Group Title: relationship of teacher praise to teacher orientation and their relationships to students' perceived competence, intrinsic motivation, and achievement /
Title: The Relationship of teacher praise to teacher orientation and their relationships to students' perceived competence, intrinsic motivation, and achievement /
Full Citation
Permanent Link:
 Material Information
Title: The Relationship of teacher praise to teacher orientation and their relationships to students' perceived competence, intrinsic motivation, and achievement /
Physical Description: viii, 170 leaves : ill. ; 28 cm.
Language: English
Creator: Denny, Patricia L., 1954-
Publication Date: 1985
Copyright Date: 1985
Subject: Teacher-student relationships   ( lcsh )
Praise   ( lcsh )
Learning, Psychology of   ( lcsh )
Curriculum and Instruction thesis Ph. D
Dissertations, Academic -- Curriculum and Instruction -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis (Ph. D.)--University of Florida, 1985.
Bibliography: Bibliography: leaves 163-169.
Statement of Responsibility: by Patricia L. Denny.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00099583
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000875745
notis - AEH3311
oclc - 014696897


This item has the following downloads:

PDF ( 6 MBs ) ( PDF )

Full Text







This dissertation is dedicated to
my husband, John, and my son, Bryan.


The years I have spent at the University of Florida

have been ones of intellectual and personal growth. This

growth is due in large part to interactions with several

professors, friends, and my family. This dissertation

is the culmination of my experiences with these individuals.

I would like to thank the members of my committee:

Dr. Linda Lamme, for her excellent guidance and support

during my graduate career, for the interest she has

expressed in my academic growth, and for the patience she

exercised when our ideas and timetables were at odds.

Dr. Patricia Ashton, for her intellectual stimulation

in classes and discussions, for her insight and knowledge

of children, and for her belief in me.

Dr. Dorene Ross, for her active participation in

helping me write and think, for her encouragement of

academic excellence, and for her friendship.

Dr. Jamie Algina, for his insight, for his help in

data analysis, and for his patience.

Dr. Suzanne Krogh, for her encouragement and interest,

and for her valuable suggestions about data collection and


I would like to thank my friends:

Marilyn McAuliffe, for her gentle encouragement and

reminders without which this disseration may have remained

unfinished, for her caring, and for her friendship.

Dwight Rogers, for his many hours of thoughtful discus-

sions throughout our graduate careers and for his example

of sincere interest and concern for children and their


Beth Clark, for her encouragement and humor during

our graduate studies.

Darla Beaty, Maria Bolanos, Beth Goodridge, and Mary

Gouge, for their generous help as observers and the many

hours they spent learning and scoring the observation forms.

Karen Kilgore, for her help in testing the children.

I would like to thank my family:

John, my husband, for his love, understanding, and

support always;

Mom and Dad, for their love and confidence in me.



ACKNOWLEDGEMENTS..... ........................ iii

ABSTRACT...................................... vii


I INTRODUCTION............................. 1

Statement of the Problem................. 4
Definition of Terms ...................... 5
Need for the Study....................... 6
Limitations of the Study................. 11
Summary.................................. 12

II REVIEW OF THE RESEARCH................... 13

Introduction............................. 13
Praise and Achievement................... 15
Praise and Intrinsic Motivation.......... 21
Praise and Competence.................... 29
Functional Aspects of Praise............ 37
Summary.................................. 62

III METHODOLOGY.............................. 64

Design of the Study...................... 65
Subjects.................................. 67
Instrumentation.......................... 68
Data Collection.......................... 84
Summary.................................. 88

IV RESULTS.................................. 90

Introduction............................. 90
Findings.................................. 99
Summary.................................. 106


Summary of Findings...................... 109
Discussion............ .................... 11l
Suggestions for Further Research.......... 127
Implications for Teaching................ 129
Conclusion............................... 131




MEASURES................................. 146

OTHER MEASURES.............. .............. 154


F PRAISE STATEMENTS........................ 158

REFERENCES............................... .... 163

BIOGRAPHCIAL SKETCH........................... 170

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy



Patricia L. Denny

May 1985

Chairperson: Dr. Linda L. Lamme
Cochairperson: Dr. Patricia Ashton
Major Department: Curriculum and Instruction

This observational correlational study investigated

the relationship between frequencies of various types of

teacher praise and teacher orientation toward supporting

autonomy in versus controlling children. Further, this

study investigated the relationships between the frequencies

of various types of teacher praise and children's feelings

of competence, intrinsic motivation, and achievement.

The relationships between teacher orientation and the three

children's measures were also studied. The subjects were

154 third-grade children in middle-ability reading groups

and their 22 teachers in Alachua County, Florida.

Prior to classroom observation, the teachers' orienta-

tion was assessed. The teachers' use of praise was recorded

during middle-ability reading groups by four different

observers on four separate occasions of 30 minutes each.


The children's feelings of competence and intrinsic

motivation were assessed in small group test situations

while the study was in progress. Achievement data were

collected at the end of the academic year.

A series of correlations was computed to analyze the

data. The findings indicated that teacher orientation and

teacher praise had few relationships with the children in

the study. No significant relationship was found between

teachers' orientation and teachers' use of praise. No

significant relationships were found between teachers'

orientation and children's feelings of competence, intrinsic

motivation, and achievement. No significant relationships

were found between teachers' use of praise and children's

feelings of competence and achievement. A negative correla-

tion significant at the .05 level was found between the

informational dimension of praise and children's intrinsic

motivation. Because informative praise occurred infre-

quently the negative correlation between informative praise

and children's intrinsic motivation must be interpreted

cautiously. Trends in the data, although not significant,

suggest there may be some differences between the use of

praise by control- and autonomy-oriented teachers. Control-

oriented teachers used more praise statements, more tangible

reinforcers, and more conduct-specific praise. These trends

as well as the small sample of subjects in the present

study suggest the need for further investigation.



Praise as a teacher behavior has been receiving mixed

reviews in the research literature. In a survey of 200

teachers, Zahorik (1977) found that praise was rated as

the single most important teacher behavior for facilitating

learning and elaborate praise was rated as the second most

important. Sandefur and Adams (1976) conducted a five-year

longitudinal study of the teaching effectiveness of teacher

education graduates from Western Kentucky University.

In their Evaluation of Teaching Report based on these data,

Sandefur and Adams stated, "good teachers use significantly

more praise and encouragement for the student" (p. 72).

Dunkin and Biddle (1974) reviewed a series of studies each

of which indicated that praise increased achievement.

Support for praise also can be found in numerous teacher

education texts, experimental studies, and expert opinion


Other studies, however, question the effectiveness

of teacher praise. Dershimer (1982) summarized previous

praise research in her article on pupil perceptions of

praise. She stated that several of the large correlational

studies of teacher effectiveness have concluded that teacher

praise is not a strong predictor of effective teaching.

Peng and Ashburn (1978) stated that there is little evidence

to show the importance of teacher praise. In a meta-

analysis of praise literature, Wilkinson (1980/1981) con-

cluded that teacher praise has very little relationship to

achievement. Soar and Soar (1978) stated that

positive affect [including praise] related
more consistently negatively to gain than
any other measure . it seems clear that
the "happy classroom" is not only not
necessary, but may be a liability for pupil
gain in achievement. (p. 95)

Based on findings like Soar's, Brophy (1981a) main-

tained that teacher praise was not effective and should

remain infrequent. In fact, he saw no necessity for praise

at all: "Students don't need praise to master the

curriculum, to behave or to develop healthy self-concepts"

(p. 21).

To praise or not to praise is obviously an unsettled

question. One problem with much of the research to date

is the acceptance of praise as a uni-dimensional variable.

Studies have grouped a multitude of events, attitudes,

and phrases under a single heading--praise. Rather than

continuing to view praise as a unitary phenomenon, it seems

important to acknowledge the many factors that combine

to vary the effects of praise on students. In fact, this

seems to be exactly the direction taken by several

researchers in their praise studies. For example,

investigators have found that the effects of praise vary by

factors in the situation, such as expectancy and

contingency, and by certain pupil characteristics, such

as locus of control, age, and SES (Boggiano & Ruble, 1979;

Danner & Lonky, 1981; Evertson, 1975; Swann & Pittman,


Another important dimension of praise receiving atten-

tion is the characteristic of the praiser, or rewarder.

This particular approach to studying the effects of praise

comes from the intrinsic motivation literature. Specifi-

cally, it follows from Deci's (1975) cognitive evaluation

theory. Deci proposes that every external reward, such

as praise, has both a controlling aspect and an informa-

tional aspect. The function of the controlling aspect

is to bring about a particular behavioral outcome in the

recipient. The function of the informational aspect is

to convey information about one's competence at a target

activity. The aspect of praise which is more salient to

recipients will have an effect on their intrinsic motiva-

tion. If the controlling aspect is more salient, intrinsic

motivation decreases; if less salient, intrinsic motivation

is maintained or enhanced.

Characteristics of teachers may be among the factors

that determine which types of praise administered by teach-

ers will be more salient to students (Deci, 1975, Deci &

Ryan, 1981). Teachers who are more control-oriented may be

likely to praise in controlling ways that will undermine

children's intrinsic motivation. Teachers who are more

oriented toward supporting autonomy may be likely to praise

in informational ways that enhance children's feelings of

competence and intrinsic motivation (Deci, Nezlek, &

Sheinman, 1981a). Viewed in this light, praise could func-

tion either as a detrimental or as a valuable educational

tool, depending on the orientation of teachers.

Statement of the Problem

The purpose of the present study was to investigate

informational praise and controlling praise by observing

teachers' use of praise in classrooms. Specifically, three

questions were explored:

1. Are there significant relationships between fre-

quency of various types of teacher praise and

teacher orientation toward controlling versus

supporting autonomy in children?

2. Are there significant relationships between fre-

quency of various types of teacher praise and

children's feelings of competence, intrinsic

motivation, and achievement?

3. Are there significant relationships between

teacher orientation toward controlling versus

supporting autonomy in children and children's

feelings of competence, intrinsic motivation,

and achievement?

Definition of Terms

Teacher praise. Teacher praise is a positive evalua-

tive reaction which goes beyond the level of simple

affirmation by verbally complimenting the child ("Good,"

"Fine," "Wonderful," etc.) and/or by accompanying verbali-

zation of positive feedback with expressions or gestures

connoting excitement or warmth. As used in this study,

praise is measured by behaviors on the Teacher Praise

Behavior Checklist (Appendix B).

Control-oriented teacher. A control-oriented teacher

is a teacher who exhibits the characteristic of being con-

trolling with children as indicated by a low or negative

score (below 6.98) on Deci's The Problems in Schools" Ques-

tionnaire: A Measure of "Adults' Orientations Toward Con-

trol Versus Autonomy with Children (Deci, Sheinman,

Schwartz, & Ryan, 1981b).

Autonomy-oriented teacher. A teacher who exhibits the

characteristic of supporting autonomy in children as indi-

cated by a higher score (above 6.98) on Deci's instrument

is an autonomy-oriented teacher.

Controlling praise. Teacher praise statements which

initiate a change in the students' perceived locus of

causality process are labelled as controlling praise.

Dimensions of this category are listed in Chapter III.

Informational praise. Teacher praise statements which

initiate a change in the students' perceived competence

process are labelled as informative praise. Dimensions of

this category are listed in Chapter III.

Feelings of competence. The feelings of cognitive,

social and physical competence and general self-worth of

the child are measured by Harter's Perceived Competence

Scale (Harter, 1979).

Intrinsic motivation. The motivational orientation of

the child in the situational context of classroom learning

is measured by Harter's Scale of Intrinsic Versus Extrinsic

Orientation in the Classroom (Harter, 1980).

Achievement. The achievement of the child is measured

by the Metropolitan Achievement Test, Spring administration,


Need for the Study

Cognitive evaluation theory (Deci, 1975) seems to

provide a feasible framework from which to approach the

question of the effectiveness of teacher praise. It seems

likely that previous studies of teacher effectiveness have

included both control-oriented and autonomy-oriented

teachers. If, as Deci et al. (1981a) suggest, teacher

orientation varies the effect of praise on the learner,

this may account for the inconsistent results of previous

studies. The praise of control-oriented teachers may have

undermined students' intrinsic motivation and achievement,

whereas praise of autonomy-oriented teachers may have

enhanced intrinsic motivation and achievement. The combina-

tion of these opposite effects of teacher praise on students

may account for the weak and mixed direction correlations

between praise and achievement reported in the literature

(Brophy, 1981a; Evertson, 1975; Wilkinson, 1980/1981).

Cognitive evaluation theory (Deci, 1975) suggests

that when children are informationally praised, their feel-

ings of competence are enhanced. Because they feel more

competent, at a task for example, their intrinsic motivation

to perform the task increases. Eventually, because they

are motivated to perform the task more often, their achieve-

ment increases. When children receive controlling praise,

however, their intrinsic motivation is undermined. They

may be less motivated to engage in learning tasks and their

achievement may decline.

Although the literature on cognitive evaluation theory

suggests that informational praise will increase feelings

of competence, intrinsic motivation, and, hence, achieve-

ment, no studies have been done to investigate this rela-

tionship directly. Many studies have been conducted to

establish the relationship between praise and achievement,

but because praise has been treated as a uni-directional

variable, little meaning can be assigned to the results.

Studies are needed which look at praise in terms of its

many factors in order to determine the relationship between

praise and achievement.


Several studies have been done that investigate aspects

of the relationship between one factor of praise (informa-

tional praise) and intrinsic motivation, feelings of com-

petence and achievement. In an experimental investigation

Pittman, Davey, Alafat, Wetherill, and Kramer (1980) found

that informational verbal rewards enhanced interest in a

task (Soma puzzles), whereas controlling verbal rewards did

not. The subjects were 84 college-age males and females.

Deci et al. (1981a) conducted a field study in fourth-

through sixth-grade classrooms. They considered the rela-

tionship between teacher characteristics (control vs.

autonomy orientation to children) and the competence and

intrinsic motivation of students. They found significant

positive correlations between teachers' orientation scores

and both children's measures. That is, teachers who were

more autonomy-oriented had students who were more intrin-

sically motivated and felt more confident. Deci et

al. (1981a) and Deci, Sheinman, Schwartz, and Ryan (1981b)

did not, however, observe teachers' use of praise in the

classroom. They merely stated the assumption that teachers'

orientations influenced the type of praise the teachers

used in the classroom.

Partial support for the distinction between informa-

tional and controlling praise can be found in an analysis

of praise articles written by several authors. For example,

although Evertson's (1975) results indicate that the effec-

tiveness of praise varies by SES of the student, she

cautions against telling teachers not to praise their

students. She says praise should be individual, genuine,

specific, and delivered in private. From Deci's perspec-

tive, Evertson's recommendation can be seen as calling

for informational praise rather than controlling praise.

That is, praise which is specific and communicated privately

to an individual gives that student information about

his/her competence, and in addition cannot be interpreted

by other students as trying to control their behavior (by

implying they should imitate the praised student).

Brophy's (1981a) functional analysis of praise also

adds support to Deci's formulation. Throughout his analy-

sis, Brophy refers to controlling the student. He states

that at its optimum, praise acts as reinforcement. That

is, the function of praise ought to be as a reinforcer,

which is used to control behavior. He uses as an argument

against praise that it reduces the intrinsic motivation

of students. From Deci's perspective, Brophy is addressing

controlling praise.

Although Brophy (1981a) states that praise should

function as reinforcement, he suggests that it often func-

tions in numerous other ways in the classroom. The actual

function of praise in the classroom is an area of great

concern to teachers and teacher educators, but it is also

an area with little empirical data. As Zahorik's (1977)

survey indicates, teachers highly value praise and use it

to facilitate learning. Teacher education courses are

reported most frequently as the source of teachers' praising

behavior (Zahorik, 1977). In most courses and teacher

education texts, however, praise is considered a uni-

dimensional variable. The effect of praise is assumed to be


Cognitive evaluation theory (Deci, 1975) proposes

that informational praise is beneficial but controlling

praise is not. This suggests that some instructional prob-

lems exist. Teachers who praise in controlling ways may

assume their behavior is enhancing learning when, in fact,

their praise is counterproductive. More research is needed

on the function of praise so that teachers can develop

more effective praising strategies. As an important source

of teachers' praising behavior, teacher educators are also

in need of these empirical data on the function of praise.

More knowledge of praise function would help determine

the most appropriate training and workshop content for

teachers and student teachers who praise incorrectly or

not at all. The distinction between informational and

controlling praise provides one fruitful approach to study-

ing the functions of praise.

Although partial support for Deci's distinction between

informational and controlling praise exists, the observa-

tional approach of this investigation was needed. A field

study, that has primary students as subjects, that cata-

logues the naturally occurring differences between teachers'

use of informational and controlling praise, and that

measures the relationship between teachers' praise behaviors

and children's intrinsic motivation, feelings of competence,

and achievement combines aspects of several different

studies that provide only partial support for the effec-

tiveness of informational praise.

Limitations of the Study

This study was designed as a descriptive, exploratory

study. The analyses were correlational and, therefore,

no causality was inferred. The generalizability of this

study is restricted due to the moderate sample size of

22 classrooms. Although small, this represents two-thirds

of the third-grade classrooms in Alachua County, Florida.

Controlling for age and maturation variables outweights

the small sample size deficit.

An observation instrument, the Teacher Praise Behaviors

Checklist, developed by the researcher for the present

investigation, was used to collect data on the praise styles

of third-grade teachers. The use of an instrument

restricted the richness of the data collected. Because

there were four observers, a uniform and consistent method

of data collection was needed. A further limitation was

the use of four observers. Inter-rater agreement was estab-

lished before data collection to minimize variability of

observation data due to varying degrees of observer



Finally, although the individual was the unit of analy-

sis for the teacher's orientation measure, the classroom

was the unit of analysis for all children's measures.

Pooling classroom data in this way ignores individual stu-

dent differences. It was necessary, however, because

student data in a classroom are not independent.


This study investigated whether a relationship exists

between teacher orientation and frequency of teacher praise

behaviors. In addition, this study investigated the rela-

tionship between frequency of praise behaviors and chil-

dren's feelings of competence, intrinsic motivation, and

achievement. The relationship between teacher orientation

and the three children's measures was also studied. This

study was designed to help clarify the inconsistent pattern

of results obtained in praise studies due to the research-

ers' treatment of praise as a uni-dimensional variable.

The findings from this study provide teachers with a better

understanding of the relationship between their praise

statements and children's intrinsic motivation, feelings

of competence, and achievement.



Praise as a teacher behavior is highly regarded by

practitioners (Zahorik, 1977), but its effect on learning

gains is questioned by current researchers (Brophy, 1981a,

1981b; Evertson, 1975; Soar & Soar, 1982; Wilkinson,

1980/1981). The current review of the literature related to

the effects of praise on children reveals studies with

three major outcome measures: achievement, intrinsic inter-

est, and competence. The results are often contradictory.

Some studies find praise facilitates learning, while others

find praise undermines learning.


In an attempt to clarify the findings from a number

of studies which include praise variables, three meta-

analyses have been conducted. A meta-analysis integrates

the findings from single studies and thus produces more

powerful and often more meaningful statistical results

(Glass, 1978). Even the more powerful technique of


meta-analysis yields unclear findings, however. Gage (1978)

and Lysakowski and Walberg (1981) conclude that praise has

positive effects on attitude and achievement, while

Wilkinson (1980/1981) concludes that teacher praise has very

little relationship to student achievement.

Functional Aspects of Praise

Perhaps the confusion due to confounding of the effects

of different types of praise on achievement, intrinsic

interest, and competence can be alleviated by viewing praise

as having multi-functions. The literature yields more

consistent results when the studies are grouped according

to the functional aspects of praise. Several researchers

(Bardwell, 1981; Deci & Ryan, 1981a; Deci et al., 1981a;

Meyer, Bachman, Bierman, Hemplemann, Ploger, and Spiller,

1979; Pittman et al., 1980) have investigated different

functions of praise, particularly informative and control-

ling functions. Deci and Ryan (1980) suggest that the

function of praise is determined by factors in the situa-

tion, rewardee, and rewarder.

The present review of the literature related to the

effects of praise on children covers four areas. These

include praise and achievement, praise and intrinsic motiva-

tion, praise and competence, and the functional aspects

of praise.

Praise and Achievement

The current tradition of studying the effects of praise

on the performance of primary school children was begun

by Elizabeth Hurlock in the early 1920s. Hurlock (1925)

asked the following research question:

In a classroom, do the children who constantly
receive praise for their work show more improve-
ment from day to day than do the children who
are reproved or who are completely ignored?
(p. 146)

Hurlock's approach involved dividing primary classrooms

into groups of children who received either praise, reproof,

or no comments following a test. When retested she found

that the praise and reproof groups achieved more than the

control groups with the greatest improvement shown by the

praised groups.

Investigators continue to study the same question

today, although the research paradigms are more varied.

The vast number of studies involving praise and achievement

can be classified as either experimental classroom studies,

laboratory studies, or large correlational studies of

teacher behavior.

Experimental Classroom Studies

Kennedy and Willcutt (1964) reviewed more than 50

years of experimental classroom studies of the effects

of praise and blame on the performance of school children.

Of the 33 studies reviewed, most concluded that praise

produced somewhat higher achievement than blame or control

conditions. Problems with the research included nonrandom

assignment of students to experimental conditions, no con-

trol for practice effects in test-retest situations, and

the ambiguity of praise statements used.

Thompson and Hunnicutt (1955) reported a study of

work achievement of fifth-grade pupils. They used five

different classrooms, six alternative forms of a cancella-

tion test, and specifically stated the praise and blame

conditions used. When the data were analyzed across all

subjects, either praise or blame was more effective than

no incentive.

Laboratory Studies

In a review of laboratory studies, Barringer and

Gholson (1979) compared verbal feedback with tangible

reinforcers. They found that learning was differentially

affected by various combinations and types of feedback.

In general, the laboratory findings indicated that verbal

feedback was more efficient for teaching children than

was tangible feedback. Two types of verbal feedback pro-

duced performance differences. Praise for correct answers

and silence for incorrect responses (right-blank combina-

tion) produced low efficient performance during acquisition

and little resistance to extinction. In contrast,

wrong-blank combination led to a more rapid acquisition

and greater resistance to extinction.

Current examples of studies in the laboratory tradition

found that groups who received praise showed some perform-

ance increment over control groups. Parnes (1973) found,

in a sample of 220 first graders randomly assigned to dif-

ferent incentive conditions, that the praised group per-

formed significantly better than the controls on a design

copying task. Garcia (1980) in a study of 60 third- and

fourth-grade Chicano students found that task-oriented

praise produced significantly more question-asking behavior

than either person-oriented praise or control group condi-

tions. Williams (1981), however, found that praised groups

did not perform significantly better than control groups

on a praised-associate memory task. Although Williams

performed his experiment in classrooms, his study is clas-

sified with the laboratory studies because it shares several

weaknesses with other laboratory studies.

In general the laboratory studies found that groups

who received praise showed some performance increment over

other conditions and control groups. The laboratory studies

often lacked ecological validity, however, due to novel

situations and tasks, experimenter effects, and short dura-

tion of treatment.

Classroom Observational Studies

Another approach to studying teacher praise and student

achievement has been to observe actual behavior in large

numbers of intact classrooms and correlate teacher behavior

to student outcomes. These large correlational studies

of teacher behavior are identified as process-product

research. In contrast to the generally positive influence

of praise on performance found in the experimental and

laboratory studies, the large correlational studies of

teacher behavior have yielded a generally negative view

of the effect of praise on the achievement of primary age

students. The findings from classroom observation studies

must be heeded because of their greater ecological valid-

ity. That is, these studies used actual teachers and

learners in real classroom settings with no outside manipu-

lations. What they gain through ecological validity may

be lost through threats to internal validity, however.

In addition, Wilkinson (1980/1981) states other limitations

that include data collection (frequency counts of praise

ignore many context variables), mean gain scores (means

ignore important individual differences), and teacher

behavior (the focus is on the teacher and the learner is


An example of a large correlational study that included

praise as a teacher variable was reported by Evertson

(1975). Praise data were collected over two years. In the

first year praise data were collected for 10 hours each in

31 second- and third-grade classrooms. In the second year

praise data were collected for 30 hours each in 28 second-

and third-grade classrooms. The process measure used was

an expanded version of the Brophy-Good Dyadic Interaction

System. Praise variables included positive evaluative

reactions following public response opportunities, during

study-intiated contacts, during teacher- initiated contacts,

for academic work, and for behavior. The product measure

was the students' achievement assessed by the Metropolitan

Achievement Test.

Evertson found that the effect of praise on achievement

varied by SES of the student and context of the situation.

In low SES classrooms, praise was regularly but weakly

associated with achievement but was relatively unimportant

in high SES classes. Further, praise was found to be some-

what effective when teacher-intiated and during private

interactions for both low and high SES classrooms.

Wilkinson (1980/1981) included Evertson's (1975) study

in her meta-analysis of the effect of praise on achieve-

ment. Wilkinson used as her source of studies Medley's

(1979) review of process-product research. She chose

Medley's review because of the stringent quality criteria

Medley used to screen the studies he reviewed. Wilkinson

lists the four criteria as

1. The study from which a relationship
came had to be designed so that the
relationship was generalized to some

population of teachers larger than
the sample studied.

2. The relationship had to be both reliable
enough to be statistically significant
and large enough to be practically

3. The measure of teacher-effectiveness had
to be based on long-term pupil gains in
achievement areas recognized as important
goals of education.

4. The process measure had to specify the
behaviors exhibited in such a way that
they could be reproduced as desired.
(p. 12)

Medley included 14 large correlational studies con-

ducted between 1965 and 1976. Wilkinson used these studies

and one subsequent study (which Dr. Medley, in a personal

interview with Wilkinson, indicated met his criteria but was

published after his review). In the meta-analysis,

Wilkinson aggregated data from 1,051 classrooms from kinder-

garten through grade 12. Students from both high and low

SES classrooms were included. Wilkinson reported an overall

aggregated correlation for the data set of r = .076. She

concluded, therefore, that teacher praise had little, if

any, relationship to student outcome. She did find, how-

ever, that SES membership of students, grade level of

students, and subject area had effects on the praise-

achievement correlations. These effects will be discussed

in a later section of the current review.

In summary, it seems that praise has some connection

with student achievement as indicated by the generally

positive results of experimental and laboratory studies.

Because this connection has not been confirmed in studies

of actual classroom process, however, the praise-achievement

connection does not appear to be fully explored. Other

factors must be investigated which influence the effect

of teacher praise on student achievement.

Praise and Intrinsic Motivation

Interest in intrinsic motivation as an outcome

measure of praise studies is a relatively recent devel-

opment. The trend of measuring intrinsic interest after

verbal rewards stems from the research on tangible extrinsic

rewards. Although not tangible, praise can be considered

a form of extrinsic reward. Findings that extrinsic

rewards undermine intrinsic motivation for the rewarded

activity seem quite robust (Deci & Ryan, 1980). A decrease

in intrinsic motivation in the form of less liking for the

activity or reduced time performing the activity has been

demonstrated with prizes (Ross, 1975), awards (Lepper,

Greene, and Nisbett, 1973), money (Deci, 1971, 1972), food

(Karniol & Ross, 1977), and supervision (Pittman et al.,


In terms of the preschool and primary age child, Condry

and Koslowski (1979) reported that "a major disadvantage

of extrinsic reward systems is that they motivate the child

more to get the reward than to arrive at a complete under-

standing of the educational task at hand" (p. 227). A

child is described as intrinsically motivated if he/she

performs an activity for its own sake and externally moti-

vated if the activity is performed as a means to an end,

to obtain a reward (Ross, 1975). Intrinsically motivated

students eager to learn out of curiosity and desire seem

the ideal model for education. Deci et al. (1981b) raised

a serious question about how to design learning environments

that can use rewards without undermining students' intrinsic


Some of the first studies of the effects of reward

on motivation were done by Deci (1971, 1972). Deci used

a word game "SOMA" composed of a number of blocks that

could be arranged in a variety of patterns. A series of

experiments were done using college-age students. The task

was to solve the puzzle patterns presented. The dependent

measure was the amount of time spent playing with the puzzle

during an unsupervised session. Reward conditions included

a praise condition ("that's very good"; "that's much better

than average for this configuration"), money condition

($1.00 for each configuration), and control condition (Deci,

1971), and in a later series of experiments praise in com-

bination with the money condition for males and females

separately (Deci, 1972). The results indicated that money

decreased intrinsic interest, whereas praise increased

intrinsic interest for the 1971 studies and increased

intrinsic interest for males in the 1972 study.

These Deci (1971, 1972) studies provided the paradigm

for studying the effect of praise on intrinsic motivation:

1. provide subjects with an intrinsically

interesting task,

2. praise subjects for doing the activity,

3. assess the level of intrinsic motivation

relative to a control group by observing in

a free-choice situation,

4. or by having subjects rate the extent to

which they enjoyed the activity.

Although Deci's (1971, 1972) studies seem to have

powerful implications for education (praise increases

intrinsic motivation), they were done in laboratory settings

(Condry & Koslowski, 1979). The next prototypic study

in the area of intrinsic rewards and subsequent intrinsic

interest was done in a typical nursery school setting by

Lepper et al. (1973). Although they did not use praise as

a reward condition, they did include the important dimension

of expectancy. They found it was not the "good player"

award per se which decreased interest in a drawing task,

but the expectation of receiving the award before the task

was begun. Of the three conditions, expected reward,

unexpected reward, and no reward, only the expected reward

subjects played with the magic markers less than the other

two groups.

Anderson, Manoogian, and Reznick (1976) replicated

the Lepper et al. (1973) study with 72 four- and

five-year-olds in 8 kindergarten centers. In addition

to the award condition, Anderson et al. added a praise

condition and a money condition. They also included three

comparison groups who received no reward as controls for

time or history, the presence of the experimenter, and

the personal attention given the child by the experimenter.

The task, as in the Lepper study, was free drawing with

magic markers. The dependent variable was intrinsic motiva-

tion, defined as the amount of time spent drawing. The

rewards (praise, money, or award) were presented every

two minutes. Praise statements included "you are pretty

good at this," "you really did a good job," "that picture

is real nice." Anderson et al. (1976) found that money

and awards decreased intrinsic motivation, whereas praise

sharply increased intrinsic motivation.

Swann and Pittman (1977) further explored the finding

that praise tends to increase rather than decrease intrinsic

motivation. Swann and Pittman set up their experiment

to see if it was the unexpected aspect of praise that

increased intrinsic motivation. They did so by adding

an unexpected tangible award condition. The subjects were

65 male and female first, second, and third graders from

middle and lower class families. The task was a drawing

task, and there were three experimental conditions and

two control conditions. Rewards included a "good player"

award (as in the Lepper et al., 1973, study) alone and

in combination with either praise or an unexpected star.

The praise statement, given after the award, was "Wow,

that's a really good picture; you really are a fine

artist." The dependent variable was intrinsic interest

measured as amount of time spent drawing during a free

choice period. A one-way ANOVA performed on the data showed

a main effect for reward condition. The award plus verbal-

praise group spent more time drawing and were more likely

to choose drawing than the other groups.

Swann and Pittman (1977) conclude that these results

support Deci (1972). Praise can neutralize the effects

of contingent tangible rewards. Swann and Pittman point

out that it was not the mere presence of rewards but rather

what the rewards did to self-perceived motivation that

led to changes in intrinsic motivation.

In a study of 39 preschool children in their natural

classroom setting, Krantz and Scarth (1979) investigated

teacher praise and intrinsic interest. They included four

treatment conditions and one control condition. The depen-

dent variable was measured by amount of time spent with

a target activity. The treatment conditions included com-

binations of teacher proximity, praise, and prompting.

The praise conditions resulted in more time persisting

with a task.

Rather than looking at one age level, Sarafino and

Stinger (1981) examined the effects of material and praise

rewards on intrinsic interest developmentally. Their sub-

jects were 28 kindergarten boys and girls and 28

fourth-grade boys and girls. Sarafino and Stinger chose

kindergarten and fourth grade because these two age levels

have been shown by Witryol (1971) to prefer different

rewards. In a survey of scaled reward preferences, Witryol

found a reversal occurs between kindergarten and third

grade. Kindergartners show a distinct preference for

material rewards (nickel, charm, bubble gum) over praise,

while third to sixth graders prefer praise.

The task Sarafino and Stinger used was giving "funny

endings" to riddles. Intrinsic interest was measured by

the number of unfinished riddles children chose to take

home. Each age level and sex was divided into two reward

conditions; half received a nickel and half received

praise. The fourth graders who received praise chose to

take significantly more riddles home than those who received

money, whereas the kindergartners who received money chose

to take more riddles home than the praise group, although

this difference was nonsignificant.

This study, although interesting because it attempted

a developmental perspective, has several problems. There

is a discrepancy in results between Sarafino and Stinger

(1981) and Anderson et al. (1976). The Sarafino and Stinger

study found a trend that suggested for kindergartners that

money increased intrinsic interest and praise reduced it.

Anderson et al. found that money decreased intrinsic inter-

est and praise increased it. Four factors in Sarafino

and Stinger's study make its findings questionable. First,

the Anderson et al. task, drawing, was more age appropriate

than generating multiple riddle endings. Second, Anderson

et al. praised every two minutes whereas Sarafino and

Stinger praised only at the end of the six-minute task.

Third, Sarafino and Stinger's instructions to the praise

group had a strong evaluative tone.

When you finish, I will tell you if your
endings are funnier than the ones you gave
before and if they are funnier than those
given by the children from the other
schools whom I have asked. (p. 295)

According to Deci and Ryan (1980), evaluation of subjects

lends an extrinsic character to praise and is often experi-

enced as highly controlling and aversive. The fourth prob-

lem with the Sarafino and Stinger (1981) study is the

absence of a control group. Because of these problems,

more weight must be given to Anderson et al. (1976) and

the enhancing effect of praise on intrinsic motivation.

Another study which investigated the effect of praise

on intrinsic motivation in a developmental perspective

was Danner and Lonky (1981). They reported a series of

two experiments. In Experiment I, they empirically estab-

lished what are intrinsically motivating tasks for three

developmental levels of children. Danner and Lonky divided

90 four- to ten-year-old children into three cognitive

ability groups based on their ability on a series of

Piaget's classification tasks. All three groups spent

most time with classification tasks that were just beyond

their ability levels. When asked, they also rated these

activities as most interesting. In Experiment II the task

was the same for all 90 children; thus the task was either

at, above, or below their predicted levels of intrinsic

interest. The task was a class inclusion classification

task. Based on their interest in Experiment I, this task

was highly motivating for Group 2 (just beyond their ability

level) and less motivating for Groups 1 (too difficult)

and 3 (too easy). Three reward conditions included a good

work certificate, praise, or no reward. There were 3 abil-

ity groups X 3 conditions groups or 9 groups. Praise state-

ments were given five times per session to each child and

included "That's the best work you've done so far," "Very

fine work," "Not many other children at your age have done

so well." Praise had a generally positive effect on the

children's motivation; however, it did not have a uniform

effect on all children.

Danner and Lonky (1981) thus showed a facilitating

effect of praise on intrinsic motivation for young children

as did Anderson et al. (1976), Swann and Pittman (1977),

and Krantz and Scarth (1979). Sarafino and Stinger's (1981)

methodological problems made their study less compelling

in its finding that money was more effective than praise

for kindergartners. In his review of the effects of extrin-

sic reward on intrinsic motivation Bates (1979) also con-

cluded that praise can be beneficial to intrinsic motivation

provided that it is unambiguously related to task perform-


In summary, praise seems to have an enhancing effect

on intrinsic motivation. This suggests that praise may

be a solution to what Bates (1979) calls

the apparent inability of the American
educational system to preserve and enhance
the interest in exploration and learning
that seems to be intrinsic in most children
when first entering school. (p. 557)

The major problems with the evidence on praise and intrinsic

motivation to date are the low numbers of studies and the

lack of ecological validity due to novel situations and

tasks, experimenter effects, and short duration of treat-

ment. What is needed are investigations into the effect

of praise on intrinsic motivation under normal classroom

conditions with real academic tasks.

Praise and Competence

Feelings of competence have been postulated to be

precursors of intrinsic motivation (Deci and Ryan, 1980;

Karniol and Ross, 1977). The standards with which a person

evaluates his or her competence are often extrinsically

or socially defined. Karniol and Ross (1977) thus argue

that extrinsic factors which increase one's feelings of

competence should in turn generate intrinsic interest in

the activity. Thus, when rewards provide the standard

by which an individual evaluates the quality of his/her

performance, intrinsic motivation increases. Deci (1975)

suggested that perceptions of competence are minimized

through the use of tangible rewards and maximized by the

use of praise.

Few studies manipulate both reward condition (praise

vs. tangible) and competence feedback. This section covers

three categories of experiments that each contain an aspect

of the relationship of praise to feelings of competence.

The first category covers experiments which manipulate

competence feedback but contain no praise variables. The

second category covers studies which manipulate the assumed

competency information component of the praise statements.

The third category covers experiments which investigate

praise and one aspect of feelings of competence,


Extrinsic Reward and Competence

Five experiments address the question of the connection

between extrinsic reward and competence. The first such

experiment was performed by Karniol and Ross (1977). These

investigators manipulated the competence information

imparted to the 57 subjects, aged 4-9, by making the reward

contingent or noncontingent on quality of performance and

by giving the children bogus social comparison feedback.

Their assumption was that contingent rewards impart more

information about one's competence than do noncontingent

rewards (the more rewards, the more competent). The rewards

used were marshmallows; the dependent measure was intrinsic


interest measured by the amount of time spent on the target

activity; and the task was a slide game. They found no

significant main effect for age but a significant effect

for treatment. Children, who received contingent rewards

or no reward and were told they performed well compared to

other children, showed increased interest in the task but

showed decreased interest when told they had performed

poorly. Karniol and Ross (1977) concluded that the compe-

tence information imparted by the rewards affected subse-

quent interest.

Enzle and Ross (1978) also studied the effect of con-

tingency of reward on competence information imparted.

They used 72 male college students in the puzzle-solving

paradigm initiated by Deci (1971, 1972). They used two

dependent measures of intrinsic interest, time spent on

puzzles during free time, and expressed enjoyment of the

task on a nine point rating scale. Enzle and Ross's results

agree with those of Karniol and Ross. Subjects who received

performance-contingent rewards (rewards dependent on quality

of performance) increased time spent on a task and expressed

more enjoyment than the time spent and enjoyment expressed

by subjects who received no reward or simply task-

contingent rewards (rewards dependent on simply engaging

in the task). Karniol and Ross concluded that performance-

contingent rewards signalled task competence and therefore

increased intrinsic interest in the subjects.

Schunk (1983) compared the performance of 36 children,

aged 8-11, on division problems. The children received

didactic instruction in division operations and were ran-

domly assigned to one of three groups: performance-

contingent rewards (rewards dependent on quality of per-

formance), task-contingent rewards (rewards dependent on

engaging in the task), or unexpected rewards. The dependent

measures were division skill attainment and self-efficacy

judgments. Schunk's results are consistent with those

of Karniol and Ross (1977) and Enzle and Ross (1978).

Subjects who received performance-contingent rewards

exhibited significantly higher division skill and made

significantly more efficacious judgments about their divi-

sion skill than either the task-contingent or unexpected

reward groups. Schunk discusses these results in terms

of Bandura's theory of self-efficacy as well as Deci's

cognitive evaluation theory. The performance-contingent

children viewed the rewards as conveying information about

their competence, whereas the task-contingent children

viewed the rewards as controlling.

Boggiano and Ruble (1979) referenced both Karniol

and Ross (1977) and Enzle and Ross (1978) as research sup-

port for the idea that "rewards made contingent on success-

ful performance provide cue value regarding competency

at a task, thereby maintaining intrinsic interest" (p.

1463). Their experiment was designed to test the competency

hypothesis by a) manipulating the availability of direct

information about competence and by b) means of a develop-

mental analysis. Their subjects were 147 middle-class

children, half from two private nursery schools (mean age

4.6) and half from two public elementary schools (grades

3 through 5, mean age 9.11). The task used was a hidden

picture game, the reward was candy, and the dependent meas-

ure was the amount of time spent playing with the target

activity during a free play period. There were two levels

of two conditions in the design: reward condition (either

task- or performance-contingent) and social comparison

condition (either relative competency or incompetency).

A control group received no reward or social comparison


Boggiano and Ruble found age effects, contingency

effects, and an age by social comparison interaction.

For the younger children, they found that comparative infor-

mation had no effect, but the task-contingent reward

decreased interest compared to the performance-contingent

and control conditions. The older children showed a sharply

different data pattern. For them, the contingency of reward

had no effect, but the comparative competency information

did. Task interest increased for those who received com-

petency information and decreased for those who received

incompetency information compared to the control groups.

Boggiano and Ruble (1979) concluded that the preschool

children interpreted competence information in terms of

absolute but not relative standards. They also concluded


that the results provide "strong support for the competency

hypothesis by showing that social comparison, providing

highly direct and unambiguous information about one's level

of competence, superseded reward contingency and its effects

on intrinsic interest in older children" (p. 1467).

Another study suggested that it is not contingency

per se, but whether rewards provide information about the

subjects' competence that determined how subjects reacted

to extrinsic rewards. Rosenfield, Folger, and Adelman

(1980) used 118 female college students randomly assigned

to eight conditions (two levels of contingency, two levels

of competency information, two levels of reward, and two

control). The task was a crossword game and the dependent

measures were amount of time spent with the task in a free

period, willingness to repeat task in the future, and

expressed liking for the task. Their findings support

those of Boggiano and Ruble (1979). The presence or absence

of competency feedback was the crucial determinant of the

subjects' intrinsic motivation whereas there was no signifi-

cant difference between contingency conditions.

Both Rosenfield et al. (1980) and Boggiano and Ruble

(1979) extrapolated their results in their discussion sec-

tions to the subject of praise. They hypothesized that

praise increased intrinsic motivation because it imparted

competency information to the subjects. Although no direct

investigation of this hypothesis has been carried out,,

several studies seem to support this idea. If one makes

the assumption that more specific and descriptive praise

imparts more competency information than general praise,

the following studies add weight to the competency hypothe-


Competency Information Component of Praise

Bernhardt and Forehand (1975) compared the relative

effectiveness of two types of praise content, labeled and

unlabeled, on the task performance of 40 five-year-old

white lower- and middle-class children. The task was a

marble in the hole game. Examples of the praise statements

used were "good boy" (unlabeled) and "what a good boy you

are for picking up the blocks" (labeled). Their results

indicated that the children were significantly more respon-

sive to labeled than to unlabeled praise. The labeled

praise presumably conveyed more competency information

than the unlabeled praise.

Takler (1975/1976) investigated the effect of descrip-

tive praise versus general praise on the performance of a

monotonous motor task (crossing out circles) on 60 fourth-

grade students. He found no significant differences between

the two praise conditions and a control group. This cannot

be interpreted, however, as evidence against the informative

value of descriptive praise. A task that is dull and unin-

teresting is certainly not intrinsically motivating ini-

tially. Adding extrinsic rewards, such as praise, to a

dull task would not increase intrinsic interest (Deci &

Ryan, 1980).

Scheer (1976/1977) investigated the same question but

used a more interesting classification task with 50 fifth-

and sixth-grade students. He found the descriptive praise

group performed significantly better than either the general

praise or control groups.

Praise and Self-Esteem

Feelings of self-esteem are indicators of a person's

general feelings of competence (Harter, 1982). Two studies

indicate that praise increases subjects' feelings of self-

esteem. Garcia (1980) found that task praise increased

the 60 third and fourth graders' feelings of self-esteem

as measured by self-report. Brown and Goodall (1981) found

that systematic written feedback by a fifth-grade teacher

significantly improved her students' feelings of worth

and the classroom climate compared to a control classroom.

In summary, most of the studies reviewed in this sec-

tion provide indirect support for the connection between

praise and competence. Rewards influence intrinsic motiva-

tion because of the competency information they provide

the learner (Enzle & Ross, 1978; Karniol & Ross, 1977).

Praise is hypothesized to provide unambiguous competency

information (Boggiano & Ruble, 1979; Rosenfield et al.,

1980). Some support is found for the connection between

praise and competence in the superiority of descriptive

versus general praise (Bernhardt & Forehand, 1975; Scheer,

1976/1977). Support is also found in the enhancement of a

general measure of perceived competence following praise

(Brown & Goodall, 1981; Garcia, 1980). Direct investigation

of the relationship between praise and feeling of competence

is needed rather than relying on measures of intrinsic

motivation and extrinsic rewards, especially in real class-

room situations.

Functional Aspects of Praise

The effects of praise on achievement, intrinsic motiva-

tion, and competence can be better understood when one

considers how praise is functioning in the situation.

Many current educational theorists and researchers have

begun to unravel the dimensions inherent in the construct

of praise (Bardwell, 1981; Bates, 1979; Brophy, 1981a,

1981b; Deci, 1975, Deci & Ryan, 1980; Meyer et al., 1979;

Soar & Soar, 1982).

Four Theoretical Orientations

The function of praise is interpreted differently

depending on the theoretical orientation assumed. Bates

(1979) described three theories that could be used to

explain the effects of praise. Behaviorist theory

incorporates praise into the S R paradigm. That is praise

acts as a reinforcer and should increase the incidence

of behaviors it follows. Praise does not, however, always

increase behaviors (Deci, 1972; Pittman et al., 1980;

Sarafino & Stinger, 1981; Takler, 1975/1976; Wilkinson,


Self-perception theory is another orientation discussed

by Bates (1979). This theory explains praise as an extrin-

sic reward that leads to the attribution of an external

cause of the behavior and therefore the absence of an inter-

nal one. Several studies have clearly shown that praise

increases intrinsic motivation (Anderson et al., 1976;

Danner & Lonky, 1981; Swann & Pittman, 1977). A third

orientation pointed out by Bates (1979) is the overjustifi-

cation hypothesis. This orientation is an off-shoot of

the self-perception framework and was used by Lepper et al.

(1973) to explain their results. In the overjustification

hypothesis praise functions as an oversufficient reward

when applied to an interesting task. That is, praise would

shift the locus of control for performing a task from

internal to external. This is refuted by the same evidence

that refutes self-perception theory's explanation of praise.

A fourth theoretical orientation toward praise appears

to be more consistent with empirical evidence than the

previous three. Cognitive evaluation theory (Deci, 1972,

1975; Deci & Ryan, 1980) seems to provide a framework for

interpreting the discrepant results of praise studies.

Deci (1975) formulated the cognitive evaluation theory

to account for the effects of extrinsic rewards (including

praise) on intrinsic motivation.

Cognitive Evaluation Theory and Praise

Deci (1975) asserts that there are two psychological

processes through which praise can affect a person's intrin-

sic motivation. The first is a change in the perceived

locus of causality process. The perceived locus of causal-

ity process is based on feelings and perceptions of self-

determination (Deci & Ryan, 1980). When people are rewarded

in a controlling way for an activity, they perceive an

instrumental relationship between the activity and the

reward. The behavior which had been in the domain of the

intrinsic motivational subsystem moves to the domain of

the extrinsic subsystem.

The second process through which one's intrinsic

motivation can be affected is the change in perceived com-

petence process (Deci, 1975). [Deci originally included

feelings of self-determination in the second process but

has since stated self-determination is part of the first

process (Deci & Ryan, 1980).] When information increases

one's perceptions of competence, it should enhance one's

intrinsic motivation; when it decreases feelings of com-

petence, it should decrease intrinsic motivation.

Deci (1975) called these two processes Propositions

I and II of his cognitive evaluation theory. Proposition

III relates directly to praise.

Proposition III of Cognitive Evaluation Theory:
Every reward [including praise] has two aspects,
a controlling aspect and an informational aspect
which provides the recipient with information
about his competence. The relative salience
of the two aspects determines which processes
will be operative. If the controlling aspect
is more salient, it will initiate the change
in perceived locus of causality process. If
the informational aspect is more salient, the
change in feelings of competence . process
will be initiated. (p. 142)

Hence, cognitive evaluation theory accounts for contra-

dictory effects of praise depending on which of its aspects

--informational or controlling--is most salient to the

recipients. From Deci's perspective, then, praise has two

functions, a controlling function and an informational


Two Functions of Praise

Support for these two functions can be found in the

praise literature even when the researchers are from dif-

ferent theoretical orientations. For example, Brophy

(1981a) explored several functions of praise. Throughout

his analysis, Brophy referred to controlling the student.

He stated that, at its optimum, praise functioned as a

reinforcer. That is, praise should be used to control

the student. Brophy argued that praise decreased intrinsic

motivation of students. In terms of cognitive evaluation

theory, Brophy emphasized the controlling aspect of praise.

Bardwell (1981), like Brophy (1981a), questioned the

reinforcing value of feedback. She demonstrated through

an analysis of feedback delay that feedback was serving

an informational function and did not follow the rules

suggested by reinforcement theory. She used a school-

related learning task with subjects in grades 4, 6, and

8. The subjects studied, were tested, and received feedback

on the learning task. One-half of the subjects received

immediate and one-half received delayed feedback. Rein-

forcement theory suggests that delayed feedback facili-

tates retention but hinders acquisition. Bardwell demon-

strated that delayed feedback facilitated both acquisition

and retention, which are the results suggested by informa-

tion theory.

In a series of six experiments, Meyer et al. (1979)

investigated the degree to which praise provides information

about others' perceptions of an acting person's ability

depending on the difficulty level of a particular task.

Their results suggested that young children perceived the

praised person's ability to be higher than a person given

neutral feedback, regardless of the difficulty level of

the task. These data are reversed for subjects sixth grade

and older. This result, in addition to Boggiano and Ruble

(1979), suggests that younger children use concrete

information (rather than comparative or relative) in per-

ceiving competence.

Soar and Soar (1982) also stated the idea that praise

functions in a more complex fashion than simply reinforcing

behavior. Although they separated praise from positive

affect rather than controlling praise from informational

praise, Soar and Soar's analysis seems consistent with

Deci's (1975) with a few semantic changes:

A possible interpretation is that praise
[controlling praise] places authority and
control in the adult, which hinders pupil
development of independence, self-reliance
in thought and complex thought processes.
Positive affect [informational praise], on
the other hand . encourages pupil
independence, complex thought processes,
and more positive attitudes. With such
different results, it seems imperative to
separate praise controllingg praise] from
positive affect [informational praise]
conceptually, as measures in research, and
as desirable teaching behaviors. (p. 18)

Soar and Soar (1982) indicated the need to separate

positive affect (praise with an informational function)

from praise (praise with a controlling function). Deci

and Ryan (1980) indicated the same need in the form of

this question, "what determines which aspect of [praise]

will be more salient? [The] answer is that there are fac-

tors in (1) the rewardee, (2) the situation, and (3) the

rewarder that will determine which aspect of the reward

is more salient" (p. 68).

The following sections of this review contain praise

studies which fall into each of these three categories.

Factors in the Rewardee

Students differ in a number of ways. Many of these

factors probably affect the way they interpret and thus

respond to praise. The factors which have been investigated

with respect to praise are SES, age, sex, and locus of


SES. Several studies have included SES as a mediator

variable in studies of praise. Although the results of

these studies on the impact of students' SES on their

response to praise seem to differ, there is a methodological

pattern. The small studies failed to find that SES had

an impact whereas the larger studies reported consistent

effects of SES on response to praise.

Bernhardt and Forehand (1975), Heller and White (1975),

and Means, Means, Osborne, and Elsom (1973) all reported

no significant differences between middle and lower SES

groups of children in response to praise. The samples

in each study were small (the largest number in any group

was 45) which may have decreased power. The treatments

were short and novel which may not have given the manipu-

lated variables time to have effect.

The larger studies show a consistent praise by SES

interaction. Brophy (1979) found that teachers of high

SES, high ability classes were more challenging and demand-

ing and less apt to praise. Teachers of low SES, low abil-

ity classes were more encouraging and praised more.

Evertson (1975) found that in low SES schools, praise was

regularly but weakly associated with learning gains. Praise

was unrelated to learning gains in high SES schools. Soar

and Soar (1982) reported Rowe's finding that praise was

contingently dispensed to high ability (high SES) students

and was just as frequently but noncontingently dispensed

to low ability (low SES) students.

Wilkinson (1980/1981) found a higher correlation

between teacher praise and student achievement for students

from low SES backgrounds (r = .107) than for high SES

students (r = .028).

Thus, although the larger studies show that praise

had a higher correlation to achievement for low SES stu-

dents, the effect is small. The possibility cannot be

overlooked that the smaller studies (experimental designs)

present more accurate findings.

Age. Several studies have looked at the effects of

praise from a developmental perspective. The subjects

in single-age studies of the effects of praise and

extrinsic reward on performance cover a wide age span,

from preschool (Anderson et al., 1976; Boggiano & Ruble,

1979; Lepper et al., 1973; Krantz & Scarth, 1979) and

elementary school levels (Boggiano & Ruble, 1979; Danner &

Lonky, 1981; Karniol & Ross, 1977; Swann & Pittman, 1977)

to college populations (Deci, 1971, 1972; Enzle & Ross,


In general these studies found that praise enhanced

the intrinsic motivation of subjects across the

developmental span. Sarafino and Stinger (1981), however,

found that kindergartners who were praised expressed less

interest than those rewarded with money. The present

investigator considered this a problem in methodology. It

could also be that because of the directions given in the

praise condition, these children felt controlled and their

intrinsic interest thus decreased.

Younger children do differ in their interpretation

of competency information, however. They seem unable to

interpret relative or comparative information about com-

petency (Boggiano & Ruble, 1979; Meyer et al., 1979).

This suggests that concrete and specific praise statements

would be appropriate for younger children (third grade

and younger).

Wilkinson (1980/1981) found a difference in the

relationship of praise to student achievement for primary

(K-3) and intermediate (4-8) students. For primary students

the overall correlation was r = .118 whereas for inter-

mediate students it was r = .064. The same trend was seen

in the separate correlations for math and reading as well.

This may suggest that teacher praise sustains an informa-

tional function through the third grade and is perceived as

more controlling as children become older.

Sex. Like the variables discussed so far, sex of

the subjects seems to mediate the effect of praise. This


effect seems to interact with age, however. Sex differences

in the effects of praise have been demonstrated with adults

but not often with children (Deci, 1972, 1975). Only one

study was found which demonstrated differences in praise

effects due to the sex of the children (Lintner & Ducette,


In the Deci (1972, 1975) studies praise increased

the intrinsic motivation of males but decreased the intrin-

sic motivation of females. In the Lintner and Ducette

(1974) study of 285 students from third, fifth, and seventh

grades of lower middle class schools, praise significantly

improved the performance of males on an ambiguous coding

task but nonsignificantly improved the performance of

females in the positive direction.

In terms of its two functions, praise tends to be

more controlling for females and more informative for

males. For females a change in the perceived locus of

causality from internal to external was initiated whereas

for males an increase in feelings of competence was ini-

tiated. One explanation for these differences may be found

in the socialization processes of males and females in

the traditional culture (Deci, 1975). Girls may have

learned to be more dependent and interpersonally focused,

and boys may have learned to be more independent and

achievement focused (Deci & Ryan, 1980). This interpreta-

tion receives some support from the work of Dweck, Davidson,

Nelson, and Enna (1978). They reported several differences

in the evaluative feedback given to elementary age boys

and girls. Praise given to boys conveyed specific

information about the adequacy of their performance, whereas

praise given girls conveyed just the general attitude of the


Another explanation of the sex differences found in

the Deci (1972) study could be the interaction of sex of

the experimenter with sex of the subject. Bates (1979)

suggested this could be the case (the experimenter in the

Deci (1972) study was a friendly young male graduate stu-

dent). Some support for this idea comes from a study by

Rothbaum, Zigler, and Hyson (1981). In a study of 96 male

and female third-grade children from a middle class school,

Rothbaum et al. found that there was an interaction between

the sex of the experimenter and sex of the subject. They

found, however, that the children were significantly more

responsive to cross-sex adults in the adult praise condi-

tion. Rothbaum et al. (1981), therefore, do not support

Bates' idea. Actually, Deci (1975) ran another experiment

to clarify his 1972 results. In the follow-up experiment

he included male and female experimenters as well as sub-

jects. He replicated the 1972 results: intrinsic interest

increased for males and decreased for females. He found

no interaction with sex of experimenter.

Condry and Koslowski (1979) offered another explanation

for the sex difference found in the Deci (1972) study.

Tasks can be considered by the participants to be

sex-appropriate, and this affects their responses to the

task. It could be that SOMA puzzles were considered by the

participants to be masculine. Deci and Ryan (1980) sug-

gested the same explanation but indicated there may have

been a change in women from the earlier to the later part

of this decade. This is supported by a recent study that

shows no differential effects of praise for males and

females (Heller & Parson, 1981).

Locus of control. The clearest evidence for the media-

tion of the effect of praise by a factor in the rewardee

is shown for the locus of control variable. Praise

increases the intrinsic motivation of internal children and

decreases it for external children. According to cognitive

evaluation theory, the informational aspect of praise is

salient for internal children. Praise increases their

perceptions of competence and, hence, enhances their

intrinsic motivation. The controlling aspect of praise is

salient for external children. Praise causes a change in

their perceived locus of causality and a decrease in

intrinsic motivation.

Three of four studies which included a locus of control

variable support this interpretation. Thompson and

Hunnicutt (1955), Anderson (1977/1978), and Danner and Lonky

(1981) all found that praise enhanced performance or intrin-

sic motivation for internal children and decreased them

for external children. In addition, Anderson (1977/1978) in

a study of 84 fifth- and sixth-grade children included

an encouragement condition. She found that internal

children preferred praise over encouragement and could

more clearly differentiate the two conditions than could

external children.

Lintner and Ducette (1974), however, found that exter-

nal children were more responsive to praise than internal

children on an ambiguous coding task. This could be due

to a problem in the study, however. It could be that the

children perceived the sorting task as dull and external

children only responded to the praise which functioned

as a controlling external reward.

Factors in the Situation

The conditions under which praise occurs seem to have

important influences on its effects. For example, Brophy

(1981b) and Evertson (1975) indicated that praise given

during teacher-initiated contacts was more effective than

praise given during student-initiated contacts. Three

factors which seem to be operating in the situation are

contingency, salience, and expectancy. These factors are

limiting conditions to the process in which a change in

perceived locus of causality leads to decreased intrinsic

motivation (Deci & Ryan, 1980).

Contingency. Several studies showed that contingent

rewards increase intrinsic motivation or performance (Enzle

& Ross, 1978; Karniol & Ross, 1977; Rosenfield et al.,


1980). Conversely, several studies showed that contingent

rewards decreased intrinsic motivation or performance (Deci,

1972; Swann & Pittman, 1977). The apparent discrepancy

in results was due in large part to a problem with the

use of the word contingent (Condry & Koslowski, 1979).

Some studies made the reward contingent on "doing the task"

while others made the reward contingent on the quality of

performance. Contingency in the behavioristic sense

indicates that the behavior is under the control of the

reward (Sharpley & Sharpley, 1981). Contingency in the

framework of cognitive evaluation theory has to do with

competence information. A contingent reward that conveys

direct and unambiguous competency information enhances

intrinsic motivation. A noncontingent reward that conveys

no competency feedback makes the control aspect of the

reward more salient and, hence, decreases intrinsic inter-

est. Rosenfield et al. (1980) showed that it was not the

contingency itself that was an important determinant of the

effect of reward on intrinsic interest. It was the presence

or absence of competency feedback. When greater rewards

indicated greater competence, intrinsic motivation

increased. When greater rewards did not indicate greater

competence, there was a decrease in intrinsic motivation

regardless of contingency.

Ryan, Mims, and Koestner (1983) also showed that infor-

mational aspects of rewards mediated the contingency

effects. They made performance-contingent rewards either

informational or controlling by varying the feedback state-

ments. The controlling performance-contingent-reward

subjects received feedback statements that were the same as

the informational performance-contingent-reward subjects

except that a "should"-related phrase was added. An example

was, "You did very well on that one, just as you should."

In addition to the two groups with performance-contingent

rewards that were informationally versus controllingly

administered, there were four other groups: two comparable

no reward groups who received the same feedback statements

and one no reward/no feedback group and one no feedback,

task-contingent reward group. The subjects were 96 college

students and the tasks were hidden-figures puzzles. The

dependent measure of intrinsic motivation was the amount

of time spent working on puzzles during a 6-minute free

choice period.

Ryan et al. (1983) showed that the informational feed-

back groups spent more time working on puzzles than the

controlling feedback groups and the no-feedback groups.

The informational feedback/no reward groups displayed more

intrinsic motivation than the comparable feedback

performance-contingent-reward group. These results are

consistent with Rosenfield et al. (1980). It is not the

contingency itself, but whether the rewards and feedback

are perceived as informational or controlling that affects

intrinsic motivation. Beyond this the Ryan et al. (1983)

study showed that performance-contingent rewards, like

all rewards, tended to lower intrinsic motivation relative

to no rewards if there is identical verbal feedback (compe-

tency information).

Salience. Ross (1975) performed an experiment with

60 preschool children in three nursery schools. The

task used was playing a drum. The treatment conditions

included a highly salient reward (a prize under a box in

full view), a nonsalient reward (prize merely mentioned),

and no reward (control). The dependent measure was

intrinsic motivation as measured by a) amount of time

spent playing the drum during free play and b) whether child

responded that the drum was the "most fun thing." The

salient reward condition significantly decreased intrinsic

motivation compared to the nonsalient and no reward groups.

Ross concluded that highly salient rewards led chidlren to

change their perceived locus of causality to external


Deci and Ryan (1980) countered, however, that the

salience of the reward was not the parameter that determined

whether rewards would decrease intrinsic motivation. The

important parameter was the salience of the informational

and controlling aspects of the reward rather than the sali-

ence of the reward per se.

Pittman et al. (1980) reported an experiment in which

they manipulated the functional aspects of praise. The

subjects were 84 college males and females randomly assigned


to the experimental and control conditions. The tasks used

were SOMA puzzles. The controlling praise given was

I haven't been able to use most of the data
I've gotten so far, but you're doing really
well, and if you keep it up I'll be able to
use yours. (p. 230)

The informational praise was

Compared to most of my subjects, you're doing
really well. (p. 230)

Pittman et al. found that compared to a no reward

control, the informational praise enhanced task interest

but the controlling praise did not.

Expectancy. Lepper et al. (1973) demonstrated that

the expectancy of receiving a reward undermined intrinsic

motivation rather than the good player award itself. To

test whether the enhancing effect of praise on motivation

was due only to its unexpected nature, Swann and Pittman

(1977) paired unexpected praise with a contingent reward

and compared its effect to an unexpected tangible reward (a

star) paired with a contingent reward. They found that the

praise group increased in interest whereas the star group


Deci and Ryan (1980) explained that the question of

expectancy can be understood within the context of the

relative salience of the two aspects of rewards. Again

it is the salience of the controlling aspect or the infor-

mational aspect of the reward (praise) that determines

whether it will increase or decrease intrinsic interest

and feelings of competence. All other things equal, when


rewards are expected they will be experienced as controlling

because one begins the activity in order to get the reward.

When rewards are unexpected, they will more likely be

experienced as an indication of good performance.

Factors in the Rewarder

The effects of praise do not appear to be uniform

across situations or across children. Praise effects seem

to vary by the situational factors of contingency, expec-

tancy, and salience. Praise effects also seem to vary

by the rewardee factors of SES, age, sex, and locus of

control. More research is needed in each area to clarify

the interaction between the function of praise (control

or information) and the situational and personological


In terms of educational practice, however, even if

the differential effects of praise were explicated for

a multitude of student characteristics, it would still

be difficult for teachers to praise differentially in the

classroom. There is usually one teacher for every thirty

children. To praise differently for every student would

be very difficult. Rather than concentrating on student

characteristics, it would seem beneficial to investigate

those factors in teachers which seem to influence whether

they praise informationally or controllingly. In this

way an attempt could be made to maximize the benefit of


teacher praise for most students by identifying the sources

of informational praise. Deci et al. (1981a, 1981b) have

begun investigating teacher characteristics in relation

to praise function. In addition, several researchers have

looked at related teacher attitudes and how students per-

ceive these teacher attitudes.

Teacher attitudes and praise. Silberman (1969) exam-

ined several different teacher attitudes toward their stu-

dents and the way in which these different attitudes

affected their classroom behavior. He interviewed 10

third-grade teachers to determine toward which students

they held attitudes of attachment, concern, indifference,

and rejection. He then observed the teachers' classroom

behavior. He designated four teacher behaviors as the

dependent variables: teacher-initiated contact, praise,

criticism, and acquiescence. Silberman found that teach-

ers' attitudes are generally revealed in their actions.

Different attitudes are translated into action in different

ways. Teachers gave more praise to those students for

whom they felt attachment and less praise to those for

whom they felt indifference.

Benninga, Guskey, and Thornburg (1981) also investi-

gated teacher attitudes and the way attitudes affected

behavior. They felt it was particularly important to exam-

ine the connection between teacher attitudes and behavior:

If attitudes and perceptions affect their
behavior and the roles they defined for
themselves, it is important to understand

these underlying beliefs, particularly since
they may have an impact on how teachers
behave toward pupils. (p. 66)

They measured the attitudes of 42 first-, second-, and

third-grade teachers. Teacher measures included authori-

tarianism, control, teaching self-concept, responsibility

for student achievement, and affect toward teaching. One

group, who liked teaching more, tended to be less authori-

tarian, felt less need to control students, and felt a

greater responsibility for positive student learning out-

comes. The other group, who had higher teaching self-

concepts, tended to be more authoritarian, were more

restrictive, and felt less responsibility for learning

outcomes. Although Benniga et al. (1981) emphasized the

need to examine connections between attitude and behavior,

they included no observation component. This study is

included in the current review as evidence that teachers

have measurably different attitudes toward children and


Deci et al. (1981a, 1981b) reported results of a field

study conducted in 36 fourth-, fifth-, and sixth-grade

classrooms in four elementary schools. The study was

designed to investigate the assumption that characteristics

of the teacher were among the factors that determined

whether the controlling or informational aspect of praise

was more salient. They predicted that if teachers were

more oriented toward controlling children, they would be

likely to praise in controlling ways that would undermine

children's intrinsic motivation. If they were oriented

toward supporting autonomy in children, they would be likely

to praise informationally in a way that would increase

children's perceived competence. The distinction between

the two types of teachers, controlling vs. supportive of

autonomy, has been indicated by Benninga et al. (1981) and

Deci et al. (1981b).

Deci et al. (1981a) looked for the connection between

teacher type and differences in children's expressed intrin-

sic motivation and competence. They measured 36 teachers'

orientation (control vs. autonomy) toward children. They

then gathered data on 610 children in these teachers' class-

rooms on several measures, two of which were perceived

competence and intrinsic orientation.

Deci et al. (1981a) found significant positive correla-

tions (from .29 to .56) between teacher orientation scores

and children's scores on scales of perceived competence

and intrinsic motivation. Deci et al. concluded that this

study showed a clear relationship between characteristics

of the teacher and intrinsic motivation and perceived com-

petence in the learner. The results could indicate that

praise will tend to undermine intrinsic motivation and

perceived competence when given by a teacher with a con-

trolling orientation but will tend to maintain or enhance

intrinsic motivation and competence when given by a teacher

who is autonomy oriented. Or the results could indicate

that the praise teachers generate differs depending on


their orientation toward controlling or supporting autonomy

in children.

Deci et al. (1981a, 1981b) assumed that when teachers

have different attitudes toward children, they behave dif-

ferently in the classroom. That is, they assumed that

control-oriented teachers praised differently from autonomy-

oriented teachers and that it was this difference in praise

and reward behavior that led to the differences measured

in their students. Deci et al. included no observational

component; however, so they offered no data to support

this assumption.

The Deci et al. (1981a, 1981b) studies measured the

teachers' personality variable, orientation toward control

versus supporting autonomy with children, that affected

whether their classrooms were more informational or more

controlling. Deci, Speigel, Ryan, Koestner, and Kauffman

(1982) manipulated conditions that led teachers to be more

controlling versus more autonomy-oriented with students.

When teachers were told that they were responsible for

their students' performing up to standards, the teachers

were more controlling. The Deci et al. (1982) study was

a laboratory study using 20 male and 20 female teachers.

Each teacher directed one student on a puzzle-solving activ-

ity. One-half of the male and one-half of the female teach-

ers were in the informational condition (no performance

standards specified). The other halves of the teacher

groups were in the controlling condition (performance

standards specified). The clearest differences between

teachers were that the controlling teachers talked more

and in more controlling ways.

It seems an important area of investigation to observe

the differences in control- and autonomy-oriented teachers'

behavior without any manipulation of their orientation

as in the Deci et al. (1982) study. It also seems important

to observe these differences in the classroom rather than

the laboratory. In this way, empirical evidence can support

the connection between teacher attitude and behavior that

Deci et al. (1981a, 1981b) assume.

Students' perceptions of teacher attitudes and praise.

Each of the studies reviewed in the previous section con-

tained a pupil perception component except the Deci et

al. (1982) study. An important issue to resolve is if

students perceive different teacher attitudes and behavior.

The research evidence indicates that students are sensitive

to and perceive teacher differences.

In the Silberman (1969) study, there was a significant

positive correlation between student perceptions of teacher

behavior and actual observations. Students were able to

predict better than chance the relative amounts of contact,

criticism, and acquiescence they received. They could

not predict with any accuracy, however, how much praise

they received.

Benninga et al. (1981) also included a pupil perception

measure in their study. They found that students could

discriminate differences in teachers. Although they did

not perceive fine distinctions between the two types of

teachers, students perceived controlling teachers as having

less rapport and interactional competence and less authori-

tarian teachers as having more rapport.

Deci et al. (1981a) measured children's perceptions

of classroom climate. They used these scores to investigate

whether children perceived a difference in classrooms of

control- versus autonomy-oriented teachers. They found

a significant positive correlation (.354) between teacher

orientation score and children's perception of classroom


The Silberman (1969), Benninga et al. (1981), and

Deci et al. (1981a) studies indicated that children were

sensitive to teachers' different attitudes. The following

two studies indicate that children are also sensitive to

teachers' use of praise. Meyer et al. (1979) investigated

how other children perceive praise given to a particular

child. They found that children's perceptions varied by

age. Children fifth grade and younger perceived praise

as absolute. That is, if a child was praised for performing

either a low ability or high ability task, that child was

perceived as having high ability. Older children and

adults, however, perceived people as having low ability

if they were praised for performing easy tasks.

Dershimer (1980, 1982) investigated pupil perceptions

of the functions of teacher praise as a part of a larger

sociolinguistic study. She observed 164 children in their

6 second-, third-, and fourth-grade classrooms. She found

several interesting patterns in pupils' perceptions of

teacher praise. Students perceived only 12 percent of

the praise that occurred. As the intensity of the praise

increased, however, there was a concomitant increase in

the proportion perceived. Those students who were high

in class participation and thus received more praise, per-

ceived praise as deserved. Those students who were low

in class participation and thus were an audience for praise

received by others, perceived praise as serving an instruc-

tional function.

Another interesting finding indicated that pupils

perceived praise as functioning not so much to reinforce

the individual student as to give information to the group.

This finding may be explained by the model of an integrated

view of classroom discourse taken from her larger study

(Dershimer, 1982): 1) teacher questions serve to identify

things one ought to know, 2) pupils respond to questions,

3) answers to questions inform pupils; one must attend

to pupil responses in order to know what should be known,

and 4) a pupil response which is praised is probably a

better presentation of information than one that is not;

so one must pay special attention to comments that draw

teacher praise.


Praise studies have typically looked at praise as

a unidimensional variable. This has led to confusion about

the effects of praise. Some studies (Anderson et al.,

1976; Garcia, 1980; Pittman et al., 1980; Swann & Pittman,

1977) reported positive effects of praise, some negative

(Evertson, 1975; Peng & Ashburn, 1978; Takler, 1975/1976;

Wilkinson, 1980/1981) and some mixed (Danner & Lonky, 1981;

Deci, 1972; Lintner & Ducette, 1974; Sarafino & Stinger,

1981). Using cognitive evaluation theory (Deci, 1975)

as a framework helps to explain the discrepant results

of the praise studies.

The effects of praise are different depending on which

aspect is more salient, the control or information dimension

(Deci & Ryan, 1980). Controlling praise tends to decrease

intrinsic motivation, whereas informational praise tends

to increase feelings of competence. Deci et al. (1981a)

suggest that there are factors in the rewardee, situation,

and rewarder that influence which aspect is salient.

Factors in the situation which tend to influence the

salient aspect include the contingency, expectancy, and

salience of praise. Factors in the rewardee include SES,

sex, age, and locus of control. Factors in the rewarder

include the orientation toward children, either controlling

or supporting of autonomy. In terms of the reality of


classroom practice, the most feasible factor to investigate

is the rewarder, or teacher.

There is some evidence to suggest that teachers' orien-

tations toward children influence the type of praise they

use and how children feel about themselves (Deci et al.,

1981a, 1981b). There have been no studies, however, which

have investigated this relationship with children younger

than fourth grade. There have also been no studies which

have included an observational component in the classroom.

The assumption that control-oriented teachers use praise

differently has not been empirically demonstrated.

There exists a need for more research on the function-

ing of praise. Specifically it seems necessary to investi-

gate further the connection between teacher orientation

and children's intrinsic motivation and competence. A

field study using primary children which includes an obser-

vational component would significantly increase the under-

standing of the effects of praise on children.


This study was designed to investigate the relation-

ships between teacher orientation and frequencies of various

types of teacher praise. Further, this study investigated

the relationships between teacher orientation and children's

feelings of competence, intrinsic motivation, and achieve-

ment. The relationships between frequencies of various types

of teacher praise and the three children's measures were also

studied. The following specific correlations were computed:

1. teacher orientation (a scale from autonomy to

controlling) with frequencies of various types

of teacher praise

2. teacher orientation with children's feelings

of competence

3. teacher orientation with children's intrinsic


4. teacher orientation with children's achievement

5. frequencies of various types of teacher praise

with children's feeling of competence

6. frequencies of various types of teacher praise

with children's intrinsic motivation

7. frequencies of various types of praise with

children's achievement.

Design of the Study

The design of the study was ex post facto. The inde-

pendent variables, teacher orientation and praise style,

were not manipulated. The naturally occurring attribute

of the teachers was measured and the naturally occurring

praise events in the classroom were recorded. Teachers

were measured using Deci's orientation measure, The "Problems

in Schools" Questionnaire. Their ranking reflected a con-

tinuum from a control orientation to an autonomy orientation

toward children. These orientation scores were then cor-

related with the teachers' frequency of various types of

praise in the classroom as recorded on the Teacher Praise

Behaviors Checklist, developed by the researcher. Thus, the

extent of the relationship between teachers' orientations

(attributes) and their use of praise in the classroom

(behavior) was determined.

These two teacher measures were then correlated with

the children's measures to determine the relationships

between teacher orientation and teacher praise and chil-

dren's scores on measures of feelings of competence, intrin-

sic motivation, and achievement. The children's measures

were considered the dependent variables.

The correlational-observational method outlined above

seemed the most appropriate design for investigating the

strength of relationships which may exist between teacher

orientation, teacher praise, and the selected student vari-

ables for several reasons. First, the intent of the present

investigation was exploratory. More data need to be gath-

ered on naturally occurring praise events in the classroom

before manipulation of variables is warranted. As Vasta

(1979) suggests, the findings from correlational studies

are most useful for indicating potential relationships which

then can be investigated using experimental methods. Gage

and Giaconia (1981) also state that the best experimental

studies of teaching are based on correlational findings

rather than on theory. Prior to the present study, Deci's

(1975) distinction between informational and controlling

praise was primarily a theoretical distinction. The present

study, as well as Deci et al.'s (1981a) study, help determine

which relationships are most useful to investigate experi-


Another reason for selecting a correlational design

was the need for greater ecological validity in praise

studies. Although Pittman et al. (1980) used an experi-

mental design to study the difference between informational

and controlling praise, their results have little appli-

cability to classrooms. They used a novel task, individual

testing, only one treatment period, and a contrived praise

statement. Before experimental designs can possess adequate


ecological validity, more about what occurs in actual class-

rooms must be investigated.


The subjects consisted of 22 third-grade teachers in

the Alachua County School District. Deci et al. (1981a)

pooled data on the relationship between teacher orientation

and student characteristics for fourth through sixth graders.

A need existed for investigating this relationship with

primary age children. To control for age effects, only one

grade level was used. Third grade was the youngest grade

level for whom Harter (1979, 1980) gathered reliability data

on two of the children's measures, the perceived competence

scale and the intrinsic motivation scale. For this reason

third grade was the primary grade chosen for this study.

Only nonteamed teachers were used in this study. It would be

difficult to separate out the effects of team-teachers on

each other and on the students. For example, if one teacher

is control-oriented and the other is autonomy-oriented, the

children may receive two types of praise in the classroom. A

clearer estimate of the relationships between teacher orien-

tation, praise, and student characteristics can be obtained

in single teacher classrooms.

In recruiting subjects for this study, the term, teacher

praise, was not used. The broader, more inclusive term of

teacher interaction was used instead. This was done to

minimize the Hawthorne effect. That is, the teachers were

unaware that their use of praise ws being observed in order

to prevent alterning their natural praise style.

Because of the need for informed consent, participation

of students and teachers was voluntary. An average sample

of 7 children from each classroom was used for a total of 154

children. These children were selected on the basis of their

membership in the middle-ability reading group. Teachers

sent out student informed consent forms to parents and then

implemented follow-up measures consisting of a follow-up

letter and then a telephone call. Informed consent reduced

the generalizability of this study to similar groups.


The "Problem in Schools" Questionnaire

Each teacher's orientation toward supporting autonomy in

versus controlling children was assessed using The "Problems

in Schools" Questionnaire developed by Deci et al. (1981b)

[see Appendix A]. The questionnaire is composed of eight

vignettes, each of which is followed by four items. Each

item represents a possible solution to the situation in the

vignette. Each item also corresponds to one of the four

subscales: highly controlling (HC), moderately controlling

(MC), moderately autonomous (MA), and highly autonomous

(HA). The teacher responded to each item by using a


seven-point scale which ranges from 1 (very inappropriate) to

7 (very appropriate). A total score was based on the sum of

the weighted averages of each of the four subscales.

Deci et al. (1981b) reported summary statistics for the

total scale score and the four subscales for 68 teachers

(K-6) as shown in Table 1. They reported split half relia-

bilities using Cronbach's alpha for the standardized scores

for the four subscales (HC, MC, MA, HA) as .73, .71, .63, and

.80, respectively. Test- retest reliability data on 19

teachers using a two-month interval were collected. Relia-

bility coefficients for the four subscales range from .77 to

.82. The test-retest reliability for the total score is .70.

Two studies were conducted to ascertain the construct

validity of The "Problems in Schools" Questionnaire. These

are reported in Deci et al.(1981a). Study 1 investigated the

relationship between teachers' and children's characteris-

tics. To be considered valid, the teacher measure should

correlate with the actual intrinsic motivation and perceived

competence of the children. Thirty-five teachers and their

610 students (4th-6th grades) were the subjects. Teacher

orientation correlated significantly with fall data on three

intrinsic motivation subscales (from .36 to .56) and on gen-

eral and cognitive perceived competence (.36 and .29, respec-

tively). These results appear in Table 2.

The second analysis used the same data base and inves-

tigated whether children perceive a difference in the class-

rooms of control- versus autonomy-oriented teachers. Teacher

Table 1

Means, Standard Deviations, Minimum Scores, and Maximum Scores for the Scale
Total and Each of the Four Subscale Scores (n=68).

Minimum Maximum
Subscale M SD Score Score

1 HC 2.72 .86 1.13 6.50
2 MC 3.36 .90 1.13 5.63
3 MA 3.67 .85 1.25 6.25
4 HA 6.05 .70 1.88 7.00
Scale Total 6.98 3.11 -10.13 12.13

Table 2

Correlations Between Teachers' Orientation and Children's Intrinsic Motivation
and Perceived Competence (n=35).*

Subscale Correlation (Fall data)

Intrinsic Orientation
Prefers challenge .41
Curiosity .56
Mastery attempts .35

Perceived Competence
General self-worth .36
Cognitive .29
*Note: The classroom is unit of analysis.


scores on The "Problems in Schools" Questionnaire were cor-

related with children's scores on deCharm's Classroom Climate

Questionnaire (1976). Deci et al. (1981a) reported a corre-

lation of .354, significant at the .05 level. Thus, it

appears children do perceive autonomy-oriented teachers as

supportive of autonomy.

Teacher Praise Behaviors Checklist

Each teacher's praise behavior during reading group

was recorded on a checklist developed by the researcher [see

Appendix B]. A compilation of the teacher praise behaviors

and contexts studied by Dershimer (1980, 1982), Dweck et al.

(1978), Evertson (1975), and Soar and Soar (1978) are included

on the checklist. The researcher drew heavily on Soar and

Soar's Climate and Control System for the form and organiza-

tion of the checklist.

The Teacher Praise Behaviors Checklist is organized

in two main sections called Behaviors and Context. The

Behaviors section contains a list of behaviors that span

three dimensions: a control dimension, an information dimen-

sion, and an affect dimension. The Context section contains

a list of possible contexts in which praise occurs. An extra

section for Additional Observer Comments was included to

record any unlisted behavior or context that occurred during

observation periods. The checklist also contains a section

for recording the actual praise statement used by the

teacher. This provided the reseacher with the opportunity

for additional praise analysis.

The observers checked as many behaviors and contexts

as were appropriate for each praise statement. The behav-

iors and contexts were not mutually exclusive. The follow-

ing are explanations of each of the 33 behaviors and con-

texts included in the checklist. The source for each

explanation is cited after the behavior or context is


Items 1-5 are part of the control dimension which varies

from very controlling to slightly controlling praise state-

ments. Two types of behaviors are included in this

dimension, contingent praise and praise with "I" referents.

1. Paired with tangibles/other rewards

Although this is not always the result (Swann &

Pittman, 1977), Deci and Ryan (1980) state that in most

cases, tangible rewards are very controlling. Tangibles

include happy faces, gold stars, candy, etc. Other rewards

include more recess time, less homework, etc. Examples:

"Good work" as teacher draws happy face.

"You've done so well that you can go outside for five

extra minutes of recess today."

2. Refers to teacher's needs, likes

By referring to an outside cause of behavior, the

teacher may shift the locus of causality for the child from

internal to exernal (Pittman et al., 1980). Examples:

"I like the way Terry is sitting."


"I need all my students to be good readers today, just

like Amy."

3. Refers to teacher's genuine opinion

It seems necessary to differentiate between "I state-

ments" that convey control from those that express genuine

opinion. Example:

"I think your story is great!"

4. Contingent on response

Praise that is contingent on responding rather than

on the quality or content of the response may be perceived

as controlling (Enzle & Ross, 1978). Praise that follows

a student response in a uniform way with no variation for

quality of responding is checked here. Example:

"Good job" following each student answer.

5. Noncontingent

Praise statements that follow incorrect answers or

mark teacher pauses or transitions (Brophy, 1981b).

Items 6-17 are part of the information dimension which

varies from vague information to very specific, individual

information about competence.

6. Contingent on quality of response

Praise that indicates level of competence varies with

the quality of the child's answers. The intensity of praise

increases as child performance increases. Praise statements

can also refer to the behavior praised. Examples:

"Nice" . "Great" . "Excellent"

"Your pronunciation deserves applause, Tom!"

7. Repeats child's response

Praise that is simple repetition is quite mild but

may convey information to the child that he/she is correct.

Dershimer (1980, 1982) includes this category.

8. Accepts child's response

This category also originates from Dershmer's (1980,

1982) study. Example:

"Good" or "Okay" or a nod after each child's response.

9. Phrase, general

Praise statements that are general and brief may

convey information about competence (Danner & Lonky, 1981).

Soar and Soar (1980) include a general praise item.


"Fine work" "Good job"

10. Sentence, general

These statements differ from #9 in their length.


"You are doing really well."

"You did a great job."

11. Specifies child's behavior/attribute

Praise that specifies what is good about a person

or his/her work may give more information about competence

(Bernhardt & Forehand, 1975; Scheer, 1976/1977). Soar and

Soar (1978) include a specific praise item. Examples:

"Great! You read well today."

"Good girl. You're so patient."

12. Specifies child's behavior/attribute with detail

These statements differ from #11 in degree of speci-

ficity. Examples:

"Great! You read twelve new works today."

"Good answer. It shows you really thought about the

meaning of the story."

13. Extended

Praise statements that go on for several sentences.

Dershimer (1980, 1982) includes this category; extended

praise is rare. Example:

"What a good reading group today. You children are

improving so much. Pretty soon we'll have to get

new books because these are getting too easy."

14. Compares child to others

Social comparison can provide highly direct and unam-

biguous information about one's competence (Boggiano &

Ruble, 1979; Pittman et al., 1980). Examples:

"You're the best reader in the group."

"Compared to my other classes, you're great!"

15. Compares child to self

The teacher may provide competency information by

comparing the child's performance to his/her past perform-

ance. Brophy (1981b) includes this behavior as desirable

praise behavior. Examples:

"Good girl! This is your best paper ever."

"You've improved your expression a lot, Tommy."

16. Attributes effort

The praise statement includes an effort attribution

(Dweck, 1978). Example:

"Good reading. You're trying very hard."

17. Attributes ability

The praise statement includes an ability attribution

(Dweck, 1978). Example:

"You are a great reader."

Items 18-23 are part of the affect dimension that

varies from congruent with the praise statement to incon-

gruent. Brophy (1981b) indicates that contradictory affect

damages the credibility and genuineness of the praise state-

ments. This implies that congruent affect may enhance

the informative function of praise, whereas incongruent

affect may enhance the controlling function.

18. Flat voice

19. No facial expression

Items 18 and 19 express incongruent affect.

20. Enthusiastic voice

21. Smiles, laughs, nods

22. Eye contact

23. Touch, pat, hug

Items 20-23 express congruent affect. Items 21 and

23 are taken from Soar and Soar (1978).

The Context section contains 10 items that indicate

details about the situational context of the praise state-

ments. The general context is uniform across all

observation periods. That is, teacher praise was observed

during small group reading time. Within this general con-

text, praise statements were addressed to an individual

in private or in public or to a group of students in private

or in public. The praise statements were given to a male

or a female. The praise statement referred to the academic

work, the conduct or a personal comment or quality of

students. The praise statement also occurred during transi-

tion times. Praise was given when a child initiated contact

or when the teacher initiated contact. The context cate-

gories are taken from Evertson's (1975) revision of the

Brophy-Good Dyadic Interaction System.

After the praise statements were scored on the above

33 items, several subscales were formed by collapsing

items. Subscale 1 (Sl) is the total of all items scored.

One difference between the praise styles of control- and

autonomy-oriented teachers could be the total amount of

praise generated. Subscale 1 would indicate this differ-

ence, whereas the individual items may not. Subscale 2

(S2) is the controlling dimension of teacher praise and

includes items 1, 2, 4, 5, 18, and 19. These are the con-

trolling praise behaviors as well as the incongruent affect

items. Subscale 3 is the information dimension of teacher

praise and includes items 3, 6, 9-16, 20-22. These are

the praise statements which convey information, excluding

items 7 and 8 which convey vague information, and the con-

gruent affect items. These three subscales were included

because they may reveal patterns in the teachers' praise

styles that individual items do not.

Scale of Intrinsic Versus Extrinsic Orientation in the

Each child completed a measure of intrinsic motiva-

tion. Harter (1980) developed the Scale of Intrinsic Versus

Extrinsic Orientation in the Classroom based on her own

theoretical model of intrinsic motivation, which stems

from White's (1959) model of effectance motivation. Charter

uses classroom learning as the situational context for

her scale. She reports the following question as the start-

ing point of the scale: "To what degree is a child's

motivation for classroom learning determined by his or

her intrinsic interest in learning and mastery, curiosity,

preference for challenge in contrast to a more extrinsic

orientation in which the child is motivated to obtain

teacher approval and/or grades, and is very dependent on

the teacher for guidance?" (Harter, 1980, p. 5).

In answering this question, Harter developed a scale

with five subscales, three of which reflect dimensions

of intrinsic motivation (preference for challenge versus

preference for easy work, curiosity/interest versus pleasing

the teacher, and independent mastery versus dependence

on teacher) and two of which reflect dimensions of internal

evaluation (independent judgment versus reliance on

teacher's judgment and internal criteria versus external

criteria). Only the three intrinsic motivation subscales

were used in this study. In their preliminary study, Deci

et al. (1981a) reported low or negative correlations between

the two internal evaluation subscales. Therefore, these

two subscales were not used.

The scale uses a structured alternative format which

decreases the social desirability response tendencies of

the true-false format. An example of the response format


Some kids like hard Other kids prefer easy
work because it's a but work that they are sure
challenge they can do

Really Sort of Sort of Really
true true true true
for me for me for me for me

Each item is scored on a scale from 1 to 4 with 1 being

most extrinsically oriented and 4 most intrinsically ori-

ented. Each subscale contains six items, three of which

are worded to begin with the intrinsic orientation and

three with extrinsic orientation. No two consecutive items

are from the same subscale and no more than two consecutive

items are keyed in the same direction.

The Kuder-Richardson Formula 20 reliability coefficient

was used to assess internal consistency using data from

third- to ninth-grade students from New York, California,

and Colorado. Reliabilities for subscales ranged from

.78 to .84, .68 to .82, .54 to .78, .72 to .81, and .75

to .83 for the challenge, mastery, curiosity, judgment,

and criteria subscales, respectively (Harter, 1981). Test-

retest reliability data for 761 third through sixth graders

in New York after a 9-month interval and for 793 third

through ninth graders in California after a one-year inter-

val ranged from .48 to .63 for the five subscales (Harter,


Validity studies included factor analyses, discriminant

as well as criterion validity. The factor analyses were

performed on data from the New York and California samples

above. Both orthoginal and oblique solutions revealed

the same five factors, reflecting the five subscales.

The average loadings for items on their designated factors

was between .46 and .53. No items cross-loaded on other

factors (Harter, 1980).

Charter (1981) reports several studies of discriminant

validity. In one study Harter (1981) compared the scores

of 26 pupils (fourth to sixth grades) in a private "open"

school from upper-middle class families which emphasized

attributes of the intrinsic pole of the scale with scores

from pupils matched for age and sex from a traditional

school. The results indicated the expected differences;

the open school students scored higher (more intrinsically)

on each subscale of the measure (see Table 3).

Table 3

Mean Scores for Open School and Traditional School Pupils
on Charter's Intrinsic Orientation Scale.

Mean Scores
Subscale Open Traditional Level

challenge 2.98 1.81 (p .001)
curiosity 3.10 2.30 (p .001)
independent mastery 2.80 2.47 (p .05)

For a test of criterion validity, Harter (1981) com-

pared the scores of a group of third through sixth graders

in Colorado on the Intrinsic Orientation Scale with teacher

ratings. Correlations between teacher and pupil ratings

were as follows: .73 for challenge, .67 for curiosity,

and .61 for mastery.

The Perceived Competence Scale for Children

Each child filled out Harter's Perceived Competence

Scale for Children (1979). Charter constructed the scale

to be sensitive to perceived competence as "an important

correlate and mediator of the child's intrinsic motivation

to be effective, to engage in independent mastery attempts

in the anticipation of a competent outcome" (p. 1). The

scale reflects the view that competence is not a unitary

domain but has several components. There are four subscales

to the Perceived Competence Scale. These are

cognitive competence--reflected primarily in academic


social competence--emphasis is on popularity with


physical competence--defined in terms of ability at

sports and outdoor games

general self-esteem--general feelings of worth which

are independent of any particular skill domain.

Only the cognitive competence and general self-esteem sub-

scales were used in this study. In their preliminary study,

Deci et al. (1981a) reported low or negative correlations

between social and physical competency, respectively, and

teacher orientation. Therefore, these two subscales were

not used.

The response format is the structured alternative

format (see Intrinsic Orientation section). Each of the

28 items is scored on a scale from 1 to 4, with a score

of 1 indicating low perceived competence and 4 high per-

ceived competence. Fourteen of the 28 items are worded

with high perceived competence reflected in the first half

of the statement.

Kuder-Richardson Formula 20 reliability coefficients

for each subscale were assessed on large samples of third-

through sixth-grade children in Colorado, Connecticut,

California, and New York. The internal consistency values

were .76 and .73 for the cognitive and general subscales,

respectively (Harter, 1979). Test-retest reliability data

for 208 Colorado students after a three-month interval

and 810 New York students after a nine-month interval were

.78, and .70 for Colorado and .78, and .679 for the New

York sample, for the two subscales (Harter, 1982).

Factor analyses of the items on the Perceived Com-

petence Scale using both orthogonal and oblique solutions

were obtained. Items have moderate to high loadings on

their designated factor and do not cross-load on other

factors. The same pattern was replicated on samples from

all four states (Harter, 1982).

Criterion validity has been established for the cog-

nitive subscale. Teacher ratings of cognitive competence

were correlated with the children's scores. For the

California sample, third through ninth grades, the following

correlations were found for grades 3 through 6, .28, .32,

.50, and .55, and for grades 7 through 9, .31, .66, .73,

respectively. The same pattern and magnitude of correla-

tions was obtained when children's scores were correlated

with the Iowa Test of Basic Skills (Harter, 1982).

Metropolitan Achievement Test

Achievement was measured by Metropolitan Achievement

Test (MAT) scores. The test has nine subtests including

word knowledge, word discrimination, reading, spelling,

language (usage, puncutation, and capitaliation, total),

mathematics (computation, problem solving, and concepts).

The reading subtest and total scores were used. The Eighth

Mental Measurement Yearbook (Buros, 1978) reports favorable

reviews for the MAT. Gronlund reports that reliability

estimates are based on measures of internal consistency.

Correlations range from .84 to .97 for the subtests.

Wingard and Bentler note that the selection of the stan-

dardization sample is a problem. A disproportionately

large number of low-income and low-ability pupils were

erroneously included. The national norms, therefore, cannot

be relied upon. Comparisons among schools and districts

can, however, be made with confidence. In his critique,

Wolf concludes the MAT are high quality instruments despite

minor technical flaws and their use is well justified.

Data Collection

This study involved 22 third-grade teachers in Alachua

County, Florida, and 154 children, an average sample of

7 children from each classroom. Participation of teachers

and children in the study was determined by their willing-

ness to sign the informed consent forms. Follow-up attempts

by the researcher to increase their participation level

included personal contacts, letters, and phone contacts.

The use of volunteer subjects decreased the generalizability

of the present study to similar groups. It was important

to use teachers from a variety of schools so that the

discipline procedures and social climate of any one school

did not overly influence the data collected. Follow-up

attempts to increase teacher participation focused on

including different schools in the county. Eight schools

were included.

Once the participating teachers were identified, the

researcher contacted the teachers at each school and admin-

istered Deci's The "Problems in Schools" Questionnaire

(1981) within one week. Group administration of Deci's

instrument promoted faster and more independent teacher

responses. Arrangements were made at this time for the

children to be tested and for observers to collect praise

data in each classroom. Copies of each teacher's daily

and weekly schedules were obtained to facilitate observation


Four observers were trained to the criterion level

of 75% interrater agreement on the Teacher Praise Behaviors

Checklist, developed by the researcher. The mean inter-

rater agreement over four observation periods was 76.7%.

The agreement ranged from 71% to 81%. Two agreement checks

were made before the study began, and two additional agree-

ment checks were made during the study. The calculations

were based on the number of agreements divided by the total

number of praise statements for each observation period.

Training consisted of approximately four hours of viewing

and discussing videotapes while learning the instrument

and an additional two hours in the field observing and

recording teacher praise episodes. Videotape training


provided the opportunity for praise episodes to be evaluated

and discussed repeatedly until observer consensus was

reached. Field training provided the opportunity to prac-

tice accurate recording of praise episodes of teachers

in real contexts. That is, observers recorded teacher

praise while experiencing the distractions of a noisy class-

room and while sitting only a few feet from the reading

group. An attempt was made to find teachers who displayed

frequent and readily discernible styles of praise for the

practice sessions. Practice took place in second, rather

than third, grade classrooms to avoid any overlap between

practice and experimental classrooms.

It has been shown that more frequent, shorter episodes

of observation are more reliable than longer but fewer

periods (Rowley, 1978). For this reason, each teacher

was observed four times for 30 minutes, or a total of 120


To facilitate comparability of data among teachers,

observation of teacher praise occurred during small group

reading time. This area of the curriculum was chosen

because of the similarity in format across classrooms,

frequency of occurrence, high probability of teacher feed-

back to student behavior, verbal nature of the task, and

the high priority accorded reading in the elementary school


Reading groups were generally 30 minutes long. An

observer began recording praise statements from the time


the teacher called the children together until she dismissed

them. In the event a reading group lasted a shorter period

of time than 30 minutes, the time the group began and ended

was recorded. The observer then repeated the praise

observation process with the next group, noting when the

30-minute period ended but recording praise statements

until the group ended. In other words, praise-recording

episodes were the length of reading groups. It was impor-

tant to record the entire group session in the event praise

style changed as the lesson progressed.

Each teacher was observed for a minimum of 120 min-

utes. For the majority of teachers this was for four read-

ing groups of 30 minutes each. For a few teachers with

shorter reading groups, this was more than four groups.

In both cases, observers recorded praise statements in

reading groups of middle-ability students. Only the middle-

ability reading group in each classroom was observed because

the middle-ability reading group was the largest group

per classroom and provided enough student subjects for

this study. The same ability level was observed in each

classroom to facilitate comparability of the data


Observation times, of course, coincided with reading

group times. The time of observation, therefore, depended

upon the teachers' schedules and was not made uniform by

the researcher. It was common for reading groups to occur

in the mornings every day of the week. No teacher was

observed more than once per day. Each teacher was observed

four times for 30 minutes each. Each of the four observers

recorded praise statements from each of the 22 teachers.

During the observation weeks, but at times when obser-

vation was not going on (afternoons, for example), the

researcher and one other proctor administered the children's

measures, Intrinsic Orientation Scale and Perceived

Competence Scale. The administration time was uniform

across children to control for fatigue. A comparable set-

ting was chosen for testing. The measures were administered

in groups of 16-24 children, all the subjects per school.

When more than 24 children in any one school participated,

the group was divided in half for testing. Total testing

time was no longer than 40 minutes, including instructions.

Achievement data were obtained in the form of Metro-

politan Achievement Test scores, Spring Administration,



The present study was designed to investigate the

relationships between teacher orientation toward controlling

versus supporting autonomy in children and frequencies

of various types of praise. Further, the study investi-

gated the relationships between these two teacher variables

and children's feelings of competence, intrinsic motivation,

and achievement. The design was correlatonal-observational

because the study was exploratory and because of the need

for ecological validity.

The subjects consisted of 22 third-grade teachers

and 154 children in their middle-ability reading groups.

Observation of praise occurred during reading group sessions

on four occasions of 30 minutes' duration. Reading groups

were chosen because of the similarity in format across

classrooms, frequency of occurrence, high probability of

teacher feedback to the student behavior, verbal nature

of the task, and the high priority accorded reading in

the curriculum.

Teacher orientation was assessed in the weeks prior

to classroom observation, and the children's feelings of

competence and intrinsic motivation were assessed in the

afternoons concurrent with classroom observation. Achieve-

ment data were collected after the school year was ended.



The purpose of this study was to explore the relation-

ships between teachers' use of praise and their orientation

toward controlling versus supporting autonomy in children.

The relationships between these two teacher measures and

children's feeling of competence, intrinsic motivation,

and achievement were also explored. One hundred fifty-four

third-grade children and their teachers in 22 classrooms

participated in this study. The children's scores were

aggregated and averaged per classroom. The classroom was

thus the unit of analysis for this investigation.

Specifically, the study sought to answer three


1. Were there significant relationships between frequen-

cies of various types of teacher praise and teacher

orientation toward controlling versus supporting

autonomy in children?

2. Were there significant relationships between frequen-

cies of various types of teacher praise and children's

feelings of competence, intrinsic motivation, and


3. Were there significant relationships between teacher

orientation toward controlling versus supporting

autonomy in children and children's feelings of

competence, intrinsic motivation, and achievement?

To answer these questions, correlations were calculated

between the two teachers' measures (frequency of teacher

praise and teacher orientation toward controlling versus

supporting autonomy in children) and between each teacher

measure and the three children's measures (children's feel-

ings of competence, intrinsic motivation, and achievement).

The following teacher measures were used: to measure

frequency of teacher praise, the Teacher Praise Behavior

Checklist (35 items plus 3 collapsed subscales), and to

measure teacher orientation toward controlling versus sup-

porting autonomy in children, The "Problems in Schools"

Questionnaire (1 score).

The following children's measures were used: to meas-

ure children's feelings of competence, the Cognitive and

General subscales of the Perceived Competence Scale for

Children (2 subtest scores); to measure children's intrinsic

motivation, the Preference for Challenge, Independent

Mastery and Curiosity/Interest subtests of the Scale of

Intrinsic Versus Extrinsic Orientation in the Classroom

(3 subtest scores); and to measure children's achievement,

the reading subtest and the total scores on the Metropoli-

tan Achievement Test (2 scores).

Descriptive Statistics for Teachers' and Children's Measures

Descriptive statistics for the teacher orientation

measure are reported in Table 4. Deci et al. (1981b) indi-

cate that teachers who score below 6.98 on The "Problems

in Schools" Questionnaire are control-oriented. Teachers

who score above 6.98 are autonomy-oriented. In the present

study 15 teachers scored as control-oriented and 7 teachers

scored as autonomy-oriented.

Table 4

Descriptive Statistics for The "Problems in Schools"

Standard Minimum Maximum
N Mean Deviation Score Score

22 5.00 2.87 -0.80 10.00

Descriptive statistics for the various types of praise

are presented in Table 5. There were 1,712 praise state-

ments generated by the 22 teachers for the 120 minutes

of observation time each. The teachers' use of praise

ranged from a minimum of 2 statements to a maximum of 46

statements per 30 minute lesson for an average of 19 state-

ments per 30 minute lesson.

The observers checked as many items as were appropriate

for each praise statement; that is, the items were not

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs