Academic Emotions in Students’ Self-Regulated Learning and
Achievement: A Program of Qualitative and
Quantitative Research
Reinhard Pekrun, Thomas Goetz, and Wolfram Titz
Department of Psychology
University of Munich, Germany
Raymond P. Perry
Department of Psychology
University of Manitoba, Canada
Academic emotions have largely been neglected by educational psychology, with the exception
of test anxiety. In 5 qualitative studies, it was found that students experience a rich diversity of
emotions in academic settings. Anxiety was reported most often, but overall, positive emotions
were described no less frequently than negative emotions. Based on the studies in this article,
taxonomies of different academic emotions and a self-report instrument measuring students’
enjoyment, hope, pride, relief, anger, anxiety, shame, hopelessness, and boredom (Academic
Emotions Questionnaire [AEQ]) were developed. Using the AEQ, assumptions of a cogni-
tive-motivationalmodelofthe achievementeffects ofemotions, andof acontrol/value theoryof
their antecedents (Pekrun, 1992b, 2000), were tested in 7 cross-sectional, 3 longitudinal, and 1
diarystudyusing samplesofuniversity andschoolstudents. Resultsshowed that academicemo-
tions are significantly related to students’ motivation, learning strategies, cognitive resources,
self-regulation,andacademic achievement,as wellas topersonalityand classroomantecedents.
The findings indicate that affective research in educational psychology should acknowledge
emotional diversity in academic settings by addressing the full range of emotions experienced
by students at school and university.
In the following analysis, we present an overview of our re-
search on students’ academic emotions. This research origi-
nated through Reinhard Pekrun’s interests in students’ test
anxiety and motivation,and RaymondPerry’s studies on aca-
demic control and motivation (e.g., Pekrun, 1991, 1992a;
Perry, 1991; Perry & Magnusson, 1989). Within the perspec-
tive of test anxiety and motivation research, we have found
that test anxiety has been researched extensively sincethe be-
ginning of the 1950s (Mandler & Sarason, 1952) and even
prior to that (cf. Stengel, 1936), whereas students’ academic
emotions, other than anxiety, have been largely neglected.
The only major exception is the attributional research on
achievement emotions undertaken by Weiner (cf. Weiner,
1985). Based on innovative theorizing, this research has
produced a sizable number of studies as well as consistent,
cumulative empirical evidence on the cognitive antecedents
of achievement-related emotions.
Table 1 presents the results of a literature search of rele-
vant studies. For the time span from 1974 to 1990, this search
was originally published in Pekrun and Frese (1992) and was
updatedhere for thedecade ofthe 1990s, using thePsycINFO
database. The search extends beyond the educational domain
andincludes any empiricalstudies linking theemotions listed
in Table 1 to the topics of learning, test, performance, work,
or achievement. It appears that test anxiety has continued to
attract many researchers, whereas other achievement-related
emotions have received much less attention. This pertains to
negativeemotions other thanjust anxiety,but even more soto
positive achievement-related emotions. For example,
whereas more than 1,000 studies have addressed achieve
-
ment-related anxiety to date, we were not able to locate more
Requests for reprints should be sent to Reinhard Pekrun, Department of
Psychology, University of Munich, Leopoldstrasse 13, 80802 Munich, Ger
-
many. E-mail: pekrun@edupsy.uni-muenchen.de
First publ. in: Educational Psychologist 37 (2002), 2, pp. 91-106
Konstanzer Online-Publikations-System (KOPS)
URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-138856
than 9 studies on the complementary emotion of students’
hope. Also, there is a dearth of studies on students’ regula-
tionoftheiremotions(cf.Schutz&Davis,2000).Moststud-
ies were conducted from the perspective of specific
traditionsof research and addressedno morethanone or two
emotions, implying that a broader perspective on a range of
emotions was largely lacking in much of the empirical re-
search to date.
Theoretically, it may be assumed that students experience
a rich diversity of emotions in academic settings. Academic
learning and achievement are among the most important top-
ics across the life span in our society today, especially be-
cause educational and professional careers, social relations,
and the allocation of many kinds of resources are largely de-
pendent on individual achievement. This implies that learn-
ing and achievement are singularly important and thus major
sources of human emotions today, instigating a variety of
self-referenced, task-related, and social emotions (cf. also
Scherer, Wallbott, & Summerfield, 1986). Furthermore,
judging from the general functions of emotions for human
agency, it may be assumed that emotions influence students’
cognitiveprocesses and performance,as well astheir psycho-
logical and physical health.
In particular, this perspective is warranted for emotions
that are directly linked to academic learning, classroom in-
struction,andachievement(e.g.,enjoymentoflearning, pride
of success, or test-related anxiety). We propose to use the
term academic emotions to denote such emotions. By defin-
ing academic emotions inthis way,the term academic is used
as is commonlydone with termssuch as academic motivation
or academic self-concept. In doing so, the domain of aca
-
demic emotions would include students’ achievement emo
-
tions experienced in school or university settings, but goes
beyond emotions relating to success and failure by also cov
-
ering, for example, emotions relating to instruction or to the
process of studying (see Table 2).
The following research questions guided our studies on
academic emotions:
1. Which emotions do students experience in academic
settings when attending class, studying, and taking tests and
exams? Furthermore, what are the elements of these emo-
tional experiences and how are they structured?
2. How can we measure students’ academic emotions?
3. How do these emotions affect learning, academic
achievement, and students’ health?
4. What are the origins of these emotions within students’
personality and in their environments?
5. What can we do to foster positive academic emotions
and to help students avoid negative emotions, or to cope with
negative emotions in flexible ways once they emerge?
To date, our research has mainly addressedthe firstfour of
these questions, which may be regarded as basic and have to
beanswered before educationalapplications canbe designed.
In addressing these questions, we used qualitative as well as
quantitative methods. The logic underlying our strategy im-
plies that qualitative, exploratory analysis may be best suited
to develop insights into the range and phenomenology of stu-
dents’ emotions and to generate hypotheses, whereas quanti-
tative assessment is needed for the more precise analysis of
effects and causes.
In the following, we give an overview of our research per-
taining tothese four questions. Specifically, in the sections of
this article we address the following parts of our research in
turn:
1. In a series of qualitative case studies, we explored the
occurrence and phenomenological structures of academic
emotions.
2. Based on our exploratory evidence, we developed a
quantitative self-report instrument measuring nine academic
92
TABLE 1
Literature
Search 1974–2000: Studies Linking Emotions to
Learning and Achievement
Emotion 1974–1990 1991–2000 Research Tradition
Joy 32 29 Mood research
Enthusiasm 9 7 Teacher enthusiasm
Hope 0 9
Relief 2 1
Pride 17 10 Achievement motivation
Gratitude 2 1
Admiration 0 0
Sadness 10 5 Mood research
Anger 31 33 Type A personality
Anxiety > 700 > 500 Test anxiety
Hopelessness 2 12 Hopelessness theory
Shame and Guilt 24 20 Achievement motivation
Disappointment 2 0
Boredom 27 16 Job monotony
Envy 5 1
Contempt 0 0
Surprise 6 1 Attributional theory
TABLE 2
The Domain of Academic Emotions: Examples
Positive Negative
Task-related and self-related
Process Enjoyment Boredom
Prospective Anticipatory joy Hopelessness
Hope Anxiety
Retrospective Joy about success Sadness
Satisfaction Disappointment
Pride Shame and guilt
Relief
Social
Gratitude Anger
Empathy Jealousy and envy
Admiration Contempt
Sympathy and love Antipathy and hate
emotions experienced by students (the Academic Emotions
Questionnaire [AEQ]).
3. In a number of quantitative studies using the AEQ, we
tested assumptions underlying Pekrun’s (1992b) cogni
-
tive-motivational model concerning the effects of emotions
on students’ self-regulated learning and achievement.
4. Finally, we developed and tested a social cognitive,
control–value theory on theindividual andsocial antecedents
of academic emotions.
In conclusion, we emphasize the reciprocal nature of link-
ages between academic emotions, their effects, and their an-
tecedents. Also, we discuss the need for strengthening the
studyofemotionsin educational psychology, and thevalue of
using multiple research strategies to analyze this domain of
human affect.
EXPLORING THE OCCURRENCE AND
STRUCTURES OF ACADEMIC
EMOTIONS: QUALITATIVE CASE
STUDIES
How do students feel at school and university, and what are
the contents, sources, and functions of the emotions they ex-
perience? In a series of five exploratory studies, we analyzed
the emotions of school and university students by means of
qualitative interviews and questionnaires. We asked our par-
ticipants about their emotional experiences in academic set-
tings when in class, studying, and taking tests and exams.
Also,we asked aboutthe quality andthe elementsof these ex-
periences and about their origins and consequences. In the
firststudy,students were asked torecall typical academic epi-
sodes from their autobiographical memories and to report the
emotionsexperienced within theseepisodes (Pekrun, 1992c).
The other fourstudies used amore situated approachby inter-
viewing participants immediately after classroom instruc-
tion, an exam, or a daily period of studying (cf. Spangler,
Pekrun, Kramer, & Hofmann,in press;Titz, 2001). As part of
each study, we presented a fixed sequence of questions, but
answers were less constrained, the focus being on individual
descriptions of emotional episodes. In addition, in two of
these studies assessing emotions during exams, video recon-
struction procedures were used and physiological activation
recordings were made (Spangler et al., in press).
Frequencies of Different Emotions
In accordance with theoretical expectations, the results of the
five studies showed that students experience a wide range of
emotions in academic settings. There was virtually no major
human emotion not reported by our participants, disgust be
-
ing the only notable exception. Emotional diversity implies
thattheory-drivenapproaches to students’ emotionsthat limit
the range of emotions considered for theoretical reasons may
be in danger of missing important parts of students’ affective
life. For example, this may apply to test anxiety research and
to traditional conceptions of hope, fear, pride, and shame as
relating to achievement motivation.
Different categories of discrete emotions appeared with
differing frequencies, depending in part on the type of aca-
demic situation addressed. Overall, anxiety was the one emo-
tion reported most often, accounting for 15% to 25% of all
emotions reported in our studies. Anxiety was mentioned
most often not only in relation to taking exams, but also with
reference to being in class or studying at home. Achievement
pressure and expectancies of failure were reported as major
contributors to emotional arousal,thus pointing to the need to
enhance our students’ psychological well-being by giving
them chances to cope with excessive demands, and by in-
creasing opportunities for success. However, despite the fre-
quency of reports about anxiety, positive emotions were
described about as often as negative emotions. Aside from
anxiety, emotions reported most often were enjoyment of
learning, hope, pride, and relief, as well as anger, boredom,
and shame. In addition, there were accounts of several less
frequently reported emotions (e.g., the social emotions of
gratitude, admiration, contempt, and envy). Hopelessness
was also reported less often, but some accounts of this emo-
tion involved reports about personal tragedies including sui-
cidal ideation relating to failing academic exams.
Tracing Unexpected Phenomena
The qualitative accounts given by our participants also en-
abledus to detect phenomenathat otherwisewould havegone
unnoticed. A case in point was students’ meta-emotions, that
is, their feelings about their own emotions. For example, a
number of students gave detailed accounts of experiencing
angeraboutbeinganxiousbeforeexams.Insomeofthesestu-
dents, this anger helped them to find ways to cope with the
anxiety, thus implying that meta-emotions may facilitate stu-
dents’ coping with negative emotions, something educators
may wish to consider when trying toassist studentsin dealing
withtheiraffectiveexperiences.An example pertaining to the
sources of academic emotions is our participants’ reports
about academic boredom. We had expected that boredom
should occur when demands are too low, as in the case of
high-ability students who are taught in regular classes. Many
of our participants’ reports about boredom conformed to this
analysis: When self-evaluations of abilities were high and in-
structional demands low, students reported that they felt
bored and daydreamed. However, a sizable number of other
students told us that they had experienced boredom when
feeling unable to keep up with demands, implying that bore
-
dom was connected to low self-evaluations of abilities and
high evaluations of demands. Apparently, boredom may
servefunctions of escaping behaviorallyor mentallyfrom sit
-
93
uations not providing sufficient stimulation, but may also
serve as an escape from situations having demands that are
beyond the student’s capabilities.
Constructing Taxonomies of the Internal
Structures of Academic Emotions
From a moresystematic perspective,we usedour exploratory
findingsas an empirical basisfor constructingtheoretical tax-
onomies of the internal structures of different academic emo-
tions. Our participants reported a broad range of affective,
cognitive, physiological, and motivational elements of their
emotional experiences that were used for creating these tax-
onomies. For example, concerning cognitive components,
three major categories of cognitions that were experienced as
being part ofacademic emotions werethoughts about thetask
(e.g., the quantity, difficulty, and relevance of tasks),
thoughts about mastery and achievement, and thoughts about
the social situation within an academic setting.
Many of these cognitive components were reported as re-
lating to more than one emotion. In a similar vein, some other
types of components were also shared by several emotions
(e.g., facets of perceived physiological arousal). This would
suggest some overlap among different emotions, as in the
case of worries about failure that were reported as being part
of exam-relatedanxiety but also as components of shame and
hopelessness. Rather than being characterized by specific,
exclusive elements, it seems that different emotions show
specific profiles of components, some of them being shared
with other emotions.One generalimplication is that it maybe
difficult to measure different emotions in nonoverlapping
ways, making it difficult to ensure discriminant validity of
these measures. In this regard, contemporary test anxiety
scales containing worry items (cf. Hodapp & Benson, 1997;
Sarason, 1984) may measure shame or hopelessness in addi-
tion to anxiety, thus making it necessary to reinterpret much
of the empirical evidence gathered using them.
Academic Emotions and
Physiological Activation
Finally, we also used our qualitative data to analyze relations
between academic emotions and stress-related physiological
processes. It is interesting that measures of emotional inten-
sity based on our participants’ qualitative reports about their
emotions experienced during an important exam were sys-
tematically related to cortisol levels before and after the
exam, whereas questionnaire-based measures showed a
much less consistent pattern. Specifically, anxiety during an
exam was positively related to postexam cortisol levels indi
-
cating physiological activation during the exam, whereas
problem-focused as well as emotion-oriented coping during
the exam were negatively related to cortisol levels, implying
that coping had reduced physiological activation. This evi
-
dence suggests that qualitative reports may prove useful even
for studying functional relations between subjective and
physiologicallevels of emotions, a fieldthathasbeenplagued
by inconsistent findings to date (Spangler et al., in press).
In sum, findings of our qualitative studies demonstrated
that students experience a rich and intense emotional life in
academic settings, suggesting that reductionist conceptions
of students’ emotions may fall short of adequately covering
this domain. Although anxiety was the single discrete emo-
tion reported most often, positive emotions were reported no
less often than negativeemotions, thuspointing to the need to
investigate them more thoroughly (cf. also Fredrickson,
2001). Furthermore, our findings allowed us to trace a num-
ber of unexpected phenomena as in the case of meta-emo-
tions, and to build empirically based conceptual taxonomies
of the internal structures of academic emotions.
QUANTITATIVE MEASUREMENT:
THE AEQ
Whereasqualitative methods maybe ideally suitedto explore
new fields, quantitative measures are needed for more rigor-
ous tests of hypotheses. From its start in the 1950s, test anxi-
ety research recognized the need for quantitative measure-
ment applications. The progress of this field was tightly
connected to advances in test anxiety measurement. Today,
there are dozens of scales assessing this single emotion (cf.
Hodapp & Benson, 1997; Zeidner, 1998). Concerning aca-
demic emotions other than test anxiety, however, explicit
measures are largely lacking as of yet. One exception is the
anger measure presented by Smith, Furlong, Bates, and
Laughlin (1998).
Previous research on academic emotions beyond anxiety
used general measures of emotions not tailored to the aca-
demic domainor sets of single rating scales assessing various
emotions following success and failure (cf. Weiner, 1985).
Measures of general emotions, however, are removed from
the academic domain and probably less predictive for aca-
demic achievement than more domain-specific scales. For
example, general trait anxiety correlates less with students’
achievement than their test anxiety (Hembree, 1988). Rating
scales based on single items, on the other hand, may be eco-
nomic and useful under specific experimental circumstances
but may lack reliability and cannot adequately represent the
detailed internal structures of academic emotions.
We therefore attempted to develop a multidimensional
instrument that measures a number of more important aca-
demic emotions in domain-specific ways, using multiple
items for each emotion to represent different components of
these emotions. In doing so, we adopted the standards that
have been set by recent advances in test anxiety measure
-
ment (Hodapp & Benson, 1997).
94
Theoretical Considerations Guiding Scale
Development
A number of conceptual decisions had to be made at the out
-
set. First, we needed a working definition of the concept of
emotion. Inline with contemporary component process mod-
els (e.g., Scherer, 1984), we view emotions as involving sets
of interrelated psychological processes. Affective, cognitive,
physiological,andmotivationalcomponentprocessesmaybe
the most important (e.g., feeling tense and uneasy, worrying,
being activated physiologically, and wanting to escape in
anxiety). Such a conception goes beyond traditional models
of test anxiety by taking motivational components into ac-
count as well. Motivational components are not represented
by recent test anxiety scales, although they have been in-
cluded in items of early instruments such as the Test Anxiety
Questionnaire (Mandler & Sarason, 1952). Concerning the
contents of the components of different emotions, we made
use of the evidence-based taxonomies derived from our ex-
ploratory data.
Second, we decided to develop scales for a limited set of
emotions that may be important in many academic situations
and that can be assumed to affect students’ learning, achieve-
ment, and health. Specifically, we used three criteria in creat-
ing this set of emotions: (a) The scales should represent those
categories of primary human emotions that play a role in aca-
demic settings—According to our exploratory data, this is
true of joy, anger, anxiety, and shame, as well as hopeless-
ness, which may be regarded as an anticipatory version of the
primaryemotion of sadness; (b)we wantedto takethose eight
emotions into account, which were reported frequently by
students in our exploratory studies (see the previous section);
and finally, (c) with regard to emotion effects, we wanted to
include positive and negative emotions, as well as both acti-
vating and deactivating emotions within these categories be-
cause the two dimensions of valence and activation may be
regarded as basic determinants of many effects of emotions
(see the next section). The resulting set of emotions contains
the positive emotions of enjoyment, hope, and pride (positive
activating emotions), as well as relief (positive deactivating),
and the negative emotions of anger, anxiety, and shame (neg-
ative activating), as well as hopelessness andboredom (nega-
tive deactivating; see Table 3).
Third, clarification concerning the situational and tempo-
ral generality of the emotion constructs to be measured was
necessary. Regarding academic situations, being in class,
studying outside of class, and taking tests and exams are the
three most important types of academic situations at school
anduniversity.Thesesituationsarecharacterizedby different
functions and social structures, implying that emotions relat-
ing to them may differ as well. For example, enjoyment of
classroom instruction may be quite different from enjoying
the challenge of an exam. We therefore decided to develop
separate scales pertaining to class-related, learning-related,
and test-related emotions. Concerning temporal generality,
we wanted to construct items so that they can be administered
using different instructions, thus making it possible to mea-
sure trait academic emotions (e.g., trait test anxiety),
course-related emotions pertaining to a specific course and
class, as well as state academic emotions experienced within
single episodes of academic life (e.g., statetest anxiety). On a
bipolar conceptual continuum representing emotional traits
versus states, trait academic emotions would be located at the
trait end of the continuum, state academic emotions at the
state end, and course-related emotions in between.
Construction of Scales
Item construction was based on the student reports gained in
our exploratory studies, on theoretical considerations, and,
concerning test anxiety, on Sarason’s (1984) Reac-
tions-to-Tests Questionnaire and Hodapp and Benson’s
(1997) Integrative Test Anxiety Questionnaire. From an ini-
tial item pool, itemswere selected for preliminary versions of
the scales by using expert judgment and criteria of redun-
dancy. Selection of items for the final versions was based on
item statistics of the preliminary versions and on results of
confirmatory factor analysis for each scale (cf. Titz, 2001).
Item Statistics, Reliabilities, and
Structures of Scales
Table 3 presents item numbers and internal reliabilities for
the trait versions of the AEQ scales. Reliabilities are based
on two samples of university students (learning-related and
class-related emotion scales: N = 230; test emotion scales:
N = 222). Coefficients imply that internal reliabilities are
quite satisfactory.
Confirmatory factor analysis showed that internal struc
-
tures of trait academic emotions can differ between emo
-
tions. Specifically, whereas items relating to affective,
95
TABLE 3
Academic Emotions Questionnaire: Reliability of the Trait Scales
Scales
Learning-Related
Emotions
Class-Related
Emotions Test Emotions
Emotion a Items a Items a Items
Enjoyment .90 14 .89 15 .90 23
Hope .86 9 .84 9 .89 16
Pride .84 9 .86 9 .92 16
Relief
a
— — — — .89 14
Anger .89 14 .85 11 .89 17
Anxiety .92 18 .89 13 .94 31
Hopelessness .93 13 .88 10 .94 21
Shame .90 14 .91 15 .93 19
Boredom
b
.93 17 .93 14 — —
a
Relief scale for test emotions only.
b
Boredom scale for learning-related
and class-related emotions only.
