six independent variables, depending on the dependent variable.
Second, the same variables across both equations signif

Supervision frequency and utility. Forty-five percent of these teachers reported that either the principal or special education director conducted formal observations in their an were additional 30% observed both classroom; super by visors. However, 17% of the resource teachers indicated
that they were not observed by any supervisor. Observa

icantly, ifmodestly, predicted teacher efficacy. Arguably the most important finding from these regressions was that,
between the two supervision composites, it was the per

ceived utility of supervision?not its frequency?that a sense to related of efficacy. Also, teacher's nificantly
er teacher efficacy was observed among women

sig high
and,

tions, when
annual basis,

they did occur, typically were conducted

although some teachers were observed

on an
semi

further, those who

resource-room

expressed
Age

higher
was

satisfaction with
to teacher

their

position.

related

efficacy;

236

The

Journal

of Educational

Research

Table

2.?Descriptive

Statistics:

Means,

Standard

Deviations,

and

Intercorrelations

Variable

1 2 3 4. 5. 6. 7 8 M SD

Total efficacy Personal efficacy Supervision Supervision Sex frequency utility

.85 .04 .17 .15 .21

.06 .16 .16 .17 .11 .17 73.67 7.75

.17 -.09 -.02 -.21 .02 .22 .81 .03 -.03 -.08 .25 .05 -.01 .01 .11 .86 .35

Age Resource-room Job satisfaction

tenure

.13 .24 127.79 12.64

.41 .08 34.24 8.30

.02 5.61 3.79 .07 .73

Note:

These

statistics analyses.

are based

on the 378 resource

teachers

for whom

complete

data were

available

for the

regression

resource-room

tenure

was

not.

Irrespective

of

resource

room experience, older sense of efficacy. higher Summary and Discussion

teachers demonstrated

a slightly

Table

3.?Mutiple

Regression

Analysis

Variable

SE(b)
Dependent variable: .79 .69 .72 1.80 Sex 4.69 Age .27 .22 .88 1.76 .08 .18 Total efficacy .04 .13 .13 .18 .07 .18 .87 2.49* 2.66* 3.30* 1.22 3.47*

Our concluding
efficacy among

remarks focus on the level of teacher

teachers in our sample; the factor

resource

Supervision Supervision

frequency utility

structure of the teacher efficacy

the vision resource-room and teacher context; efficacy. and the

instrument, modified
relation between

for
Resource-room Job satisfaction tenure 3.04

super

Teacher Efficacy Among Resource

Teachers
Supervision Supervision Sex

Dependent frequency utility

variable:

Personal

efficacy .49 .45 1.10 .05 .11 .55 .07 .13 .15 .14 .07 .11 1.33 2.44* 2.96* 2.57* 1.32 2.16*

in Our sample of resource teachers varied considerably sense the teacher of their reported average efficacy;
teacher's mean was 4.25 on a 6-point scale. The number of

there is not yet studies on teacher efficacy notwithstanding, teacher to the level of efficacy any standard by which judge in any one sample. Although 4.25 indicates that the average
resource ment teacher to more in our statements sample than expressed not, an efficacy literature senti does the extant

Age Resource-room Job satisfaction

tenure

.65 1.10 3.26 .13 .15 1.18

not allow us to appraise the relative value of this figure. As a normative question, whether 4.25 is good or high cannot
be ing answered. similar Clearly, analyses we on need similar additional studies?involv we can samples?before

Note: For total efficacy, R = .36, E(6, 371) = 9.27, p < a. For per sonal efficacy, R = .31, E(6, 371 ) = 6.80, p < a. For both equations, b is the unstandardized partial slope: ? is the standardized equiva statistics are based on the 378 resource teachers for lent. These whom complete data were available for the regression analyses. *p < a (one-tailed).

approach this question with any confidence. And this is par ticularly true with respect to studies of teacher efficacy in
the As context. education special our a of reference, point data can be compared with

their

regular-education

colleagues.

Perhaps

this preliminary

those obtained
Gibson

by Coladarci
instrument teachers

(1986), who
in Maine. Within

administered
sample that sample,

the
of a

finding reflects, in part, the differences between these two educational contexts in how instruction is planned, deliv ered, and evaluated. On the other hand, this finding also might be revealing the entering characteristics of those who
elect ture, to become of course, special must be education explored teachers. more Either systematically conjec in

and Dembo

to a representative

regular-education

mean of 4.11
The difference

(SD

.45) was obtained

these two

across the 30 items.

means (i.e., 4.25

subsequent

studies.

between

sample

to a statistically vs. 4.11) corresponds significant effect size3 of .31, ?(830) = 4.67, p < .01. That is, the sense of effi
cacy among resource teachers in Maine is, on average,

Factor Structure of Teacher Efficacy When

scale

the Gibson
for

and Dembo
use in the

(1984)
resource-room

teacher efficacy
context, a

roughly one third of a standard deviation higher than that of

is modified

March/April

1997 [Vol. 90(No. 4)]

237 gent and discriminant validity of teacher efficacy scales et al., 1991), as well as research Woolfolk Fink, 1988; (e.g., instruments the properties of modified that investigates & Passaro, 1994). (e.g., Guskey But this area of research also would profit from more studies with a decidedly qualitative orientation. For exam ple, Coladarci (1992) has called for research employing a in which teachers' thoughts are "think aloud" methodology probed as they respond to teacher efficacy items. Think aloud studies would throw needed light on the kinds of fac to
as tors, invoke considerations, as they Just standards, consider as and so on protocols forth, a that teachers efficacy interviews statements teacher and

factor structure emerges that is comparable to that found in studies of regular-education teachers (Coladarci, 1986; Gib son & Dembo, 1984). Personal efficacy. A personal efficacy factor clearly sur faced in the present study. As in both comparison studies, this factor is characterized by items that capture the
teacher's sense of personal agency (e.g., "When any of my

students
ways cy of factor

show
teaching among

improvement,
them"). resource The

it is because
presence of

I found better
effica the fruitfulness

a personal

teachers

suggests

of pursuing
those we being described

lines of teacher efficacy

conducted above. in the

research

similar
context,

regular-education

instrument.

think-aloud

General efficacy. The second factor, general efficacy, does not enjoy the same clarity of definition. Some of the items loading on this factor, like the one below, reflect one's view of the "normative teacher" (Denham & Michael, 1981, p. 41):
"When it comes right down to it, a resource-room

with test takers have clarified the meaning of derived factors regarding students' knowledge of science (Hamilton, Nuss baum, & Snow, 1995), we believe that eliciting teachers' thoughts to nominally personal and nominally general effi cacy statements will add considerably to our understanding of themeaning Supervision and import of the two efficacy constructs.

teacher [italics added] really can't do much because most

of a student's motivation and performance depends on the home environment."

and Teacher Efficacy its frequency? among

their

However, the general equivalent

the following item, which loaded on in factor the present study and, in an efficacy form, in both comparison studies: consider

The perceived utility of supervision?not teacher efficacy significantly predicted

resource teachers. That is, teachers who felt

these
supervi

"The time spent inmy [italics added] resource room pro gram has little influence on students compared to the
influence Does teacher? this The of their home reflect of environment." one's sense my of seems the normative

clearly use

the possessive

to complicate

sion was helpful tended to report a higher sense of teacher efficacy than those who reported less-positive views of the supervision they received. And this held regardless of sex, (and age, or job satisfaction, each of which significantly own in its teacher efficacy right. positively) predicted But these are weak effects. For example, only 13% of the variance in total efficacy and 10% in personal efficacy were accounted
equation, modest +.13.

such an interpretation. Our general efficacy factor is further confounded by an additional item involving self-referent
language:

for by the six independent variables. And

the regression That weight each for supervision deviation utility is, with standard

in each
was increase a

"If one of my [italics added] new resource room students

cannot remain on task for a particular assignment, there

is little that I can do to increase that student's attention."

In short, the resource this factor analysis sense of suggests efficacy that a measure the of same

teacher's

presents

in supervision utility, teacher efficacy increased only 13% of a standard deviation (other independent variables held this statistic is within the range of constant). Although effects that characterize the teacher efficacy literature to date (Coladarci, nonetheless raises 1992), its magnitude questions about the import of direct supervision for the
development However, ical factors of one that a teacher's also must sense address constrained of at efficacy. least the two methodolog relation between

problem facing those who

ular-education efficacy What ers have remains is "general argued that context. to be

study teacher efficacy

the meaning

in the reg
of general

Specifically, clarified. a measure efficacy,

arguably

efficacy" general

of? rather

Some than

research reflecting

supervision
First, by as variance:

utility and teacher efficacy

known, Where measures variance is of

in the present study.

are coefficients affected are

is well

association limited,

outcome
teacher's

(Bandura, 1977) or the normative expectations teacher (Denham & Michael, 1981), ismore indicative of a
pupil-control ideology, bureaucratic orientation,

attenuated. This doubtless

tional supervision among

is a problem
resource

in studying

instruc

teachers,

a population

toward education (Woolfolk & Rosoff, & Hoy, 1991). Others have Hoy, 1990; Woolfolk, that teacher suggested efficacy is analogous to the locus-of
control cy construct, an with external general efficacy and and personal internal effica orienta reflecting orientation

and fundamental

attitude

1991 ;Rydell, Gage, & (Breton & Donaldson, and elsewhere Coin?s, 1986) 1982) (e.g., Moya & Glenda, tends to see supervision as both insufficiently frequent and
insufficiently teacher in our useful. sample Thus, was insofar observed as only the modal once each resource year, the

that inMaine

tion, respectively
quantitative

(Guskey & Passaro,

is needed that

1994). Clearly, more

examines the conver

research

nonsignificant effects of supervision frequency could be, in part, a statistical artifact. And the same statistical principle

238 may have influenced the effects associated with supervision utility, albeit less so because the problem of variance was not as pronounced for this variable. Additional studies would be helpful for appraising this possibility. A second methodological factor to consider here reflects a limitation of the present study. Specifically, in focusing on the frequency and utility of supervision, we overlooked the important interpersonal milieu within which any superviso ry practice exists and, therefore, within which any supervi sory practice should be appraised. As Glickman and Bey (1990) argued, one should not study supervision indepen as "shared understandings, dently of such considerations clear purpose, and sensitivity to individual needs of teach ers" (p. 554). By incorporating essential aspects of the inter into a study of this kind, subsequent personal milieu researchers will move toward a better understanding of the relation between supervision and teachers' efficacy in the
resource room.

The Journal of Educational Research

sense of efficacy and commitment to teach Coladarci, T. (1992). Teachers' Education, 60, 323-337. ing. Journal of Experimental Coladarci, T, & Fink, D. R. (1995, April). Correlations among measures the same thing? Paper present of teacher efficacy: Are they measuring ed at the 1995 meeting of the American Educational Research Associa tion, San Francisco. sense of efficacy: An S. (1985). Teachers' Dembo, M. H., & Gibson, School Journal, important factor in school improvement. Elementary 86, 173-184. J. J. (1981). Teacher sense of efficacy: A def Denham, C. H., & Michael, inition of the construct and a model for further research. Educational Research Quarterly, 5, 39-63. E. A. (1989). Enhancing teacher effi H., & McDaniel, DiBella-McCarthy, cacy in special education. Teaching Exceptional Children, 21, 34-38. Evans, E. D., & Tribble, M. (1986, April). Perceived teaching problems, s elf-efficacy, to teaching among preservice and commitment teachers. at the annual meeting of the American Educational Paper presented Research Association, San Francisco. and discrim Fink, D. R. (1988). Teacher efficacy instruments: Convergent inant validity and additional correlates. Unpublished doctoral disserta of Maine. tion, University J.C, & MacPhee, D. (1995). Foster Fritz, J. J., Miller-Heyl, J., Kreutzer, and classroom ing personal teaching efficacy through staff development activities. The Journal of Educational Research, 88, 200-208. Gibson, S., & Dembo, M. H. (1984). Teacher efficacy: A construct valida tion. Journal of Educational 76, 569-582. Psychology, C. D. (1990). Supervision Glickman, of instruction: A developmental (2nd ed.). Boston: Allyn & Bacon. approach C. D., & Bey, T. M. In W. R. Houston Glickman, (1990). Supervision. New (Ed.), Handbook of research on teacher education (pp. 549-566). York: Macmillan. M. L., Anderson, R. N., & Loewen, P. S. (1988, April). Relation ' ' teachers and students ships among thinking skills, sense of efficacy, and student achievement. of the Paper presented at the annual meeting American Educational Research Association, New Orleans. of the responsibility teachers assume Guskey, T. R. (1981). Measurement for academic successes and failures in the classroom. Journal of Teacher Education, 32(3), 44-51. variables that affect measures of teacher Guskey, T. R. (1987). Context Research, 81, 41-47. efficacy. The Journal of Educational and attitudes toward Guskey, T. R. (1988). Teacher efficacy, self-concept, the implementation of instructional innovation. Teaching and Teacher 4, 63-69. Education, Greene, Guskey, T R., & Passaro, P. D. (1994). Teacher struct dimensions. American Educational 627-643. efficacy: A study of con Research Journal, 31,

we believe our results point to the promise of pursuing teacher efficacy research within the special education context. Some of this promise is in the form of as the emergence of a clear personal effi born fruit?such In conclusion,
cacy factor?whereas some of this promise represents a

additional correlates of teacher effi challenge?exploring and the cacy meaning of general efficacy. tackling NOTES
statis 1. The modified 30-item teacher efficacy scale, with item-level tics, is available from the authors. we of these perceptions, the variability 2. To more vividly convey reduced each item to a disagree/agree responses by collapsing dichotomy or strong across the gradations of the perception (i.e., slightly, moderately, for each of the two resulting categories. ly) and reporting the percentage is computed by dividing 3. Effect size, a standardized mean difference, standard deviation the mean difference by the pooled within-group (Hedges & Olkin, 1985, p. 79).

REFERENCES
Allinder, R. M. (1994). The relationship between efficacy and the instruc Teacher tional practices of special education teachers and consultants. Education and Special Education, 17, 86-95. Ashton, P. T. (1984). Teacher efficacy: A motivational paradigm for effec tive teacher education. Journal of Teacher Education, 35, 28-32. and the teacher's sense of efficacy. In C. Ashton, P. T. (1985). Motivation on motivation in education: The Ames & R. Ames (Eds.), Research Press. classroom milieu (pp. 141-171). New York: Academic a difference: Teachers' sense P. T., & Webb, R. B. (1986). Making of efficacy and student achievement. New York: Longman. Toward a unifying Bandura, A. (1977). Self-efficacy: theory of behavioral Review, 84, 191-215. change. Psychological A Social foundations A. Bandura, (1986). of thought and action: Cliffs, NJ: Prentice-Hill. theory. Englewood social-cognitive M. (1977). Federal programs Berman, P., & McLaughlin, supporting edu and continu cational change (Vol. 3): Factors affecting implementation CA: The Rand Cor Santa Monica, ation (Report No. R-1589/7-HEW). Ashton, poration. G. A., Jr. (1991). Too little, too late? The Breton, W. A., & Donaldson, resource room teachers. Journal of Special Edu of Maine supervision 114-125. cation, 25(1), Coladarci, T. (1986, April). Teacher efficacy and school climate. Paper pre Research of the American Educational sented at the annual meeting Association, San Francisco.

E. M., & Snow, R. E. (1995, April). Alterna L. S., Nussbaum, Hamilton, tive interview procedures for validating science assessments. Paper pre of the American sented at the annual meeting Educational Research San Francisco. Association, Hedges, L. V, & Olkin, Academic Press. I. Statistical methods for meta-analysis. Orlando:

J. S. (1987). Parent K. V, Bassler, O. C, & Brissie, Hoover-Dempsey, Contributions of teacher efficacy, school socioeconomic involvement: American Educational Research status, and other school characteristics. Journal, 24, 417-35. of student teachers. A. E. (1990). Socialization Hoy, W K., & Woolfolk, American Educational Hoy, W. K., & Woolfolk, health organizational 355-372. Research Journal, 27, 279-300. sense of efficacy and the A. E. (1993). Teachers' School Journal, of schools. The Elementary 93,

in teacher effi J. S. (1989). Change Midgley, C, Feldlaufer, H., & Eccles, cacy and student self- and task-related beliefs inmathematics during the transition to junior high school. Journal of Educational 81, Psychology, 247-248. in spe E. A. (1989). Enhancing teacher efficacy Miller, R., & McDaniel, of student performance. Academ cial education through the assessment ic Therapy, 25, 171-181. G. M., Walker, D? Wakefield, P., & Solberg, S. (1994). Teacher Morrison, to efficacy for for collaborative relationships: Relationship preferences in the Schools, domains. Psychology 31, teaching in prevention-related 221-231. of special education teachers. S., & Glenda, G. (1982). Evaluation Moya, and Special Education, Teacher Education 5, 37-41.

March/April

1997 [Vol. 90(No. 4)]

239
to individual and psychological antecedents teacher Organizational Research Journal, 25, 1-30. change. American Educational to Soar, R. S., & Soar, R. M. (1982, March). A process-product approach of the teachers' sense of efficacy. Paper presented at the annual meeting American Educational Research Association, New York City. Soodak, L. C, & Podell, D. M. (1993). Teacher efficacy and student prob lem as factors in special education referral. The Journal of Special Edu cation, 27, 66-81. in preservice teachers. Spector, J. E. (1990, April). Efficacy for teaching at the annual meeting of the American Educational Paper presented Boston. Research Association, A. E., & Hoy, W. K. (1990). Prospective teachers' sense of effi Woolfolk, 82, cacy and beliefs about control. Journal of Educational Psychology, 81-91. sense of effi A. E., Rosoff, B., & Hoy, W. K. (1991). Teachers' Woolfolk, students. Teaching and Teacher cacy and their beliefs about managing Education, 6, 137-148.

Parkay, F. W., Olejnik, S., & Proller, N. (1986, April). A study of the rela tionships among teacher efficacy, locus of control, and stress. Paper pre of the American Educational Research sented at the annual meeting San Francisco. Association, Podell, D. M, & Soodak, L. C. (1993). Teacher efficacy and bias in spe cial education referrals. The Journal Research, 86, of Educational 247-253. Poole, M. G., Okeafor, K., & Sloan, E. C. (1989, March). Teachers'inter actions, personal efficacy, and change implementation. Paper presented at the annual meeting of the American Educational Research Associa tion, San Francisco. Rose, J. S., & Medway, over their control F. J. (1981). Measurement of teachers' beliefs student outcome. The Journal of Educational in

Research, 74,185-190. Rydell, L, Gage, B., & Coin?s, A. (1986). Teacher recruitment and reten tion inMaine. Augusta, ME: Office of Policy and Legal Analysis. function of staff development: Smylie, M. A. (1988). The enhancement

If you have a basic aptitude for math and the desire to help others, you can get a lot of satisfaction by volunteering your time and skills to people who

need help doing their taxes. 80,000 people already have. Join them. To find out about the free 1RS train ing program, call 1-800-424-1040 now.

A PublicService of & This Publication

\folunteer new. And you'll make someone's taxes less taxing later.