See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225698867

Effects of different evaluative feedback on students’ self-efficacy in learning

Article in Instructional Science · January 2010

DOI: 10.1007/s11251-008-9077-2

CITATIONS READS

120 4,296

2 authors:

Joanne Chung-Yan Chan Shui-fong Lam

The Chinese University of Hong Kong The University of Hong Kong
14 PUBLICATIONS 754 CITATIONS 70 PUBLICATIONS 3,287 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Shui-fong Lam on 03 February 2014.

The user has requested enhancement of the downloaded file.

Instr Sci (2010) 38:37–58
DOI 10.1007/s11251-008-9077-2

Effects of different evaluative feedback on students’

self-efficacy in learning

Joanne C. Y. Chan Æ Shui-fong Lam

Received: 9 October 2007 / Accepted: 1 September 2008 / Published online: 23 September 2008
Ó Springer Science+Business Media B.V. 2008

Abstract Two studies examined the effects of four types of teachers’ evaluative feedback
on Chinese students’ self-efficacy in English vocabulary acquisition. In Study 1, a random
sample of Grade 8 students (N = 79) learned prefixes and received either formative or
summative feedback after failure in test. The results showed that students who received
summative feedback showed a larger decrease in their self-efficacy than those who
received formative feedback. In Study 2, a random sample of Grade 7 students (N = 77)
went through similar procedures as in Study 1 except that students received either self-
referenced or norm-referenced feedback. The results showed that self-referenced feedback
was more beneficial to students’ self-efficacy than norm-referenced feedback. The influ-
ences of teachers’ evaluation and feedback on students’ self-efficacy are discussed.

Keywords Self-efficacy
Control
Teacher feedback
Evaluation
Goal theory

Introduction

In social cognitive theory, personal factors, behavior, and external environment form a
triadic reciprocal causation (Bandura 1986). Self-efficacy theory is concerned with the
factor of personal causation within this interdependent structure and provides guidance for
individuals to exercise influence over their way of living (Bandura 1997). By definition,
self-efficacy is the belief in an individual’s capabilities to execute behavior that is required
to achieve prospective outcomes (Bandura 1977). Self-efficacy theories distinguish among
the agents of control, the means of control, and the ends of control (Skinner 1996).
Outcome expectations are individuals’ estimates that a certain behavior (means) can

Joanne C. Y. Chan is now at The Hong Kong Polytechnic University. This paper is based on Joanne C.Y.
Chan’s doctoral thesis completed at The University of Hong Kong. Parts of this paper have been presented at
the 17th Annual Convention of the American Psychological Society, Los Angeles, May 2005 and the 30th
Annual Conference of the Australian Psychological Society, Melbourne, September 2005.

J. C. Y. Chan (&)
S.-f. Lam

The University of Hong Kong, Pokfulam Road, Hong Kong, China
e-mail: [email protected]

123
38 J. C. Y. Chan, S.-f. Lam

achieve the desired outcomes (ends) whereas efficacy expectations are individuals’ beliefs
of whether they (agent) can produce the behavior (means) which can produce the desired
outcomes (ends) (Bandura 1977). According to Bandura (1997), though certain behaviors
will attain desired outcomes, individuals will not exert control unless they believe that they
can produce these behaviors. In this research, we endeavored to test empirically how
teachers can manipulate the ends and means of control to affect students’ self-efficacy.
Self-efficacy affects one’s level of motivation, affective states, actions, thought patterns
and resilience (Bandura 1997; Pajares 1996). Individuals with high self-efficacy invest
more effort and persist longer than those with low self-efficacy, especially when they face
setbacks (Bandura 1989). Individuals with high self-efficacy are also more likely to
attribute failure to insufficient effort or situational factors, whereas those with low self-
efficacy tend to attribute failure to lack of ability (Bandura 1999). Given the powerful
influence of self-efficacy, researchers and educators are eager to understand what con-
tributes to its development and maintenance. It is important to note that self-efficacy
beliefs are multi-dimensional rather than uni-dimensional (Zimmerman and Kitsantas
1997). The present research focused on the domain of academic self-efficacy.

Academic self-efficacy and sources of efficacy information

Perceived academic self-efficacy is defined as the personal judgment of one’s capabilities

to organize and execute the courses of action to attain specific types of educational per-
formances (Zimmerman 1995). Research has shown that self-efficacy is related positively
to motivation, memory, persistence, stress management, cognitive engagement, use of self-
regulatory strategies and achievement (Linnenbrink and Pintrich 2002; Zimmerman 1995).
Bandura (1997) postulated four fundamental sources of information that consolidate
self-efficacy, namely mastery experience, social persuasion, vicarious experience, and
physiological and affective states with mastery experience being the most influential. In
general, success increases self-efficacy and failure lowers it (Bandura 1999). As self-
efficacy judgments and performance are reciprocally linked, previous performance can
alter one’s efficacy judgments and one’s efficacy can affect subsequent performance
(Cervone 1993).
Besides performance, social persuasion influences individuals’ self-efficacy. Social
persuasion that confirms the potential of individuals strengthens their self-efficacy but an
individual’s subsequent poor performance may decrease it (Schunk 2003). Furthermore,
social persuasion that convinces individuals of their lack of ability also leads them to avoid
challenge and give up easily (Bandura 1999). In an academic setting, teachers and peers are
the immediate sources of social persuasion for students. In the current research, we looked
at how teachers affect students’ self-efficacy through manipulation of mastery experience
and social persuasion.

Effects of teachers’ evaluative feedback on students’ self-efficacy

Teachers partially determine students’ mastery experience by the evaluative standards they
set. Consider a real case scenario in Hong Kong where two schools adopt different passing
standards for a class test: one school may set the passing score at 60% and another school

123
Effects of different evaluative feedback 39

may set it at 50%. Thus, students with the same absolute score of 55% would obtain a pass
in the former but a fail in the latter. Different reference standards yield different inter-
pretations of success that subsequently influence students’ evaluation of their self-efficacy.
In addition, teachers’ evaluative feedback constitutes social persuasion that affects stu-
dents’ self-efficacy. A teacher who confirms students’ capabilities may enhance their self-
efficacy but a teacher who tells students that they are incapable may weaken their self-
efficacy.
From the perspective of social cognitive theory, teachers’ feedback constitutes an
environmental variable that influences self-efficacy, a personal variable (Schunk and
Zimmerman 1997; Schunk 2003). Individuals react to feedback by maintaining successful
strategies and modifying unsuccessful ones (Zimmerman 1997). Research has shown that
feedback that informed students of their mastery of learning strategies (Zimmerman and
Kitsantas 2002) and those that linked students’ success with their mastery of strategies both
contributed positively to self-efficacy (Schunk and Rice 1987, 1992, 1993). Schunk and
colleagues conducted a series of experiments to examine the effects of feedback and goals
on students’ self-efficacy and found that progress feedback together with learning goal are
beneficial to their self-efficacy (Schunk and Rice 1991; Schunk and Swartz 1993a, b).
There is also evidence that the sequence of ability and effort attribution feedback affected
self-efficacy (Schunk and Rice 1986).
Given the evidence that feedback affects self-efficacy, we endeavored to identify how
feedback affects self-efficacy through Ellen Skinner’s (1996) construct of control.
Adopting Skinner’s terminologies, in an academic setting, students are the agents of
control, learning strategies are the means of control and designated educational goals are
the ends of control. Feedback that convinces students that they, as the agent, can exercise
control over the means enhances efficacy expectation. Moreover, feedback that provides
students with the means to achieve their goals, or the ends, strengthens outcome expec-
tations. The goal of this paper is to provide empirical evidence of how the triadic construct
of control can further our understanding of the effects of teachers’ feedback on students’
self-efficacy.

Importance of achievement goals and feedback on self-efficacy

Teachers’ evaluative feedback can influence students’ self-efficacy through its impact on
students’ achievement goals. Achievement goals are the ends in the construct presented in
Fig. 1. According to Ames and Ames (1984), different structures of the learning envi-
ronment can make different goals salient. Although researchers use different terminologies
(e.g. Ames and Archer 1987, 1988; Dykeman 1994; Nicholls 1979; Schunk and Rice 1989;
Schunk and Swartz 1993a), achievement goals have been generally dichotomized into
learning versus performance goals (Dweck 1986; Dweck and Leggett 1988; Elliott and
Dweck 1988).
Learning goals define competence in absolute terms (Elliot and McGregor 2001).
Students who adopt learning goals are oriented towards learning and improvement based
on a self-referenced standard. They tend to attribute outcome to effort (Ames 1992). Since
students can control how much effort they invest in a given task, they will have more
confidence over the agent–means relation, that is, higher self-efficacy. Learning goals can
be further differentiated into mastery-approach goal that target at approaching success and
mastery-avoidance goal that targets at avoiding failure (Elliot and McGregor 2001).

123
40 J. C. Y. Chan, S.-f. Lam

Fig. 1 Self-efficacy and the Agent

triadic relation among agent,
means and ends of control
efficacy expectation

Means Ends
outcome expectation

Performance goals define competence in normative terms (Elliot and McGregor 2001).
Students with performance goal usually focus on using ability to achieve success and
impress others. Since most individuals believe that ability is relatively stable and unal-
terable (Weiner 1986), they may perceive less control over the means of achievement,
hence lower their self-efficacy. There is the differentiation between performance-approach
goal that aims at approaching success and performance-avoidance goal that aims at
avoiding failure (Elliot and Harackiewicz 1996; Elliot and McGregor 2001; Skaalvik
1997). Students may also adopt a failure-avoiding pattern of motivation when they are
unsure if others will evaluate their performance positively (Ames 1992).
Some researchers endorse a multiple goal perspective which claims that learning and
performance goals coexist in an individual’s goal system (Barron and Harackiewicz 2001).
Urdan and Maehr (1995) also proposed the concept of social goals to describe the social
purposes underlying academic achievement. For the purpose of this paper, we focused on
mastery-approach goal and performance-avoidance goal. The focus was on these two goals
because they have the most opposite consequences after setback. Mastery goal is resilient
to failure and still elicits mastery behaviors and approach mentality after negative feed-
back. In contrast, performance goal in face of failure will turn into performance-avoidance
goal that elicits helpless behaviors and avoidance mentality. For the parsimony and con-
venience of presentation, we will refer to the former as learning goal and the latter as
performance goal throughout the paper.

Four types of evaluative feedback and two achievement goals

Achievement goals and evaluative feedback influence each other. Given that each
achievement goal creates its own framework for processing information (Elliott and Dweck
1988), students with different achievement goals may interpret feedback differently.
Evaluative feedback, on the other hand, orients students towards different achievement
goals. In the present research, we will examine four types of feedback: formative, sum-
mative, self-referenced and norm-referenced.
Formative feedback and summative feedback act on the means of control (see Fig. 1).
Formative feedback provides students with the means to achieve a goal and thus orients
students towards a learning goal (McAlpine 2004). As a result, students may invest more
effort or adopt different strategies when they face failure (Elliott and Dweck 1988). On the
contrary, summative feedback focuses on the outcome (McAlpine 2004) without
empowering students with the means to achieve the goal. This focus facilitates a perfor-
mance orientation and students are more likely to associate their performance with ability
evaluation (Ames 1992) and attribute failure to low ability. According to Schunk and
Zimmerman (2007), students’ conceptions of ability may influence self-efficacy. When
students believe that ability is not malleable (Elliott and Dweck 1988), they perceive little
control over ability as the means of goal achievement and lower their self-efficacy.

123
Effects of different evaluative feedback 41

Besides directly acting on the means of control, evaluative feedback can also affect the
ends of control, which inevitably influences the means. Classroom structures affect stu-
dents’ perceptions of cognitive abilities mainly by emphasizing social-comparative versus
self-comparative appraisal (Bandura 1997). Schunk (1991) stated that self-referenced
comparison helps students concentrate on their progress and adopt learning goal. Since
students’ improvement in the mastery of specific skills or areas of knowledge is being
assessed, they are more likely to focus on using effort to make improvement. Therefore,
this sense of control may enhance self-efficacy.
On the contrary, norm-referenced evaluation urges students to demonstrate their ability
by outperforming others (Popham 2001). It leads students to focus on comparative per-
formances and encourages outcome attribution to ability (Ames 1992). However, as
students lack the means of controlling the performance of others, their self-efficacy
weakens. After failure, students may adopt performance-avoidance goal to avoid social
comparison that may reveal their incapability.
Past research has shown that students in self-referenced groups demonstrated higher
self-efficacy and better skill than those in social-referenced groups (Shih and Alexander
2000). It was reasoned that the self-referenced groups could concentrate on self-
improvement and engaged actively in the task but the social-referenced groups were
preoccupied with comparing their abilities with others, resulting in anxiety that interfered
with both cognitive engagement and self-efficacy.

Contributions of the current research

As Bandura (1986) pointed out, one of the objectives in academic research is to study how
different types of educational practices and structures affect the development of social and
cognitive competencies. Over time, it is believed that students who have developed a
strong sense of self-efficacy are better equipped to progress with their own initiatives.
Therefore, it is of both theoretical and practical importance to study self-efficacy in aca-
demic settings.
Against this backdrop, this research aimed at responding to the call for a better
understanding of adolescents’ self-efficacy in specific settings, which has been underrep-
resented in the literature (Anderman and Maehr 1994; Pajares 1996). We looked into the
self-efficacy of Chinese adolescents in English vocabulary acquisition. With reference to
the triadic relation among the agents, means, and ends in control, we investigated the
effects of teachers’ evaluative feedback on students’ self-efficacy.
To bring about a greater capacity to generalize the findings, we chose four types of
evaluation feedback that are commonly used in classrooms, namely summative feedback,
formative feedback, norm-referenced feedback, and self-referenced feedback. We also
used normal students as participants to complement the substantial research that looked at
students with special needs (e.g. Garcı́a and de Caso 2004; Graham et al. 2005; Schunk
1982; Schunk and Swartz 1993b; Schunk and Rice 1986, 1987, 1989, 1993).
We chose vocabulary acquisition as the subject of investigation because it is an essential
element in language acquisition (Penno et al. 2002) and vocabulary knowledge is a well-
established predictor of reading comprehension (e.g. Cunningham and Stanovich 1997;
Spearitt 1972) as well as academic progress (Walker et al. 1994).
The present research employed experimental controls in a field setting with Chinese-
Confucian Heritage which has a strong emphasis on effort and endurance (Hau and Salili

123
42 J. C. Y. Chan, S.-f. Lam

1996). Classrooms with this heritage are also known for their competitiveness and high
prevalence for normative evaluation (Biggs 1996; Shih and Alexander 2000). Therefore,
this research would shed light on how different types of feedback can affect students’ self-
efficacy in a competitive learning environment.
The two studies in this research were conducted in actual classrooms to advance their
internal and external validities. The findings strive to enable instructors to design and
implement evaluative feedback that can improve students’ self-efficacy and consequently
enhance their learning motivation and achievements.

Overview of the two studies

The present research investigated the effects of four types of evaluative feedback on
students’ self-efficacy. The first study focused on the manipulation of the means in the
concept of control. Using formative and summative feedback, we compared the effects of
providing and withholding the means of control on students’ self-efficacy. The second
study focused on the manipulation of the ends in the concept of control. Using self-
referenced and norm-referenced feedback, we compared the effects of two different ends,
namely, outperforming self and outperforming others, on students’ self-efficacy.

Study 1

In Study 1, we compared the effects of summative and formative feedback on students’

self-efficacy. Formative feedback that provides strategies for subsequent improvement
gives students the means to achieve a goal explicitly while summative feedback that
addresses past performance that is no longer controllable withholds such means from
students. As self-efficacy is concerned with an agent’s perceived control over the means
towards a designated goal, we hypothesized that the self-efficacy of students in the for-
mative feedback condition would be higher than that in the summative feedback condition.

Method

Participants

Participants were 79 Chinese students in Grade 8 who joined a vocabulary builder pro-
gram. They were recruited from a secondary school in Hong Kong that uses English as the
medium of instruction. About a quarter of the secondary schools in Hong Kong use English
as the medium of instruction so the sample is considered to be representative of these
schools. Invitation letters were sent to the parents of all students in the eight classes of
Grade 8. Students were assigned to these eight classes based on their English proficiency.
Parents were informed that their children might be selected randomly to participate in a
‘‘Vocabulary Builder Program’’ during which they would learn vocabulary acquisition
strategies. They were also told that the researchers would study what instructional strat-
egies would be most effective. All parents gave their consent for their children’s
participation. The English teachers of the eight classes were asked to randomly select five
male and five female students from their respective class. As a result, 80 students were

123
Effects of different evaluative feedback 43

recruited but one student was absent on the day of the experiment so only 79 students
participated.

Experimental design

The 79 students were randomly assigned to either the summative feedback condition
(n = 39) or the formative feedback condition (n = 40). The instructor and sequence of
instruction were the same for both conditions. The only difference was that one condition
received summative feedback and the other received formative feedback in the evaluation
of the tests in the program.

Vocabulary acquisition program

The program was held in the school computer room during an English lesson. Two
assistants, unaware of the experimental procedures, distributed the questionnaires. Upon
arrival at the classroom, the instructor assigned each student a desktop computer in a carrel
with partitions to separate students. Each computer was installed with a program designed
for the experiment. The program presented students with instructions, exercises, and tests
of prefix usage. The software automatically recorded the students’ test answers.

Procedures

During the program, the students went through two 20-minute instructional sessions that
included exercises and tests. After each session they answered a questionnaire. In the first
session the instructor introduced six prefixes with examples to illustrate their usage. For
example, the prefix ‘‘mis’’ means ‘‘bad’’ or ‘‘wrong’’; therefore, ‘‘misbehave’’ means
behave badly and ‘‘miscount’’ means count wrongly. The examples were in a multiple-
choice format (e.g., ‘‘I misplace my books’’, the word ‘‘misplace’’ means (a) place cor-
rectly, (b) put in the wrong place, (c) read, (d) lost). Students tried to answer them and the
instructor later revealed the correct answers.
After teaching the first three prefixes, the instructor taught students how to differentiate
between real and pseudo prefixes, that is, words that start with the spelling of a prefix. Then
the instructor continued to teach the three remaining prefixes. After that, the instructor
taught another strategy. After the teaching session, students responded to the first ques-
tionnaire (Questionnaire A) that tapped their self-efficacy with regard to the subsequent
teaching session and test. Then students did Test 1 and had 30 s to answer each of the 16
questions.
At the end of Test 1, the computers provided students with feedback on their perfor-
mance on Test 1. Students in the summative condition received summative feedback
whereas students in the formative condition received formative feedback. After students
had viewed the results, the instructor started the second instructional session and taught
another six prefixes and two strategies. Prior to Test 2, the instructor asked students to
respond to the second questionnaire (Questionnaire B). The questions tapped their sense of
mastery with regard to the result of Test 1 and their self-efficacy about the subsequent
teaching session and test. Manipulation check items were also included. The students then
did Test 2 that had the same format as Test 1. After that, the computer generated genuine
feedback based on students’ performance. The instructor then informed students that there
was not enough time for further instruction. At the end of the school day, we debriefed

123
44 J. C. Y. Chan, S.-f. Lam

students from both conditions and informed them that, for the purpose of the study, every
student was looking at identical test results for Test 1 that did not reflect their performance
accurately. We gave students a package that included all the prefix teaching material, test
items, and an answer key for them to assess their actual level of mastery as well as for
future reference.

Feedback manipulation

After Test 1, the instructor told students in the summative condition that the computers
would show them the results, including the number of correct and incorrect answers, the
overall percentage of accuracy, and a list of the correct and incorrect items. She also
reminded students that it was most important to do well in the test as it could indicate their
ability. The results presented by the computers were pre-programmed. All students in the
summative condition received identical feedback irrespective of their actual performance.
To students in the formative condition, the instructor said that the computers would show
them the results, including the number of correct and incorrect answers and some sug-
gestions that were allegedly tailor-made for the students so that they could make
improvement. She also reminded the students that it was most important to learn from the
suggestions as the test could help them find ways of improvement. The students were led to
believe that the computers had analyzed their work and made suggestions for their
improvement. For example, participants would see the following feedback, ‘‘For questions
4, 9, 12, you should try to use the strategy of understanding the meaning of the whole
sentence before choosing the answer.’’ These suggestions were identical for all students in
the formative condition irrespective of their actual performance. We manipulated the
feedback so that the students in both conditions got seven correct answers out of 16
questions in Test 1. According to the grading system in most schools in Hong Kong, such a
result would be regarded as a failure because the accuracy rate was below 50%.

Measures

Self-evaluation of success

As mastery experience is the most important source of efficacy information (Bandura

1997), we asked students to evaluate how the results of Test 1 made them feel in Ques-
tionnaire B. The ratings were made on a seven-point scale from 1 for ‘‘not successful at
all’’ to 7 for ‘‘very successful.’’

Self-efficacy

Students evaluated their self-efficacy twice subsequent to learning sessions and tests. In the
first questionnaire (Questionnaire A) they evaluated their confidence in and control over
learning well in the first learning session and getting good results in Test 1. In the second
questionnaire (Questionnaire B) they responded to the same questions with regard to the
second learning session and Test 2. Specifically, students indicated their response about:
(1) How confident are you to do well in the next test? (2) How much control do you have
over the result of the next test? (3) How confident are you to learn all the prefixes in the
next lesson? (4) How much control do you have over how well you learn in the next
lesson? Seven-point scales from 1 for ‘‘not confident at all’’ or ‘‘no control at all’’ to 7 for

123
Effects of different evaluative feedback 45

‘‘very confident’’ or ‘‘very much control’’ were used. The Cronbach’s alpha of these four
items was .88 before Test 1 and .89 before Test 2; showing high internal consistency.

Manipulation checks

In Questionnaire B, we asked students to assess the quality of instruction and classroom

management to see whether there was any difference between the two conditions. The
evaluation of the clarity of teaching was on a 7-point Likert scale from 1 for ‘‘not clear at
all’’ to 7 for ‘‘very clear.’’ For noise level of the classroom, students rated from 1 for ‘‘not
noisy at all’’ to 7 for ‘‘very noisy.’’ We also asked students to indicate if they had learn
more about prefixes after the two lessons to evaluate the effectiveness of our instruction.
Moreover, we asked them to indicate the number of correct answers they had in Test 1.

Results

Performance

We compared students’ actual performance in Test 1. The summative condition (X ¼ 6:55;
SD = 1.72) and the formative condition (X ¼ 5:82; SD = 1.57) did not show a significant
difference, t(75) = -1.95, p = .06, ns, Cohen’s d = .45. In Test 2, the performance of the
students in the formative condition (X ¼ 11:05; SD = 1.45) were also similar to that of the
students in the summative condition (X ¼ 11:03; SD = 1.73), t(75) = .07, p = .95, ns,
Cohen’s d = .01.

Self-evaluation of success

After Test 1, students in the formative condition (X ¼ 3:59; SD = .72) felt more successful
than their counterparts in the summative condition (X ¼ 3:15; SD = 1.15), t(77) = 2.04,
p = .045, Cohen’s d = .46. The difference was statistically significant and the effect size
was medium according to statistical conventions (Urdan 2001).

Self-efficacy

We compared students’ self-efficacy before Test 1 and Test 2 (see Table 1). We performed
a mixed model Univariate Analysis of Variance (ANOVA) with self-efficacy (Test 1 vs.
Test 2) as the within-participant factor and condition (formative feedback vs. summative

Table 1 Means of self-efficacy by condition for Study 1

Self-efficacy Conditions t-value Cohen’s d

Formative, n = 39 Summative, n = 40

Test 1 5.22a (.80) 5.37b (.87) -.77 .18

a
Test 2 4.85 (.72) 4.69b (.80) .38 .21
Difference of Tests 1 & 2 .38c (.54) .68c (.77) -.20 .45

Note: The means sharing the same superscript are significantly different from each other. The numbers in
parentheses are standard deviations

123
46 J. C. Y. Chan, S.-f. Lam

feedback) as the between-participant factor. The ANOVA yielded a statistically significant

main effect for self-efficacy, F(1, 77) = 49.41, p \ .01, partial g2 = .39, but not for
condition, F(1, 77) = .001, p = .98, g2 \ .01. However, the interaction between self-
efficacy and condition was at the .05 statistical level, F(1, 77) = 3.92, p = .05, g2 = .05.
To understand this two-way interaction, we compared students’ self-efficacy between
Test 1 and Test 2 (see Table 1). Participants in both conditions experienced a decrease in
self-efficacy from Test 1 to Test 2. The formative condition decreased from 5.22
(SD = .80) to 4.85 (SD = .72), t(38) = 4.34, p \ .01, Cohen’s d = .50. The effect size
was medium. The summative condition also decreased from 5.37 (SD = .87) to 4.69
(SD = .80), t(39) = 5.57, p \ .01, Cohen’s d = .81. It is noteworthy that the effect size
was large. The summative condition showed a larger decrease in self-efficacy (X ¼ :68;
SD = .77) than the formative feedback condition (X ¼ :38; SD = .54), t(77) = -.20,
p = .05, Cohen’s d = .45. The difference was significant and the effect size was medium.

Manipulation checks

The summative condition (X ¼ 5:75; SD = 1.10) and the formative condition (X ¼ 5:97;
SD = .78) both rated the teaching as quite clear, t(70.2) = 1.04, p = .30, ns, Cohen’s
d = .23. Both the summative condition (X ¼ 1:90; SD = 1.15) and the formative condi-
tion (X ¼ 1:56; SD = .96) also rated the noise level of the classroom as very low
t(77) = -1.39, p = .17, ns, Cohen’s d = .32. In the formative condition, 92.3% of the
students indicated that they had learned more about prefixes after the two lessons while
90% of the students in the summative condition claimed so, v2(79) = .13, p = .72, ns.
When asked to indicate their results on Test 1, 94% of the formative condition and 94.6%
of the summative condition accurately recalled that they got seven correct answers,
v2(76) = .003, p = .96, ns. This showed that the majority of students remembered that
they got less than half of the questions correct in Test 1.

Discussion

The results showed that though students in the formative and summative conditions did not
show significant differences in their self-efficacy before Test 1 and Test 2, their change in
self-efficacy was significantly different with a medium effect size of .45. As seen, not-
withstanding that both conditions had identical learning environments and poor test
performances, different evaluative feedback had different effects on students’ self-efficacy
and sense of success.
In the formative condition, the feedback presented students with concrete means for
improvement that strengthened the means–ends relation in the construct of control.
Moreover, by telling students that the suggestions for improvement were tailor-made for
them, the agent–means relation could also be reinforced. It was possible that, as a result,
the consolidation of the triadic relation among the agents, means and ends of control
alleviated the decrease in students’ self-efficacy after failure. In the summative condition,
feedback was presented to students subsequent to their performance without any sugges-
tion for improvement. Since nothing could be done to alter past performance, the means–
ends relation would be weak. Students as the agents also had no control over the means to
alter the quality of the end product, thus the agent–means relation would also be weak. As a
result, students’ self-efficacy decreased.

123
Effects of different evaluative feedback 47

We assumed that students interpreted the result of having only seven correct answers
out of 16 questions as a failure because it was below the normal passing standard of 50% in
most secondary schools in Hong Kong. It was possible that some students might be
attempting to predict their own success and did not see the pre-programmed results as
credible. Nevertheless, the actual performance of the students in both summative condition
(X ¼ 6:55; SD = 1.72) and formative condition (X ¼ 5:82; SD = 1.57) were in fact even
lower than the pre-programmed feedback. Therefore, the feedback would have boosted
their self-efficacy instead of dampening it. To improve the accountability of the feedback
manipulation, we could have asked students to indicate the credibility of the feedback. To
address this limitation, we included this manipulation check in Study 2.
To rule out the possibility that students’ self-evaluation and self-efficacy were affected
by their actual performance, we compared their actual performance on both Test 1 and
Test 2. Students in the two conditions had similar performances on both tests. Therefore,
different feedback did not affect actual performance. Nevertheless, students in the for-
mative condition showed a higher sense of success than those in the summative
condition. Their self-efficacy also did not decrease as much as that of their counterparts
in the summative condition. Different types of evaluative feedback might have had
different effects on the mastery experience of students that consequently affected self-
efficacy.
In Study 1, we provided the means of control to some students by giving them formative
feedback and withheld it from others by providing summative feedback. To further explore
the effects of the manipulation of the ends of control on students’ self-efficacy, we
investigated another two types of feedback in Study 2.

Study 2

In Study 2, we used different feedback to highlight different achievement goals as the ends
in the construct of control. We used self-referenced feedback to highlight learning goal and
norm-referenced feedback to highlight performance goal. Oriented towards a learning goal,
students aim to develop their own competence and are more inclined to make effort
attribution. As effort is a controllable factor and students can exercise effort to attain their
goal, their self-efficacy can be strengthened. The performance goal, however, exerts
pressure on students to outperform others and induces ability attribution. As students lack
the means to control their own ability as well as the performance of others, their self-
efficacy would be at risk. Hence, we hypothesized that the self-efficacy of students in the
norm-referenced feedback condition would be lower than that in the self-referenced
feedback condition.

Method

Participants

Participants were 77 Chinese students in Grade 7 who joined a vocabulary builder pro-
gram. They were recruited using the same procedures from the same school as in Study 1.
All parents gave their consent for their children’s participation. As a result, 80 students
were recruited. However, only 77 students showed up on the day of the experiment.

123
48 J. C. Y. Chan, S.-f. Lam

Experimental design

The 77 students were randomly assigned to either the self-referenced feedback condition
(n = 38) or the norm-referenced feedback condition (n = 39). The instructor and sequence
of instruction were the same for both conditions. The only difference was the evaluative
feedback for the tests. The self-referenced condition received feedback in terms of the
number of correct answers and the percentage of accuracy. In contrast, the norm-refer-
enced condition received feedback in terms of the number of correct answers and the
percentile ranking when compared to other students.

Vocabulary acquisition program

Except where otherwise stated, we used the same material for instruction, test format,
questionnaires, and procedures as in Study 1.

Procedures

The experiment took place during a normal English lesson in the school computer room.
First, students did a baseline vocabulary test for subsequent evaluation of whether the initial
performances of the two groups were comparable. Then the students went through two 20-
minute instructional sessions; each included a set of exercises and a test. After completing
each test, students answered a questionnaire on the computer. Since the computers mal-
functioned in our trial run before the experiment, we also asked the students to record their
answers in pencil on the hard copies of the questionnaires. In the first session, the instructor
presented the meaning of six prefixes and examples to illustrate their usage.
After learning six prefixes, students did Test 1. After they completed Test 1, the
computers gave feedback about their performance. After viewing the feedback students
responded to the first questionnaire (Questionnaire A) both on the computer and on paper.
The questions tapped their sense of success about their previous performance and their
self-efficacy with regard to the subsequent instructional session and test (i.e. the second
instructional session and Test 2).
After students had completed the questionnaire, the instructor started the second session
to introduce another six prefixes. Then students did Test 2. The format of Test 2 was the
same as the previous two tests. At the end of Test 2, the computers provided the two
conditions with different feedback. Then students completed the second questionnaire
(Questionnaire B) both on the computer and on paper. It tapped their sense of success about
their previous performance and their self-efficacy with regard to the subsequent instruc-
tional session and test (i.e. the third instructional session and Test 3). After collecting the
questionnaires, the instructor informed students that there was not enough time for the
subsequent instructional session and Test 3. The instructor then debriefed both groups as in
Study 1 and told the norm-referenced condition that comparison with others was actually
less important than the mastery of new skills.

Feedback manipulation

At the end of Test 1, the instructor reminded the students in the norm-referenced condition
that their relative performance in comparison with other students was most important. The
instructor reminded the students in the self-referenced condition that making improvement
in comparison with their previous performance was most important. The computers then

123
Effects of different evaluative feedback 49

displayed the number of correct answers for students. We manipulated the feedback so that
students in both conditions had six questions correct in the baseline test and ten questions
correct in Test 1. In addition, the computers showed the percentage of accuracy for the self-
referenced condition and the percentile ranking of the individual’s performance when
compared to other students for the norm-referenced condition. Moreover, the computers
displayed a bar chart that compared the results between the baseline test and Test 1. For the
self-referenced condition, the bar chart showed 38% accuracy in the baseline test vs. 63%
accuracy in Test 1. For the norm-referenced condition, the bar chart showed 38th percentile
ranking in the baseline test vs. 63rd in Test 1.
At the end of Test 2, the computers showed the results for both groups in a similar
format as Test 1. In addition, the computers displayed a bar chart of the comparison among
the results of the baseline test, Test 1, and Test 2. We manipulated the feedback so that
students in both conditions had 12 correct answers in Test 2. For the self-referenced
condition, the bar chart showed the percentage of accuracy among the three tests which
were 38%, 63%, and 75% respectively. For the norm-referenced condition, the bar chart
showed the percentile ranking among the three tests which were 38th, 63rd and 35th
respectively. The fact that an increase in the absolute number of correct answers corre-
sponded to a drop in ranking in Test 1 but an increase in Test 2 was a reflection of the fact
that percentile ranking can fluctuate despite improvement in actual scores.

Measures

Self-evaluation of success

In both questionnaires students were asked how the results of the previous test (i.e.
Questionnaire A asked about Test 1 and Questionnaire B asked about Test 2) made them
feel. The ratings were made on a seven-point scale from 1 for ‘‘not successful at all’’ to 7
for ‘‘very successful.’’

Self-efficacy

Students rated their self-efficacy with regard to the subsequent lessons and tests (i.e. after
Test 1, they evaluated their self-efficacy with regard to Test 2 and the learning before that
test; after Test 2, they rated that for Test 3 and the learning before that test). We used the
same four questions as in Study 1. The Cronbach’s alpha of these four questions was .79
before Test 2 and .85 before Test 3, demonstrating high internal consistency.

Manipulation checks

In both questionnaires, we asked students whether the previous test result (i.e. Question-
naire A asked about Test 1 and Questionnaire B asked about Test 2) could show how much
they had learned about prefixes. This provided us insight into whether students perceived
the artificial feedback to be credible. The ratings were made on a seven-point scale, from 1
for ‘‘not at all’’ to 7 for ‘‘very much.’’ In Questionnaire B, students indicated whether they
had learned more prefixes after the two lessons and the number of correct answers they got
for Test 2. The purpose was to see if they recalled that they had the majority of the
questions correct. Students also evaluated the clarity of teaching and noise level of the
classroom on a seven-point scale as in Study 1.

123
50 J. C. Y. Chan, S.-f. Lam

Results

Performance

We compared students’ actual performances on the two tests. Unfortunately, some com-
puters had technical problems and failed to record the test results of about one-third of the
students. In the self-referenced condition, the results of 25 students were intact and 13 were
lost. In the norm-referenced condition, the results of 26 students were intact and 13 were
lost. To ensure that the group whose test results were lost was comparable to the group
whose test results were retained, we examined the manipulation checks of these two groups
of students. Within each condition, these two groups of students had no difference in their
perception of the credibility of their test results, the clarity of the instruction, the noise
level of the classroom, whether they have learned more about prefixes, and the number of
correct answers they got in Test 2, all p [ .05. The analyses indicated that the attrition in
test results had not produced a biased sample.
In the baseline test, the performance of the self-referenced condition (X ¼ 8:28;
SD = 3.29) and the norm-referenced condition (X ¼ 8:33; SD = 2.47) were comparable,
t(50) = -.067, p = .95, ns, Cohen’s d = .02. In Test 1, the performance of the self-
referenced condition (X ¼ 8:20; SD = 1.98) and the norm-referenced condition (X ¼ 9:00;
SD = 1.14) were also comparable, t(50) = -1.80, p = .08, ns, Cohen’s d = .51. In Test
2, the performance of the students in the self-referenced condition (X ¼ 9:20; SD = 2.25)
and the norm-referenced condition (X ¼ 8:67; SD = 2.08) also did not differ, t(50) = .89,
p = .38, ns, Cohen’s d = .31.

Self-evaluation of success

Besides performance, the rest of the data was intact. We compared students’ self-evalu-
ation of their success after Test 1 and Test 2. The self-referenced condition (X ¼ :43;
SD = 1.27) felt more successful than the norm-referenced condition (X ¼ :24;
SD = 1.36) and the difference was statistically significant with a medium effect,
t(70) = 2.16, p = .03, Cohen’s d = .52.

Self-efficacy

We performed a mixed model Univariate Analysis of Variance (ANOVA) with self-

efficacy (Test 2 vs. Test 3) as the within-participant factor and condition (self-referenced
vs. norm-referenced feedback) as the between-participant factor. The ANOVA yielded a
significant main effect for condition, F(1, 73) = 9.31, p \ .01, partial g2 = .11. The main
effect for self-efficacy was not significant, F(1, 73) = 3.75, p = .06, partial g2 = .05 but
the two-way interaction between self-efficacy and condition was significant, F(1,
73) = 45.45, p = .01, g2 = .39.
To understand this two-way interaction, we decomposed it into a series of t-tests,
examining participants’ self-efficacy for Test 2 and Test 3 (see Table 2). The self-refer-
enced condition experienced a significant increase in self-efficacy from Test 2 (X ¼ 4:83;
SD = .84) to Test 3 (X ¼ 5:39; SD = .79), t(36) = 6.55, p \ .01, Cohen’s d = .67. The
effect size was medium. In contrast, the norm-referenced condition suffered a significant
decrease in self-efficacy from Test 2 (X ¼ 4:72; SD = .78) to Test 3 (X ¼ 4:41;
SD = .89), t(37) = -3.22, p \ .01, Cohen’s d = -.37. The two conditions started out
with no difference in self-efficacy for Test 2 (4.83 vs. 4.72), t(73) = .59, p = .56, Cohen’s

123
Effects of different evaluative feedback 51

Table 2 Means of self-efficacy by condition for Study 2

Self-efficacy Conditions t-value Cohen’s d

Self-referenced, n = 37 Norm-referenced, n = 38

Test 2 4.83a (.84) 4.72b (.78) .59 .14

ac
Test 3 5.39 (.79) 4.41bc (.89) 5.08 1.18
Difference of Tests 2 & 3 .56d (.52) -.32d (.59) 6.74 1.56
Note: The means sharing the same superscript are significantly different from each other. The numbers in
parentheses are standard deviations

d = .14. However, their self-efficacy for Test 3 after the feedback showed a significant
difference with a large effect size (5.39 vs. 4.41), t(74) = .59, p \ .05, Cohen’s d = 1.18.
The difference between self-efficacy for Test 2 and Test 3 was .56 (SD = .52) for the self-
referenced condition but -.32 (SD = .59) for the norm-referenced condition. The differ-
ence between the two conditions was statistically significant with a large effect size,
t(73) = 6.74, p \ .01, Cohen’s d = 1.56.

Manipulation checks

Students indicated the extent to which the test results reflected their prefix acquisition. The
self-referenced condition rated both tests positively (above the neutral point of four) for
showing how much they had learned (for Test 1, X ¼ 4:66; SD = 1.26 and for Test 2,
X ¼ 4:97; SD = 1.48) and the same applied to the norm-referenced condition (for Test 1,
X ¼ 4:73; SD = 1.28, and for Test 2, X ¼ 4:49; SD = 1.35). There was no significant
difference between the two groups’ evaluation for both Test 1, t(70) = -.24, p = .81,
Cohen’s d = .06, and Test 2, t(74) = 1.49, p = .14, Cohen’s d = .34. Thus we inferred
that the results of both tests appeared to be credible for both conditions.
There were 92.1% of students in the self-referenced condition and 89.7% of students in
the norm-referenced condition who indicated that they had learned more about prefixes
after the two lessons. The difference between the two conditions was not statistically
significant, v2(77) = 1.32, p = .52. Furthermore, 94.9% of the self-referenced condition
and 94.6 % of the norm-referenced condition remembered accurately that they had 12
answers correct in the previous test and there was no statistically significant difference,
v2(76) = .003, p = .96.
The two conditions only differed in their evaluation of the clarity of instruction. The
norm-referenced condition rated the teaching as above the level of being ‘‘quite clear’’
(X ¼ 5:44; SD = 1.12), while the self-referenced condition gave an even higher rating
(X ¼ 6:08; SD = 1.01) and the difference was statistically significant with a medium
effect, t(74) = 2.63, p \ .05, Cohen’s d = .61. With regard to the noise level of the
classroom, the self-referenced condition (X ¼ 2:41; SD = 1.21) and the norm-referenced
condition (X ¼ 2:85; SD = 1.50) showed no significant difference, t(74) = -1.41,
p = .16, Cohen’s d = .33.

Discussion

In Study 2, as hypothesized, students who received norm-referenced feedback showed

lower self-efficacy than those who received self-referenced feedback. The difference in the

123
52 J. C. Y. Chan, S.-f. Lam

change in self-efficacy between the two conditions was significant with a large effect size
of 1.56.
In the self-referenced condition, students as the agent of control used effort as the means
to reach their learning goal, the ends of control. This belief in the agent–means relation
strengthened self-efficacy. On the contrary, students in the norm-referenced condition were
constantly being compared to other students. As it was impossible for them to perceive
control over the performance of others, they had no means to outperform others. Thus, their
self-efficacy decreased. Despite the fact that students observed their progressive
improvement in terms of the absolute number of correct answers, the fluctuation in their
percentile ranking had reinforced their helplessness. This finding echoes the observation of
Elliott and Dweck (1988) that when individuals believe that they have low ability, they will
react in a learned helplessness manner in face of performance goal.
There is concern that the evaluative feedback might have presented a success scenario
for the self-referenced condition but a failure scenario for the norm-referenced condition.
However, it is important to note that both groups received identical results in terms of the
number of correct answers throughout the program. Nevertheless, the type of evaluative
feedback administered did have an effect on the students’ self-evaluation of success. The
self-referenced condition showed an overall increase in the sense of success while the
norm-referenced condition showed an overall decrease, with a medium effect size of .52.
This difference may be because the norm-referenced condition perceived no means to
control others’ performance, and thus their own percentile ranking. Consequently, they
were more likely to interpret the experience as a setback.
To rule out the possibility that the self-efficacy and sense of success of students were
influenced by their actual performance, we compared their actual performance in Test 1
and Test 2. Within the two-thirds of data retained, there was no difference between the two
conditions.
In the manipulation check, the norm-referenced condition rated the clarity of teaching as
above ‘‘quite clear,’’ while the self-referenced condition gave an even higher rating. Given
that both groups perceived the clarity of teaching as above ‘‘quite clear,’’ we had confi-
dence that the instructions were clear enough.

General discussion

The present paper investigated how different types of evaluative feedback affected stu-
dents’ self-efficacy in vocabulary acquisition. Through different types of evaluative
feedback, teachers oriented students towards different achievement goals. Formative and
self-referenced feedback encouraged students to adopt a learning goal that emphasized
incremental self-improvement; on the contrary, summative and norm-referenced feedback
pointed towards a performance goal that defined success as outperforming others.
The results were consistent with findings in the literature on goal-theory of achievement
motivation (Ames 1992). In Study 1, formative feedback provided strategic information for
students to make improvement. Learning strategies, as the means, promised a sense of
control over academic goals and strengthened the means–ends relation in control. Thus,
students, as agents, would perceive more control in the agent–means relation that could
either enhance their self-efficacy or lessen the decrease in self-efficacy when they face
setbacks. In contrast, summative feedback prompted students to produce a quality end
product without providing strategic information. Students were less likely to perceive

123
Effects of different evaluative feedback 53

control over the means–ends relation, as well as the agent–means relation. Hence, it would
be more difficult for them to maintain their self-efficacy.
These findings imply that students’ self-efficacy can benefit from teachers’ use of
formative feedback. Specifically, instead of summarizing the strengths and weaknesses of
students at the end of their performance, teachers can try to provide strategies for students
to make improvement during a task. This may be coupled with other techniques, for
example, verbal persuasion like ‘‘You can do it’’ (Schunk and Zimmerman 2007) to
convince students of their capability.
The findings of Study 2 also cast legitimate doubts on the appropriateness of norm-
referenced feedback wherein student performance is evaluated comparatively (Marsh et al.
2000). In reality, teachers with their expertise and power (Oettingen 1995) control the
assessment criteria and mode of evaluation (Buunk et al. 1990) and students passively
receive information about how their capabilities are compared to their peers’ (Bandura
1999). As explained earlier, such information affects students’ self-appraisal of ability, and
thus their academic self-efficacy (Oettingen 1995). Therefore, teachers should consider
using self-referenced feedback instead.

Limitations and merits

The present study demonstrated how formative, summative, norm-referenced, and self-
referenced feedback could deliver immediate impact on students’ self-efficacy despite the
brief evaluative session which might not be representative of an actual school experience.
Though this limited timeframe of intervention showed effects of different feedback, it fell
short of demonstrating whether these effects might be maintained over time. Nevertheless,
if such feedback were to be administered regularly over time, there is reason to believe that
there would be long-term effects on students. The brief experiments, which consisted of
only two consecutive class periods, limited any prolonged detrimental effects that the
norm-referenced and summative measures might have on the participants.
Students within the same experimental condition received identical and pre-pro-
grammed feedback throughout both studies. The only exception was the last test in Study 1
in which students received genuine feedback about their performance. The pre-
programmed feedback might have led some students to question the credibility of the
results. However, the manipulation checks in Study 2 indicated that students generally
found the results credible and reflective of their actual performance.
With reference to previous experimental studies (Butler 1993; Elliott and Dweck 1988),
the present research highlighted learning and performance goals through verbal instruction
and evaluation objectives. Measuring goal orientation was duly considered, but ultimately
rejected for the following reasons: First, previous research has shown that even when
students come to an experiment with learning goal, experimental manipulation in the form
of competition, for example, that activates performance goal might still affect learning
motivation and result in the worsening of self-evaluation after failure (Lam et al. 2004).
This provides evidence that the teacher-assigned goals are likely to override the individual
goals that students bring to the classroom. Second, the specificity and transient nature of
classroom goals or task goals that were the major foci of the present research made the
stable goal orientation inadequate as it might not reflect the students’ on-task goals.
In the present studies, the scales that measured self-efficacy had students evaluate their
level of confidence in or control over learning and performance. The purpose was to
highlight the agent–means relation in the concept of control. In fact, self-efficacy is the
agent’s perceived controllability over the means of goal achievement. We took into

123
54 J. C. Y. Chan, S.-f. Lam

account several criteria in developing this scale. The first criterion was that judgments of
self-efficacy are task and domain specific so measures of self-efficacy should be tailored
specifically to the task being assessed (Pajares and Miller 1995). Hence, the questions
prompted students to evaluate their self-efficacy with regard to ‘‘the next lesson’’ and ‘‘the
next test.’’ The aim was to focus their attention on evaluating their self-efficacy with regard
to the task alone.
The second criterion was that self-efficacy is about prospective situations (Bandura
1997). Hence, all of the questions on self-efficacy were written in the future tense to
highlight the concern about future performance rather than past or present performance.
Students were asked to evaluate their self-efficacy specifically with regard to the learning
of the prefixes and doing well in the upcoming lessons and tests. The high internal con-
sistency of the self-efficacy items was reflected in the Cronbach’s alpha coefficients
ranging from .79 to .89.
In response to the demand for more studies on self-efficacy in actual classrooms
(Schunk and Zimmerman 1997), we endeavored to conduct tightly controlled experiments
in a classroom during school hours to maximize internal and external validities. We chose
vocabulary acquisition as the subject of learning by virtue of its value in language
acquisition. Despite its importance, this area has been neglected while research on aca-
demic self-efficacy has examined reading and comprehension (e.g. Schunk and Rice 1986,
1987, 1989), writing (e.g. Schunk and Swartz 1993a, 1993b), learning mathematics (e.g.
Hackett 1985; Reyes 1984; Schunk 1982; Zimmerman and Martinez-Pons 1990), and
computer skill acquisition (Schunk and Ertmer 1999). Previous research on students’ self-
efficacy in literary activities seldom looked at how students acquire a second language.
More specifically, this study focused on how Chinese students learned English vocabulary.

Cultural factors

The present research employed experimental controls in a field setting with Chinese stu-
dents as participants. Contemporary Chinese students, teachers, and parents are all aware
of the competition that prevails on the path to higher levels of schooling (Stevenson and
Lee 1996). Students in this culture are accustomed to the competition and high prevalence
of normative evaluation in classrooms (Biggs 1996; Shih and Alexander 2000). Never-
theless, the participants were still not immunized against the negative effects of norm-
referenced feedback in Study 2. Teachers should endeavor to minimize such negative
effects when they design and conduct evaluation.
Besides Chinese culture, other cultures also rely heavily on normative assessment
(France-Kaatrude and Smith 1985). For instance, American middle schools emphasize
relative ability and social comparison in learning (Midgley et al. 1995) where the success
and failure of students are determined comparatively (Levine 1983; Richer 1976). Hence
future studies can investigate the effects of teachers’ feedback on students’ self-efficacy
across cultures under similar educational regimes.

Future research directions

Schunk and Zimmerman (1997) called for more research on the role of social origins such
as models, verbal description, social guidance and feedback in students’ academic com-
petence. In response, this paper investigated the effects of teachers’ feedback as an
important source of students’ mastery experience and social persuasion on students’ self-
efficacy. Future research would entail studies to gain a better understanding of other

123
Effects of different evaluative feedback 55

sources of self-efficacy and the relations among them. For example, researchers have
already shown that enhancing the physiological and psychological well-being of students
and reducing their negative emotions can strengthen self-efficacy (Usher and Pajares
2006). It is therefore desirable to obtain more understanding about how students can
improve their physical and emotional well-being as well as how to empower students to
interpret their physiological and psychological experiences positively.
According to Ames (1992), teachers’ own goal orientation can influence their beliefs
about the efficacy of instructional decisions and teaching methods. Future research should
address the impact of teachers’ goal orientation and their preference for evaluative feed-
back on students’ self-efficacy, including advantages and disadvantages of alternative
evaluations. As teachers play an important role in students’ goal-setting and self-efficacy,
more research needs to address the effects of different classroom practices and the inter-
action between teachers and students.
Adolescents spend a major portion of their time at school, where they are constantly
exposed to school-related stress such as academic standing, examinations, competition and
conflict among peers, and relationships with teachers (Fanshawe and Burnett 1999; Thuen
and Bru 2004). As self-efficacy is related positively to positive schooling outcomes such as
motivation and achievement, more research is needed to address the self-efficacy of
adolescents at school.

Conclusion

For decades, educators have struggled with the challenge of making the learning envi-
ronment effective for early adolescent students (Urdan et al. 1995). Considering how to
enhance students’ self-efficacy may be a step towards building an effective learning
environment. In this research, we showed that teachers, through different evaluative
feedback, affected the self-efficacy of students. Formative and self-referenced feedback
that enabled students to perceive a sense of control over their progress were beneficial to
their self-efficacy. Conversely, summative and norm-referenced feedback that weakened
students’ perceived control over the achievement outcome threatened their self-efficacy.
According to Bandura (1997), students’ appraisals of their capabilities are largely affected
by the relative emphasis on social comparative versus self-comparative appraisal they
receive in the classroom. ‘‘Socially competitive grading practices convert educational
experiences into ones where many are doomed to failure for the high success of a few’’
(Bandura 1997, p. 175). Having the responsibility to design and provide feedback, teachers
could employ more self-referenced evaluation and provide more learning strategies in their
feedback to empower the self-efficacy of students.

References

Ames, C. (1992). Classrooms: Goals, structures, and student motivation. Journal of Educational Psychol-
ogy, 84, 261–271. doi:10.1037/0022-0663.84.3.261.
Ames, C., & Ames, R. (1984). Systems of student and teacher motivation: Toward a qualitative definition.
Journal of Educational Psychology, 76, 535–556. doi:10.1037/0022-0663.76.4.535.
Ames, C., & Archer, J. (1987). Mothers’ beliefs about the role of ability and effort in school learning.
Journal of Educational Psychology, 79, 409–414. doi:10.1037/0022-0663.79.4.409.

123
56 J. C. Y. Chan, S.-f. Lam

Ames, C., & Archer, J. (1988). Achievement goals in the classroom: Students’ learning strategies and
motivation processes. Journal of Educational Psychology, 80, 260–267. doi:10.1037/0022-
0663.80.3.260.
Anderman, E. M., & Maehr, M. L. (1994). Motivation and schooling in the middle grades. Review of
Educational Research, 64, 287–309.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review,
84, 191–215. doi:10.1037/0033-295X.84.2.191.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs,
NJ: Prentice Hall.
Bandura, A. (1989). Human agency in social cognitive theory. The American Psychologist, 44, 1175–1184.
doi:10.1037/0003-066X.44.9.1175.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W.H. Freeman & Co.
Bandura, A. (1999). Exercise of personal and collective efficacy in changing societies. In A. Bandura (Ed.),
Self-efficacy in changing societies (pp. 1–45). New York: Cambridge University Press.
Barron, K. E., & Harackiewicz, J. M. (2001). Achievement goals and optimal motivation: Testing multiple
goal models. Journal of Personality and Social Psychology, 80, 706–722. doi:10.1037/0022-
3514.80.5.706.
Biggs, J. (1996). Western misperceptions of the Confucian-Heritage learning culture. In D. A. Watkins & J.
B. Biggs (Eds.), The Chinese learner: Cultural psychological and contextual influences (pp. 45–67).
Hong Kong: Comparative Education Research Centre.
Butler, R. (1993). Effects of task- and ego-achievement goals on information seeking during task
engagement. Journal of Personality and Social Psychology, 65, 18–31. doi:10.1037/0022-3514.65.1.18
.
Buunk, B. P., Collins, R. L., Taylor, S. E., VanYperen, N. W., & Dakof, G. A. (1990). The affective
consequences of social comparison: Either direction has its ups and downs. Journal of Personality and
Social Psychology, 59, 1238–1249. doi:10.1037/0022-3514.59.6.1238.
Cervone, D. (1993). The role of self-referenced cognitions in goal setting, motivation, and performance. In
M. Rabinowitz (Ed.), Cognitive science foundations of instruction (pp. 57–96). Hillsdale, NJ: Law-
rence Erlbaum.
Cunningham, A. E., & Stanovich, K. E. (1997). Early reading acquisition and its relation to reading
experience and ability 10 years later. Developmental Psychology, 33, 934–945. doi:10.1037/0012-
1649.33.6.934.
Dweck, C. S. (1986). Motivational processes affecting learning. American Psychologist, 41, 1040–1048.
doi:10.1037/0003-066X.41.10.1040.
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and personality. Psy-
chological Review, 95, 256–273. doi:10.1037/0033-295X.95.2.256.
Dykeman, B. F. (1994). The effects of motivational orientation, self-efficacy, and feedback condition on test
anxiety. Journal of Instructional Psychology, 21, 114–120.
Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation and achievement. Journal of
Personality and Social Psychology, 54, 5–12. doi:10.1037/0022-3514.54.1.5.
Elliot, A. J., & Harackiewicz, J. M. (1996). Approach and avoidance achievement goals and intrinsic
motivation: A mediational analysis. Journal of Personality and Social Psychology, 70, 461–475.
doi:10.1037/0022-3514.70.3.461.
Elliot, A. J., & McGregor, H. A. (2001). A 2 9 2 achievement goal framework. Journal of Personality and
Social Psychology, 80, 501–519. doi:10.1037/0022-3514.80.3.501.
Fanshawe, J. P., & Burnett, P. C. (1999). Students’ and teachers’ perceptions of adolescents’ problems and
coping strategies. Australian Journal of Guidance & Counselling, 9, 93–107.
France-Kaatrude, A.-C., & Smith, W. P. (1985). Social comparison, task motivation, and the development of
self-evaluative standards in children. Developmental Psychology, 21, 1080–1089. doi:10.1037/0012-
1649.21.6.1080.
Garcı́a, J.-N., & de Caso, A. M. (2004). Effects of a motivational intervention for improving the writing of
children with learning disabilities. Learning Disability Quarterly, 27, 141–159. doi:10.2307/1593665.
Graham, S., Harris, K. R., & Mason, L. (2005). Improving the writing performance, knowledge, and self-
efficacy of struggling young writers: The effects of self-regulated strategy development. Contemporary
Educational Psychology, 30, 207–241. doi:10.1016/j.cedpsych.2004.08.001.
Hackett, G. (1985). The role of mathematics self-efficacy in the choice of math-related majors of college
women and men: A path analysis. Journal of Counseling Psychology, 32, 47–56. doi:10.1037/0022-
0167.32.1.47.

123
Effects of different evaluative feedback 57

Hau, K. T., & Salili, F. (1996). Prediction of academic performance among Chinese students: Effort can
compensate for lack of ability. Organizational Behavior and Human Decision Processes, 65, 83–94.
doi:10.1006/obhd.1996.0008.
Lam, S. F., Yim, P. S., Law, J. S. F., & Cheung, R. W. Y. (2004). The effects of competition on achievement
motivation in Chinese classrooms. British Journal of Educational Psychology, 74, 281–296.
doi:10.1348/000709904773839888.
Levine, J. M. (1983). Social comparison and education. In J. M. Levine & M. C. Wang (Eds.), Teacher and
student perceptions: Implications for learning (pp. 29–55). New York: Erlbaum.
Linnenbrink, E. A., & Pintrich, P. R. (2002). Motivation as an enabler for academic success. School
Psychology Review, 31, 313–327.
Marsh, H. W., Kong, C.-K., & Hau, K. T. (2000). Longitudinal multilevel models of the big-fish-little-pond
effect on academic self-concept: Counterbalancing contrast and reflected-glory effect in Hong Kong
schools. Journal of Personality and Social Psychology, 78, 337–349. doi:10.1037/0022-3514.78.2.337.
McAlpine, L. (2004). Designing learning as well as teaching. A research-based model for instruction that
emphasizes learner practice. Learning in Higher Education, 5, 119–134. doi:10.1177/
1469787404043809.
Midgley, C., Anderman, E., & Hicks, L. (1995). Differences between elementary and middle school teachers
and students: A goal theory approach. The Journal of Early Adolescence, 15, 90–113. doi:10.1177/
0272431695015001006.
Nicholls, J. G. (1979). Development of perception of own attainment and causal attribution for success and
failure in reading. Journal of Educational Psychology, 75, 94–99. doi:10.1037/0022-0663.71.1.94.
Oettingen, G. (1995). Cross-cultural perspectives on self-efficacy. In A. Bandura (Ed.), Self-efficacy in
changing societies (pp. 149–176). Cambridge: Cambridge University Press.
Pajares, F. (1996). Self-efficacy beliefs in academic settings. Review of Educational Research, 66, 543–578.
Pajares, F., & Miller, M. D. (1995). Mathematics self-efficacy and mathematics performances: The need for
specificity of assessment. Journal of Counseling Psychology, 42, 190–198. doi:10.1037/0022-
0167.42.2.190.
Penno, J. F., Wilkinson, I. A. G., & Moore, D. W. (2002). Vocabulary acquisition from teacher explanation
and repeated listening to stories: Do they overcome the Matthew effect? Journal of Educational
Psychology, 94, 23–33. doi:10.1037/0022-0663.94.1.23.
Popham, W. J. (2001). Teaching to the test? Educational Leadership, 58, 16–20.
Reyes, L. H. (1984). Affective variables and mathematics education. The Elementary School Journal, 84,
558–582. doi:10.1086/461384.
Richer, S. (1976). Reference-group theory and ability grouping: A convergence of sociological theory and
educational research. Sociology of Education, 49, 65–71. doi:10.2307/2112394.
Schunk, D. H. (1982). Effects of effort attributional feedback on children’s perceived self-efficacy and
achievement. Journal of Educational Psychology, 74, 548–556. doi:10.1037/0022-0663.74.4.548.
Schunk, D. H. (1991). Goal setting and self-evaluation: A social cognitive perspective on self-regulation.
Advances in Motivation and Achievement, 7, 85–113.
Schunk, D. H. (2003). Self-efficacy for reading and writing: Influence of modeling, goal setting, and self-
evaluation. Reading & Writing Quarterly, 19, 159–172. doi:10.1080/10573560308219.
Schunk, D. H., & Ertmer, P. A. (1999). Self-regulatory processes during computer skill acquisition: Goal
and self-evaluative influences. Journal of Educational Psychology, 91, 251–260. doi:10.1037/0022-
0663.91.2.251.
Schunk, D. H., & Rice, J. M. (1986). Extended attributional feedback: Sequence effects during remedial
reading instruction. The Journal of Early Adolescence, 6, 55–66. doi:10.1177/0272431686061005.
Schunk, D. H., & Rice, J. M. (1987). Enhancing comprehension skill and self-efficacy with strategy value
information. Journal of Reading Behavior, 19, 285–302.
Schunk, D. H., & Rice, J. M. (1989). Learning goals and children’s reading comprehension. Journal of
Reading Behavior, 21, 279–293.
Schunk, D. H., & Rice, J. M. (1991). Learning goals and progress feedback during reading comprehension
instruction. Journal of Reading Behavior, 23, 351–364.
Schunk, D. H., & Rice, J. M. (1992). Influence of reading comprehension strategy information on children’s
achievement outcomes. Learning Disability Quarterly, 15, 51–64. doi:10.2307/1510565.
Schunk, D. H., & Rice, J. M. (1993). Strategy fading and progress feedback: Effects on self-efficacy and
comprehension among students receiving remedial reading services. The Journal of Special Education,
27, 257–276.
Schunk, D. H., & Swartz, C. W. (1993a). Goals and progress feedback: Effects on self-efficacy and writing
achievement. Contemporary Educational Psychology, 18, 337–354. doi:10.1006/ceps.1993.1024.

123
58 J. C. Y. Chan, S.-f. Lam

Schunk, D. H., & Swartz, C. W. (1993b). Writing strategy instruction with gifted students: Effects of goals
and feedback on self-efficacy and skills. Roeper Review, 15, 225–230.
Schunk, D. H., & Zimmerman, B. J. (1997). Social origins of self-regulatory competence. Educational
Psychologist, 32, 195–208. doi:10.1207/s15326985ep3204_1.
Schunk, D. H., & Zimmerman, B. J. (2007). Influencing children’s self-efficacy and self-regulation of
reading and writing through modeling. Reading & Writing Quarterly, 23, 7–25. doi:10.1080/
10573560600837578.
Shih, S. S., & Alexander, J. M. (2000). Interacting effects of goal setting and self- or other-referenced
feedback on children’s development of self-efficacy and cognitive skill within the Taiwanese class-
room. Journal of Educational Psychology, 92, 536–543. doi:10.1037/0022-0663.92.3.536.
Skaalvik, E. M. (1997). Self-enhancing and self-defeating ego orientation: Relations with task and avoid-
ance orientation, achievement, self-perceptions, and anxiety. Journal of Educational Psychology, 89,
71–81. doi:10.1037/0022-0663.89.1.71.
Skinner, E. (1996). A guide to constructs of control. Journal of Personality and Social Psychology, 71, 549–
570. doi:10.1037/0022-3514.71.3.549.
Spearitt, D. (1972). Identification of subskills of reading comprehension by maximum likelihood factor
analysis. Reading Research Quarterly, 8, 92–111. doi:10.2307/746983.
Stevenson, H. W., & Lee, S.-Y. (1996). The academic achievement of Chinese students. A function of
socialization processes and the context of learning? In M. Bond (Ed.), The handbook of Chinese
psychology (pp. 124–142). Hong Kong: Oxford University Press.
Thuen, E., & Bru, E. (2004). Coping styles and emotional and behavioral problems among Norwegian Grade
9 students. Scandinavian Journal of Educational Research, 48, 493–510. doi:10.1080/
003138042000272140.
Urdan, T. (2001). Statistics in plain English. London: Lawrence Erlbaum.
Urdan, T. C., & Maehr, M. L. (1995). Beyond a two-goal theory of motivation and achievement: A case for
social goals. Review of Educational Research, 6, 213–243.
Urdan, T., Midgley, C., & Wood, S. (1995). Special issues in reforming middle level grades. The Journal of
Early Adolescence, 15, 9–37. doi:10.1177/0272431695015001003.
Usher, E. L., & Pajares, F. (2006). Sources of academic and self-regulatory efficacy beliefs of entering
middle school students. Contemporary Educational Psychology, 31, 125–252. doi:10.1016/
j.cedpsych.2005.03.002.
Walker, D., Greenwood, C. R., Hart, B., & Carta, J. (1994). Special children and poverty. Child Devel-
opment, 65, 606–621. doi:10.2307/1131404.
Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer.
Zimmerman, B. J. (1995). Self-efficacy and educational development. In A. Bandura (Ed.), Self-efficacy in
changing societies (pp. 202–231). New York: Cambridge University Press.
Zimmerman, B. J. (1997). Becoming a self-regulated writer: A social cognitive perspective. Contemporary
Educational Psychology, 22, 73–101. doi:10.1006/ceps.1997.0919.
Zimmerman, B. J., & Kitsantas, A. (1997). Developmental phases in self-regulation: Shifting from process
goals to outcome goals. Journal of Educational Psychology, 89, 29–36. doi:10.1037/0022-
0663.89.1.29.
Zimmerman, B. J., & Kitsantas, A. (2002). Acquiring writing revision and self-regulatory skill through
observation and emulation. Journal of Educational Psychology, 94, 660–668. doi:10.1037/0022-
0663.94.4.660.
Zimmerman, B. J., & Martinez-Pons, M. (1990). Student differences in self-regulated learning: Relating
grade, sex, and giftedness to self-efficacy and strategy use. Journal of Educational Psychology, 82, 51–
59. doi:10.1037/0022-0663.82.1.51.

123
View publication stats