Affect, Epistemic Emotions, Metacognition, and Self-Regulated Learning

The concept of self-regulated learning (SRL), when first introduced in educational research in the 1980s, brought a new perspective to theories of learning (e.g., Zimmerman & Schunk, 1989). The emphasis in cognitive theories of learning was on acquisition of content knowledge (e.g., conceptual learning) and the constraints on it—developmental or cognitive. SRL theories changed the emphasis from content knowledge to goal-driven and strategic regulation of knowledge. Motivation and goal setting are critical components of SRL, and so are the metacognitive monitoring and control of learning processes. The person is an active agent in his or her learning and responsible for the outcomes of it. Τhe person decides whether to initiate learning activities or not, direct his/her thinking, carry out the needed cognitive processing, monitor and control cognitive processing through the use of learning, cognitive, or metacognitive strategies, and evaluate and reflect on the outcome of cognitive processing (e.g., Schunk & Zimmerman, 1994). The evaluative phase is critical for subsequent cycles of learning because it offers the ground for a personal account of the factors that might have impacted the current learning outcomes.

The SRL conception of learning affect (in the sense of learning-related affective responses or emotions) plays a role particularly at the beginning of a learning event (e.g., interest or anxiety) or at the evaluation of the learning outcomes phase (e.g., pride or shame). The role of emotions is often reduced to their interactions with motivation such as achievement goals, and attributions about learning outcomes (Pekrun, 2006; Weiner, 2014). This makes sense because SRL is conceived as a top-down process, and emotions are presumably implicated in goal setting and evaluation of the outcomes of a SRL cycle that sets the scene for next ones. Pekrun (2006) extended the range of emotions implicated in learning to include emotions during learning as well (e.g., when studying or taking exams). Emotions during learning are related to final performance (Pekrun, Elliot, & Maier, 2009), but it is not clear if and how they are implicated in the regulation of learning besides their effect on motivation.

However, take the case of a student who made an error as he worked on a problem. Monitoring of errors is a metacognitive process. But is awareness of an error a “cold” metacognitive event (purely informational experience) or a subjective experience that encompasses an evaluative component as well? Making an error has implications for the self and our goals, and hence is appraised as such. This entails that metacognitive awareness of an error is “hot” metacognition that includes both an informational and an affective component (Aarts, De Houwer, & Pourtois, 2013). When we become aware of having committed an error, we experience negative feelings, both negative affect and distinct emotions (e.g., embarrassment, confusion, anxiety), and this may influence our control decision to quit, ignore, or continue processing, ask for help, increase effort, use control strategies, and so on.

Respectively, a student finishes a test and is happy because she responded correctly. Knowing that you did well on a test is a metacognitive process that conveys subjective feedback about one’s response (in the absence of external feedback). This information is in the form of a feeling of confidence about one’s response. In our example, the student is happy about her performance based on the metacognitive feeling of confidence. These examples suggest not only that there are interrelations between metacognition and affect but also that subjective experiences in the form of metacognitive feelings or affect and emotions provide information that is critical for bottom-up regulation of learning (e.g., Nussinson & Koriat, 2008). Hence, affect plays a major regulatory role in SRL, and this is done in interaction with metacognition.

The aim of this chapter is to lighten the role of emotions and affect in the regulation of learning and the interactions of affect with metacognition. This will hopefully contribute to a revision of current conceptualizations of SRL in the direction of a more substantive role of affect in SRL. The claim being made here is that affect in SRL can have effects independent of cognition or metacognition but also in interaction with them. Moreover, although the main models of emotions in learning (e.g., attribution theory, Weiner, 1985, 2014; control-value theory of emotions, Pekrun, 2006) emphasize achievement emotions, I will discuss the interactions of metacognition with two epistemic emotions, namely surprise and curiosity, that arise in response to cognitive states such as discrepancies, inconsistencies, or interruptions of processing. Epistemic emotions play a major role in the change of cognitive processing mode (e.g., from automatic to analytic) or search for new information that can resolve a gap or uncertainty in our knowledge.

In what follows, I will first give a brief description of the metacognitive and affective model of SRL (MASRL; Efklides, 2011), which provides a theoretical framework that envisages interrelations between metacognition and affect. Then I will give a brief overview of the relations of affect and emotions with cognition and metacognition in SRL. Then I will present evidence on interrelations between metacognition and epistemic emotions, and discuss the implications of such evidence for the mechanism underlying SRL.

DEFINITIONAL ISSUES

Before presenting the MASRL model, and for conceptual clarity, I will briefly provide the definitions of affect and metacognition. Affect refers to subjective experiential states that have a pleasant/unpleasant valence. It is a generic term that encompasses emotions, feelings, mood, self-esteem, attitudes, and so on (Forgas, 1994). Specifically, emotions take discrete forms in terms of subjective experience and autonomic reactivity (e.g., fear, anger, joy), are short lived, and are triggered by stimuli or events that are relevant to one’s goals or concerns. They can have behavioral manifestations (e.g., facial expressions or bodily posture) and action tendencies (i.e., approach-avoidance), and they often involve appraisals about the triggering stimuli/events (Frijda, 1986; Pekrun, 2006). Feelings are the experiential part of emotions or subjective states that inform about the good functioning of the organism or processing (Frijda, 1986). Moods, on the other hand, are what is left after an emotion has ceased to exist and last longer than emotions (Forgas, 1994; Frijda, 1986).

Metacognition is cognition about cognition (Nelson, 1996), or the monitoring and control of cognition (Flavell, 1979). It takes the form of metacognitive experiences, metacognitive knowledge, and metacognitive skills (Efklides, 2008; Flavell, 1979). Metacognitive experiences comprise subjective experiences such as metacognitive feelings (e.g., feeling of difficulty or tip-of-the-tongue states), metacognitive judgments such as judgment of learning (JOL), and task-specific knowledge, that is, what the person is heeding as one is working on a task. Metacognitive knowledge is declarative knowledge or beliefs about persons (including the self), tasks, strategies, and goals, and metacognitive skills are procedural knowledge or strategies people use to control cognitive processing, such as orienting, planning, monitoring, correcting, and evaluating.

THE MASRL MODEL

The MASRL model (Efklides, 2011) stresses the interactions of metacognition with affect. It posits that the task and its context is a critical component of the SRL process. However, the task has objective characteristics, but their representation may vary across students depending on their prior knowledge or skills. The processes involved in the representation of the task and its demands function at two levels of generality: the Person level and the Task X Person level.

(a) At the Person level, the representation of task structure and processing demands is done based on the person’s capabilities or prior domain-specific knowledge, but also on metacognitive knowledge, such as beliefs about which tasks are easy or difficult (Efklides & Vlachopoulos, 2012), motivational orientations or expectancy-value beliefs, agency beliefs, and affective components such as well-developed interests, self-concept about one’s competence in various knowledge domains, and attitudes toward various knowledge domains. For example, students with low self-efficacy in mathematics avoid analyzing the task structure and processing features even of easy problems because they believe mathematics is difficult and they do not understand it (e.g., Usher, 2009).

(b) Task representation at the Task X Person level involves a closer analysis of actual task features, namely, analysis of task structure and processing demands (e.g., whether it requires memory retrieval processes). At the same time, the person, through metacognitive experiences, becomes aware of the state of cognitive processing, that is, whether it is fluent or there are interruptions, discrepancies, gaps, conflicts, or progress. Metacognitive monitoring presumably, but not exclusively, informs the need for control of processing (Efklides, 2014), the use of cognitive and metacognitive strategies, as well as effort exertion.

At the Task X Person level, besides the metacognitive loop that regulates cognitive processing, there is a different loop encompassing affect and its regulation. The decision on control processes during cognitive processing is based on metacognitive experiences and/or affect. For example, a student who does not like mathematics experiences negative affect when given a math problem; hence, when coming across a difficulty, he may decide to quit processing without effort exertion (as a response to the negative affect that accentuates the experienced difficulty). If, despite the negative affect, the difficulty experienced is judged manageable (as informed by the metacognitive loop) or worth trying to resolve it (affective loop), then control is guided jointly by metacognitive experiences and affect. As a consequence, the student exerts effort and uses cognitive or metacognitive strategies for the control of processing.

Let us take an example: Mary is a student who is good at mathematics and likes working on math tasks. One day in the classroom, she is given a challenging math task. Although at first glance, she thinks this is a difficult task (based on her metacognitive knowledge), her positive attitude toward math makes her curious about the way the problem can be solved and willing to engage in problem solving. This is a decision at the Person level. Following this decision, she revisits the problem and starts its processing (Task X Person level). As she is working on the problem, she experiences difficulty. Her judgment of learning is that the probability of solving the problem is low. Her initial positive feelings decrease, and she realizes that she needs to revise the initial approach to the problem. As she tries various ways to tackle the problem, she becomes aware that the original representation of the problem was incorrect, and there is another representation that can lead to the solution. She feels a pleasant surprise, relief, and renewed interest. She starts working with the novel task representation and solves the problem. While carrying out the solution, she feels that processing is running fluently. She also feels confident that the solution is correct and satisfied with her response. When reflecting on the whole learning experience, she feels proud that she overcame the difficulty and solved the problem.

This example suggests that the student started with goal setting in a top-down mode of regulation of learning based on her previous domain-specific skills and metacognitive knowledge and the affective response to the problem. As the metacognitive experiences and affect changed during problem solving, she became aware of the lack of progress in processing (monitoring), and this led to a decision to change the problem representation (control). However, it was the pleasant surprise and affect that reinforced the decision to start a new SRL cycle because it denoted that the previous negative metacognitive and affective experiences were not valid anymore. Therefore, awareness of one’s thinking, monitoring of the fluency of processing, and evaluation of the progress made as conveyed by metacognitive experiences and affect during problem solving, that is, “hot” metacognition, led to bottom-up regulation of problem solving until the goal was achieved.

AFFECT, COGNITION, AND METACOGNITION

Since the 1980s, there has been a lot of research on the interrelations between cognition and affect (e.g., Bower, 1981), showing that affect activates memory information, and memory information activates affective responses. Furthermore, positive affect is associated with holistic, more creative, and less critical thinking, whereas negative affect is associated with more analytic processing (Alter, Oppenheimer, & Epley, 2013). Affect is also at the heart of decision making, intuition, and evaluative judgments (Schwarz, 2005; Schwarz & Clore, 1983). All of the above constitute what is called “hot cognition” (Thagard, 2006).

There is also a relationship between emotions and cognition. Curiosity, surprise, and confusion are epistemic emotions in the sense that they focus on knowledge states (Muis, Psaradellis, Lajoie, Di Leo, & Chevrier, 2015); their role is to turn attention to information that is most relevant to current cognitive processing. D’Mello (2013) showed that epistemic emotions are present during learning along with achievement emotions.

In addition to the relations between affect and cognition, there are also relations between affect and metacognition that have implications for SRL. As Efklides (2016) showed in a literature review, there is evidence suggesting that metacognition, in the form of metacognitive experiences, interacts with affect. Specifically, metacognition can have effects on affect, affect can have effects on metacognition, and both of them may be triggered by knowledge states such as conflict, interruption of processing, or discrepant events.

Specifically, metacognitive experiences impact appraisals of learning-related or achievement emotions (Pekrun, 2006; Weiner, 1985, 2014). For example, appraisals of task value or controllability during cognitive processing presuppose subjective evaluations of task difficulty, effort to be exerted or already exerted, and expectancies about the learnability of the task material or processing outcomes. Such subjective evaluations take the form of metacognitive experiences such as feeling of difficulty, feelings of effort, judgments of learning (JOLs), or confidence. One could argue that feelings of effort provide information on cost (a form of value appraisal; Eccles & Wigfield, 2002), whereas feeling of difficulty or JOLs inform about the controllability of the situation through the expectancies they create. Confidence, on the other hand, provides subjective feedback on the quality of processing outcome, namely success or failure (see Tornare, Czajkowski, & Pons, 2015, who showed that outcome-related emotions were predicted by self-concept and performance, but the effects were mediated by metacognitive experiences such as feeling of difficulty and feeling of success). Metallidou and Efklides (2001) showed that feeling of difficulty, along with estimate of effort or time spent on the task, predicted attributions of task difficulty, whereas feeling of confidence predicted attributions of competence. Moreover, reattributing feeling of difficulty to task difficulty rather than lack of competence changes the emotions experienced and subsequent SRL (Autin & Croizet, 2012).

Conversely, affect can have effects on metacognitive feelings, such as feeling of difficulty, and metacognitive judgments such as JOLs. For example, Efklides and Petkaki (2005) showed that negative mood increased the self-reported feeling of difficulty in math problem solving, and Koriat and Nussinson (2009) showed that induced feeling of difficulty through the contraction of the corrugator muscle (i.e., frowning), which is indicative of effort exertion, impacted subsequent JOLs.

Furthermore, there are relations between metacognition and core affectivity, that is, positive and negative affect. Such relations originate from cognitive states such as fluency in processing and response formation. Fluency triggers positive affect and disfluency negative (Winkielman & Cacioppo, 2001). It also impacts metacognitive experiences such as feeling of difficulty and feelings of effort exertion (Efklides, 2016; Efklides, Schwartz, & Brown, in press).

Similarly, for SRL, it is important to monitor the rate of progress toward our goal (Carver, 2015; Carver & Scheier, 1998). If it is as expected, no specific affect is present. If the rate of progress toward our goal is faster than anticipated, we experience positive affect; if the rate is slower than expected, negative affect is experienced. This informs us that more effort is needed to attain our goal. From a metacognitive point of view, feeling of difficulty or ease of processing along with monitoring of effort and time spent on a task provide information about the velocity with which we approach our goal. Therefore, affect and metacognition jointly determine engagement with/disengagement from our goal.

To sum up, affect is an essential component of SRL, and its effects are not limited to achievement emotions. Moreover, affect exerts its effects on SRL in interaction with cognition and metacognition. This is particularly true for epistemic emotions, as shown in the following section.

EPISTEMIC EMOTIONS

Epistemic emotions offer awareness of knowledge states such as conflict, incongruence, discrepancy, interruption of processing, or gaps in one’s knowledge (Muis et al., 2015). Their role is to facilitate action that can help restore processing. This can be done by focusing attention to particular aspects of a situation (e.g., discrepant information), seeking of new information, or construction of new schemas. Specifically, curiosity denotes a gap in one’s knowledge (Litman, 2010). It activates behaviors such as exploration that can fill in the gap. Surprise, on the other hand, denotes the presence of unexpected and discrepant events (Mandler, 1975; Meyer, Reisenzein, & Schützwohl, 1997); its role is to refocus attention on the discrepant information and activate analytic processing (Topolinski & Strack, 2015). Confusion is experienced when there is conflict of response tendencies, and therefore no response can be formed. It prompts behaviors that can reduce confusion by resolving the conflict (for the constructive role of confusion in learning, see Muis et al., 2015).

From an SRL point of view, both epistemic emotions and metacognition have as their object knowledge states and cognitive processing. Therefore, there should be relations between them. Two studies that investigated the association of metacognitive experiences with surprise and curiosity, respectively, are presented in the following sections.

SURPRISE

According to Mandler (1975, 1984), surprise is associated with cognitive interruption. Cognitive interruption occurs unexpectedly and whenever the available cognitive schemas fail to handle task demands. Interruptions trigger surprise when something unexpected, mismatching, or discrepant to our schemas occurred. Depending on the implications of the discrepant event, be they positive or negative, surprise is pleasant or unpleasant. At the cognitive level, interruptions due to discrepant events activate schema revision processes; this is an effortful process. Hence, at the metacognitive level, interruptions should be associated with feeling of difficulty and increased effort. This entails that feeling of difficulty due to cognitive interruption will be correlated with surprise. Touroutoglou and Efklides (2010) tested this hypothesis. The predictions were:

1.

Self-reported surprise and feeling of difficulty will be increased in interrupted, as compared with noninterrupted, cognitive processing.

2.

Self-reported surprise will correlate with self-reported feeling of difficulty.

3.

Self-reported surprise will decrease as stimuli with interruptions are repeated across trials whereas feeling of difficulty will remain unchanged across trials.

Method

Ten psychology students of both genders were individually tested on a computerized task that involved three blocks of trials, each comprising three types of number sequences. Each of the number sequences followed a simple arithmetic rule. Participants had to fill in the number that followed the last number of the sequence. The three rule types were:

(a) No interruption in the processing of the number sequence. A schema was established with the first two numbers (based on a rule), and the sequence continued with the same rule:
e.g., 2 4 6 8 10 12 ---
(b) Interruption due to repetition of a number. A schema with a simple rule was initially established, but then a different number than expected followed:

e.g., 2 4 4 6 6 8 ---
(c) Interruption due to an intervening number:
e.g., 2 13 4 13 6 13 ---

Results

The results showed that there were no differences in performance in the three types of number sequences. However, reaction time was higher in the interrupted sequences compared with the uninterrupted. As predicted, feeling of difficulty and surprise were higher in the interrupted sequences compared with the uninterrupted. Also, there was high correlation (about .90) between feeling of difficulty and surprise, but, whereas surprise decreased from the first to the following blocks of trials, feeling of difficulty remained at similar levels across blocks.

The high correlation between feeling of difficulty and surprise could mean that the two measures tapped the same underlying mechanism, and, in essence, feeling of difficulty was similar to surprise. We carried out two other studies with number sequences that involved the four arithmetic operations (e.g., 2, 6, 18, 54, 162, 486, ---; Touroutoglou & Efklides, 2010). The results showed that whereas feeling of difficulty increased in sequences involving multiplication and division, as compared with addition and subtraction, surprise did not differ between the four arithmetic operations sequences. Moreover, there was no correlation between feeling of difficulty and surprise. The second study tested the effect of working memory load on feeling of difficulty and surprise. We used arithmetic operations with three-digit numbers, but the size of the numbers differed. Half the sequences involved numbers ranging from 300 to 400 (e.g., 421, 418, 415, 412, 409, 406, ---) and the other half involved numbers ranging from 700 to 900 (e.g., 921, 918, 915, 912, 909, 906, ---). The results showed an effect of number size on feeling of difficulty but not on surprise. Therefore, sheer processing difficulty or working memory load increases the experienced difficulty but not surprise when there is nothing unexpected or discrepant from the activated schemas. On the contrary, in interruptions of processing due to discrepant events, there is surprise as well as increased effort to restore the interrupted processing.

Finally, in the study with interruptions of processing, the correlations between feeling of difficulty and surprise were positive, suggesting that surprise was associated with negative affect, as feeling of difficulty is. Touroutoglou (2009), however, showed that the correlation between feeling of difficulty and surprise can also be negative, depending on whether the resolution of the interrupted number sequence is based on a difficult or easy rule. Specifically, she presented a number sequence that followed a complex rule. Then she showed that the same number sequence could be produced with the application of a very simple rule. This increased surprise but not feeling of difficulty, suggesting a negative relation between the two unlike the positive in the initial study.

Discussion

Feeling of difficulty and surprise are triggered by different mechanisms, but in the case of cognitive interruption, the two are related. Topolinski and Strack (2015), using physiological measures of facial expressions, found that surprise is associated with cognitive interruption and increased effort/difficulty that elicits negative affect. Negative affect in turn initiates cognitive tuning by switching current cognitive processing from automatic and heuristic to systematic and more analytic. Thus, surprise and feeling of difficulty are a good example of the synergy between metacognitive feelings and affective responses as one works on a task. However, surprise can also be pleasant (Touroutoglou, 2009). It is plausible that phasic change of affect from negative to positive signals that analytic processing or mental effort can be relaxed.

To sum up, for successful SRL, both surprise and metacognitive experiences are important, but their regulatory effects seem to be mediated by negative affect. To decipher these interactions, more research is needed.

Curiosity

Curiosity is “a desire for acquiring new knowledge and new sensory experience that motivates exploratory behavior” (Litman & Spielberger, 2003, p. 75; see also Berlyne, 1954). Curiosity arises when there is an inconsistency or gap in knowledge we have and knowledge we would like to have (Loewenstein, 1994). Essentially, what underlies curiosity is the expectation that a piece of information can resolve uncertainty in our knowledge states. This explains why curiosity is most obvious when one is in tip-of-the-tongue state (TOT) (Litman et al., 2005), and less so in the “not-know” state, in which one is aware of a gap in knowledge (absolute gap) that has little probability to be resolved (Litman, Hutchins, & Russon, 2005; Lowenstein, 1994).

From the MASRL model’s point of view, curiosity as exploratory behavior is manifested at the Task X Person level (state curiosity). However, curiosity can also be measured as person characteristic, that is, as trait. There is perceptual (PC) and epistemic curiosity (EC) (Berlyne, 1954; Litman & Spielberger, 2003). Perceptual curiosity captures seeking of new perceptual information, as when we visit new places. Epistemic curiosity is desire for knowledge that motivates an individual to seek new ideas, eliminate gaps in knowledge, and resolve intellectual problems (Litman, 2008; Mussel, 2010). EC can be associated with the pleasure of discovering new ideas (interest-type curiosity) or with an effort to reduce negative affect arising from uncertainty about, or gaps in, one’s knowledge (deprivation-type curiosity; Litman, 2010).

Sideropoulou (2015)¹ investigated whether epistemic and perceptual curiosity as trait are related to state curiosity and whether state curiosity is associated with metacognitive experiences and positive or negative affect during problem solving. She used mathematical tasks (i.e., number sequences). At the Person level, she studied the interrelations between epistemic and perceptual curiosity as traits with mathematics self-concept, and attitude toward mathematics (positive, negative). The inclusion of self-concept and attitudes toward mathematics in the design was guided by previous research showing relations between these two person characteristics and metacognitive experiences (Dina & Efklides, 2009; Efklides & Tsiora, 2002). The assumption was that self-concept and attitude toward mathematics would not be related to curiosity as trait.

At the Task X Person level, state curiosity was measured as seeking feedback on the correct answer to the items of the number sequence tasks. There were also measures of performance, feeling of difficulty, and feeling of confidence for each item. Measures of positive and negative affect were collected at the beginning, during, or after the processing of the number sequence tasks. Measures of affect were included because interest-type curiosity is associated with positive affect, whereas deprivation-type curiosity is associated with negative affect.

The hypotheses were:

1. Epistemic curiosity, but not perceptual curiosity, will be positively related to state curiosity. Neither epistemic nor perceptual curiosity will be related to performance on the number sequences tasks, affect, or metacognitive experiences.

2. Self-concept will not be related to state curiosity. On the contrary, positive attitude toward mathematics will be positively related to state curiosity because it taps interest in mathematics; negative attitude will be negatively related to state curiosity.

3. State curiosity will be negatively related to feeling of confidence but not necessarily to feeling of difficulty. It will be positively related to positive affect.

Method

Testing took place in two phases: In the first phase, 283 university students of psychology responded to questionnaires tapping epistemic and perceptual curiosity as trait, mathematics self-concept, and attitude toward mathematics (positive, negative). In the second phase, 100 students (80 of whom came from the initial pool) responded to a computerized number sequence task.

Questionnaires

Epistemic and perceptual curiosity. The Litman and Spielberger (2003) questionnaire of trait curiosity (epistemic and perceptual) was used. The factor analysis of the Greek version revealed 13 items loading the epistemic curiosity factor (Cronbach’s α = .84) and 4 items loading perceptual curiosity (Cronbach’s α = .63). Example items: Epistemic curiosity: “It is interesting to think of controversial issues”; “I am interested to know how different people would react to a crisis situation.” Perceptual curiosity: “I like to discover new places to visit.”

Self-concept in mathematics (Dermitzaki & Efklides, 2000). The questionnaire was adapted for mathematics. It comprises 22 items tapping self-perception, self-efficacy, self-esteem, and others’ perception of one’s self. One factor was abstracted (Cronbach’s α = .95). Example item: “I am pleased with my abilities in math.”

Attitude toward mathematics (Aiken, 1996). It comprises 20 items tapping positive and negative attitude toward mathematics. Cronbach’s α for positive attitude was .97 and for negative attitude was .91. Example items: Positive attitude: “Mathematics is fascinating and amusing”; negative attitude: “I always feel extreme pressure in the mathematics lessons.”

Relations between person characteristics. Correlations between the factors of the questionnaires showed that epistemic and perceptual curiosity were moderately correlated, r = .34, p < .006. Perceptual curiosity was negatively related to positive attitude toward mathematics, r = -.33, p < .006. Self-concept did not correlate with any measure of trait curiosity. It only correlated with attitude toward mathematics: positive attitude, r = .81, p < .001; negative attitude, r = -.63, p < .001.

Tasks. There were 48 number sequence items in eight blocks of six items each. There were two types of number sequences: one following an arbitrary rule and one with arithmetic operations. There were also two levels of difficulty for each type of rule:

Arbitrary/easy: 1, 3, 1, 3, 1, 3, ---
Arbitrary/difficult: 1, 3, 1, 5, 1, 7, ---
Arithmetic/easy: 143, 140, 137, 134, 131, 128, ---
Arithmetic/difficult: 899, 896, 893, 890, 887, 884, ---

Affect was measured eight times: before starting the solution of the number sequence task, after each block of items, and at the end of the task. Responses were on a 4-point Likert scale. There were three items loading the positive affect factor (happy, aroused, bored [inversely coded]), and three items loading the negative affect factor (angry, disappointed, calm [inversely coded]).

Metacognitive experiences were measured before and after each item. Each item was first presented for 3 seconds, and a prospective judgment of feeling of difficulty was made, e.g., “How much difficulty do you think you will have to find the answer?” Then the item was presented again to be solved. Immediately after the solution, there was a retrospective measure of feeling of difficult, for example, “How much difficulty did you have to find the answer?” There was also a measure of feeling of confidence, after the response was given, for example, “How confident are you that your response is correct?” Responses were on a 4-point Likert-type scale, ranging from 1 (not at all) to 4 (very much).

State curiosity was measured after the response to the metacognitive experiences. The question was: “How much would you like to know the correct response?” Response was on a 4-point Likert-type scale. There was no feedback on the correct answer because this might interfere with the exploratory behavior.

Results

The correlations between person characteristics and state curiosity showed that both epistemic and perceptual curiosity were negatively related to state curiosity, r = -.32, p = .004, r = -.23, p = .05, respectively. This negative relationship was not expected and suggests that trait curiosity is not always conducive to state curiosity. Moreover, perceptual curiosity also correlated with state curiosity even though the task was purely cognitive. Interestingly, what supported state curiosity was the positive attitude toward mathematics. The respective correlation was positive, r = .28, p = .05. This could mean that a positive attitude toward a knowledge domain is a source of interest-type curiosity. Self-concept in mathematics and negative attitude toward mathematics were not related to state curiosity. Therefore, Hypotheses 1 and 2 were partly confirmed.

Hypothesis 3, however, was confirmed. Mean scores of performance, metacognitive experiences, and state curiosity are given in Table 1. Performance on easy arbitrary rules was higher than on easy arithmetic rules, and performance on difficult arithmetic rules was higher than performance on difficult arbitrary rules. However, performance was in general high in all cases except the difficult arbitrary rule items.

Table 1. Means (and Standard Error) of Performance, Metacognitive Experiences, and State Curiosity as a Function of Task Type and Level of Task Difficulty

State curiosity was lowest in the easy arithmetic rules and highest in the difficult arbitrary rule items. Feeling of difficulty (prospective and retrospective) and feeling of confidence were also highest in the difficult arbitrary rule items. Moreover, state curiosity was negatively related to feeling of confidence, r = -.24, p = .01, but there was no significant correlation with feeling of difficulty or performance. However, feeling of difficulty negatively correlated with feeling of confidence, rs ranging from -.22 to r = -.28, p = .01., and with performance, r = -.30, p = .01.

Most important, there was a positive relationship between positive affect and state curiosity, r = .42, p = .003, at the beginning of task processing (first block) that continued at following blocks, r = .26, p = .05, and r = .23, p = .05, at the fourth and the eighth block, respectively. Moreover, state curiosity was not related to negative affect. Therefore, state curiosity was associated with lower confidence as in deprivation-type curiosity, and with positive affect as in interest-type curiosity, as predicted by Hypothesis 3.

Discussion. The findings regarding curiosity are very interesting because they show that trait curiosity is distinct from other person characteristics, and state curiosity is distinct from metacognitive experiences. Trait curiosity, both epistemic and perceptual, was negatively related to state curiosity, contrary to Hypothesis 1. It seems that being curious about a particular knowledge domain can undermine state curiosity in another domain. Specifically, the epistemic curiosity measure used by Sideropoulou, although based on Litman and Spielberger’s (2003) pool of items, comprised mainly items tapping interpersonal relations and puzzling events that require an explanation. It was more a measure of interpersonal curiosity (Litman & Pezzo, 2007) than general epistemic curiosity. Obviously, interpersonal curiosity is not the kind of curiosity underlying the understanding of mathematical relations.

If, however, one likes mathematics (i.e., has a positive attitude toward mathematics), state curiosity is triggered. Of course, the nature of the task is also playing a role. State curiosity was highest in tasks that were novel, that is, number sequences that followed an arbitrary rule. Performance on such items was moderate, which suggests that curiosity is triggered in situations in which the response is not easy to find but not impossible. Finally, it seems that positive attitude is a potent factor for the triggering of state curiosity, not only as a precursor, but also as a resource for positive affect during task processing. When one likes a knowledge domain and enjoys dealing with tasks in this domain, this positive affect supports state curiosity even if the task is not particularly interesting by itself. This was the case for most of the tasks used by Sideropoulou, which involved simple arithmetic rules or simple arbitrary rules that were easily decoded and did not require any further information seeking. This kind of evidence lays the groundwork for the deciphering the conditions that activate state curiosity or prevent it.

EPISTEMIC EMOTIONS AND SRL—REVISITED

The aim of this chapter was to highlight the role of affect in SRL. Besides the unique effects of emotions on behavior and the interaction of emotions with motivation in the context of learning, I tried to show that affect also interacts with cognition and metacognition. The close relations of metacognitive feelings with affect render metacognitive experiences “hot” metacognition and suggest that the effects of monitoring on control in SRL may involve affective paths. One such path is epistemic emotions.

Surprise is an epistemic emotion that is not associated with some particular trait as curiosity is. For surprise, what is critical is the discrepancy between existing cognitive schemas and external events. Its role is to inform about the need for change of processing mode, from automatic to controlled. This is done, according to Topolinski and Strack (2015), through the interruption of processing and the unpleasant affect associated with the disfluency produced by interruption. However, whereas Topolinski and Strack stress the nonconscious effects of disfluency on affect and connect it to changes of processing mode, Touroutoglou and Efklides (2010) demonstrated that the person becomes aware of disfluency through feeling of difficulty; hence, change of processing mode can be a conscious metacognitive control decision. That is, negative affect triggered by disfluency is not an independent feeling state but part of the experience of feeling of difficulty. The combination of feeling of difficulty (disfluency), negative affect, and surprise (due to a discrepant event) informs the person about the source of interruption of processing and the need for control, such as change of processing mode.

As with surprise, the mechanism underlying the functioning of curiosity, although different, involves metacognitive feelings. The metacognitive feeling in this case is confidence in one’s response, TOT, or “don’t know” state. Another difference is that curiosity, unlike surprise, can function as trait. Trait curiosity is considered strength of character in positive psychology (e.g., Toner, Haslam, Robinson, & Paige, 2012) because it opens up possibilities and broadens one’s knowledge and perspectives. From an evolutionary point of view, this is highly adaptive. However, it is also true that curiosity can “kill the cat.” This can be done in various ways: For example, curiosity can be specific to a domain or diversive (Litman & Spielberger, 2003); being focused in a domain promotes interest-type state curiosity and exploratory behavior. This is positive. But limiting one’s epistemic or perceptual curiosity only to specific areas leads to neglect of other potentially important information in other domains. This is negative and unproductive for knowledge acquisition in other domains in one’s life. The negative correlation Sideropoulou (2015) found between epistemic and state curiosity argues in this direction.

A second way in which trait curiosity can be unproductive is through an unfocused, diversive kind of curiosity and exploratory behavior. This kind of curiosity is activated by boredom or a desire for variability in knowledge, irrespective of content. Processing a broad range of potentially interesting or important information can open up possibilities (e.g., set up new goals in the presence of boredom with current activities; Elpidorou, 2014) but can be a waste of energy and resources if curiosity is distributed over a variety of stimuli that are not relevant to one’s main concerns. In such a case, curiosity distracts attention and prevents deeper processing of useful information.

On the other hand, there is curiosity as feeling-of-deprivation (Litman, 2008; Litman, Crowson, & Kolinski, 2010). Deprivation-type curiosity is experienced as a subjective state of high arousal and a compelling urge not to stop looking for the nonavailable information until it is found. This is again adaptive because it keeps one on track for the attainment of one’s goal. However, it can be counterproductive if the information sought after cannot be reached or is not available, and one needs to disengage from the search of it. How, then, can people decide when they should keep up their efforts to find what is missing in their knowledge? I would claim that affect and metacognitive experiences are critical for such a regulation of one’s exploratory behavior.

Specifically, deprivation-type curiosity is associated with low confidence in one’s response or inability to access information that already exists in memory. However, if, at the same time, the person experiences positive affect (e.g., because up to that moment, there had been progress in cognitive processing or the student likes the task), positive affect becomes a cue that effort can be maintained and exploratory behavior can go on. If, however, negative affect prevails, then the meaning is that exploratory behavior does not pay off and should be suspended. Feeling of difficulty, on the other hand, informs disfluency of processing. It can become a cue for feeling of confidence, because disfluency entails that the probability of the response being correct is not high (see the negative correlation between the two in the Sideropoulou study). Thus, although feeling of difficulty was not directly related to state curiosity, lowering confidence and increasing awareness of “not knowing,” as well as negative affect, offers the groundwork for the regulation of state curiosity and the decision to persist with exploratory behavior or not.

Self-related information, on the other hand, offers the background against which the person judges the potential for a response being correct. Although in the Sideropoulou study, self-concept was not directly related to state curiosity, it is known that self-concept impacts metacognitive experiences, including feeling of confidence (Efklides & Tsiora, 2002). Hence, indirectly, self-concept can have an impact on state curiosity via feeling of confidence. Obviously, research on curiosity and its role in SRL needs to be extended.

However, what is important for state curiosity is positive attitude toward a knowledge domain and positive affect while working on a task. One is willing to search for new information when curiosity is of interest-type. This interest is captured in positive attitude, which goes with love of learning in the domain of interest. Thus, when there is a gap in one’s knowledge, this is not perceived as a threat but as an opportunity for learning. The negative affect associated with deprivation is moderated by the positive expectancy that new learning is imminent, hence the person feels it is worth exploring new possibilities. Similarly, while working on a task that is not particularly interesting by itself, as in the Sideropoulou study, positive attitude helps enter the learning situation in a more positive affective state, and this positive affect moderates the negative effects of lack of interesting activity and/or deprivation-type curiosity in the current task. Positive affect shields against boredom and puts the cognitive system in a “standby” mode. This standby mode alerts the person when something potentially interesting occurs, such as feedback about the correct response in low confidence items or in items in which there were discrepancy and surprise. Whether surprise and curiosity are interrelated is still another challenge for research on epistemic emotions.

CONCLUDING REMARKS

The study of emotions in SRL is at its beginning. Pekrun’s (2006) control-value theory of emotions is a promising line of research as regards the relations of emotions with motivation. The study of epistemic emotions in SRL, however, is in its infancy. Yet, the potential for understanding the functioning of emotions in SRL is high, particularly when we consider the interactions of affect with metacognitive experiences and discrete emotions. This is a novel and promising area of research because it connects affect with the monitoring and control of cognitive processing and gives metacognition a hot aspect that has not been noted in past research. All in all, including affect in SRL gives a new perspective to theories of SRL and an opportunity for a more comprehensive picture of the mechanisms underlying the regulation of learning.

Note

1. The study was carried out under my supervision.

References

Aarts, K., De Houwer, J., & Pourtois, G. (2013). Erroneous and correct actions have a different affective valence: Evidence from ERPs. Emotion, 13(5), 960–973. doi:10.1037/α0032808

Aiken, L. R. (1996). Rating scales and checklists: Evaluating behavior, personality, and attitudes. London, England: Wiley.

Alter, A. L., Oppenheimer, D. M., & Epley, N. (2013). Disfluency prompts analytic thinking but not always greater accuracy: Response to Thompson et al. Cognition, 128(2), 252–255. doi:10.1016/j.cognition.2013.01.006

Autin, F., & Croizet, J-C. (2012). Improving working memory efficiency by reframing metacognitive interpretation of task difficulty. Journal of Experimental Psychology: General, 141(4), 610–618. doi:10.1037/a0027478

Berlyne, D. E. (1954). A theory of human curiosity. British Journal of Psychology, 45, 180–191.

Bower, G. D. (1981). Mood and memory. American Psychologist, 36, 129–148.

Carver, C. S. (2015). Control processes, priority management, and affective dynamics. Emotion Review, 7(4), 301–307. doi:10.1177/1754073915590616

Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge, England: Cambridge University Press.

Dermitzaki, Ι., & Efklides, Α. (2000). Aspects of self-concept and their relationship with language performance and verbal reasoning ability. American Journal of  Psychology, 113, 643–659. doi:10.2307/1423475

Dina, F., & Efklides, A. (2009). Metacognitive experiences as the link between situational characteristics, motivation, and affect in self-regulated learning. In M. Wosnitza, S. A. Karabenick, A. Efklides, & P. Nenniger (Eds.), Contemporary motivation research: From global to local perspectives (pp. 117–146). Göttingen, Germany: Hogrefe.

D’Mello, S. (2013). A selective meta-analysis on the relative incidence of discrete affective states during learning with technology. Journal of Educational Psychology, 105(4), 1082–1099. doi:10.1937/a0032674

Eccles, J. S., & Wigfield, A. (2002). Motivational beliefs, values, and goals. Annual Review of Psychology, 53, 109–132. doi:10.1146/annurev.psych.53.100901.135153

Efklides, A. (2008). Metacognition: Defining its facets and levels of functioning in relation to self-and co-regulation. European Psychologist, 13, 277–287. doi:10.1027/1016-9040.13.4.277

Efklides, A. (2011). Interactions of metacognition with motivation and affect in self-regulated learning: The MASRL model. Educational Psychologist, 46, 6–25. doi:10.1080/00461520.2011.538645

Efklides, A. (2014). How does metacognition contribute to the regulation of learning? An integrative approach. Psychological Topics, 23(1), 1–30.

Efklides, A. (2016). Metamemory and affect. In J. Dunlosky & U. Tauber (Eds.), The Oxford handbook of metamemory (pp. 245–267). New York, NY: Oxford University Press.

Efklides, A., & Petkaki, C. (2005). Effects of mood on students’ metacognitive experiences. Learning and Instruction, 15, 415–431. doi:10.1016/j.learninstruc.2005.07.010

Efklides, A., Schwartz, B. L., & Brown, V. (in press). Motivation and affect in self-regulated learning: Does metacognition play a role? In D. Schunk & J. Greene (Eds.), Handbook of self-regulation of learning and performance (2nd ed.). London, England: Routledge.

Efklides, A., & Tsiora, A. (2002). Metacognitive experiences, self-concept, and self-regulation. Psychologia, 45, 222–236.

Efklides, A., & Vlachopoulos, S. P. (2012). Measurement of metacognitive knowledge of self, task, and strategies in mathematics. European Journal of  Psychological Assessment, 28, 227–239. doi:10.1027/1015-5759/a000145

Elpidorou, A. (2014, November). The bright side of boredom. Frontiers in Psychology, 5(Article 1245), 1–4. doi:10.3389/fpsyg.2014.01245

Flavell, J. (1979). Metacognition and cognitive monitoring: A new area of developmental inquiry. American Psychologist, 34, 906–911. doi:10.1037/0003/0003-066X.34.10.906

Forgas, J. P. (1994). The role of emotion in social judgments: An introductory review and an affect infusion model (AIM). European Journal of Social Psychology, 24, 1–24. doi:10.1002/ejsp.2420240102

Frijda, N. (1986). The emotions. Cambridge, England: Cambridge University Press.

Koriat, A., & Nussinson, R. (2009). Attributing study effort to data-driven and goal-driven effects: Implications for metacognitive judgments. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, 1338–1343. doi:10.1037/a0016374

Litman, J. A. (2008). Interest and deprivation factors of epistemic curiosity. Personality and Individual Differences, 44, 1585–1595. doi:10.1016/j.paid.2008.01.014

Litman, J. A. (2010). Relationships between measures of I- and D-type curiosity, ambiguity tolerance, and need for closure: An initial test of the wanting-liking  model of information seeking. Personality and Individual Differences, 48, 397–402. doi:10.1016/j.paid.2009.11.005

Litman, J. A., Crowson, H. M., & Kolinski, K. (2010). Validity of the interest- and deprivation-type epistemic curiosity distinction in non-students. Personality and Individual Differences, 49, 531–536. doi:10.1016/j.paid.2010.05.021

Litman, J. A., Hutchins, T. L., & Russon, R. K. (2005). Epistemic curiosity, feeling of knowing, and exploratory behavior. Cognition and Emotion, 19(4), 559–582. doi:10.1080/02699930441000427

Litman, J. A., & Pezzo, M. V. (2007). Dimensionality of interpersonal curiosity. Personality and Individual Differences, 43, 1448–1459. doi:10.1016/j.paid.2007.04.021

Litman, J. A., & Spielberger, C. D. (2003). Measuring epistemic curiosity and its diversive and specific components. Journal of Personality Assessment, 80(1), 75–86.

Loewenstein, G. (1994). The psychology of curiosity: A review and reinterpretation. Psychological Bulletin, 116, 75–98. doi:10.1037/0033-2909.116.1.75

Mandler, G. (1975). Mind and emotion. New York, NY: Wiley.

Mandler, G. (1984). Mind and body: Psychology of emotion and stress. New York, NY: Norton.

Metallidou, P., & Efklides, A. (2001). The effects of general success-related beliefs and specific metacognitive experiences on causal attributions. In A. Efklides, J. Kuhl, & R. M. Sorrentino (Eds.), Trends and prospects in motivation research (pp. 325–347). Dordrecht, The Netherlands: Kluwer.

Meyer, W., Reisenzein, R., & Schützwohl, A. (1997). Toward a process analysis of emotions: The case of surprise. Motivation and Emotion, 21(3), 251–274.

Muis, K. R., Psaradellis, C., Lajoie, S. P., Di Leo, I., & Chevrier, M. (2015). The role of epistemic emotions in mathematics problem solving. Contemporary Educational Psychology, 42, 172–185. doi:http://dx.doi.org/10.1016/j.cedpsych.2015.06.003

Mussel, P. (2010). Epistemic curiosity and related constructs: Lacking evidence of discriminant validity. Personality and Individual Differences, 49, 506–510. doi:10.1016/j.paid.2010.05.014

Nelson, T. O. (1996). Consciousness and metacognition. American Psychologist, 51, 102–116. doi:10.1037//0003-066X.51.2.102

Nussinson, R., & Koriat, A. (2008). Correcting experience-based judgments: The perseverance of subjective experience in the face of the correction of judgment. Metacognition and Learning, 3, 159–174. doi:10.1007/s11409-008-9024-2

Pekrun, R. (2006). The control-value theory of achievement emotions: Assumptions, corollaries, and implications for educational research and practice. Educational Psychology Review, 18, 315–341. doi:10.1007/s10648-006-9029-9

Pekrun, R., Elliot, A. J., & Maier, M. A. (2009). Achievement goals and achievement emotions: Testing a model of their joint relations with academic performance. Journal of Educational Psychology, 101(1), 115–135. doi:10.1037/a0013383

Schunk, D. H., & Zimmerman, B. J. (Eds.). (1994). Self-regulation of learning and performance: Issues and educational implications. Hillsdale, NJ: Erlbaum.

Schwarz, N. (2005). When thinking feels difficult: Meta-cognitive experiences in judgment and decision making. Medical Decision Making, 25(1), 105–112. doi:10.1177/0272989X04273144

Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and judgments of well-

being: Informative and directive functions of affective states. Journal of Personality and Social Psychology, 45, 513–523. doi:10.1037//0022-3514.45.3.513

Sideropoulou, E. (2015). Περιέργεια ως χαρακτηριστικό προσωπικότητας και πρόθεση εξερεύνησης: Επίδραση του έργου και αλληλεπίδραση με τις μεταγνωστικές εμπειρίες και το θυμικό [Curiosity as trait and as intention for exploration: Task effects and interactions with metacognitive experiences and affect] (Unpublished MA thesis). School of Psychology, Aristotle University of Thessaloniki, Greece.

Thagard, P. (2006). Hot thought: Mechanisms and applications of emotional cognition. Cambridge, MA: MIT Press.

Toner, E., Haslam, N., Robinson, J., & Paige, W. (2012). Character strengths and well-being in adolescence: Structure and correlates of the Values in Action Inventory of Strengths for Children. Personality and Individual Differences, 52, 637–642. doi:10.1016/j.paid.2011.12.014

Topolinski, S., & Strack, F. (2015, February). Corrugator activity confirms immediate negative affect in surprise. Frontiers in Psychology, 6(Article 134), 1–8. doi:10.3389/fpsyg.2015.00134

Tornare, E., Czajkowski, N. O., & Pons, F. (2015). Children’s emotions in math problem-solving situations: Contributions of self-concept, metacognitive experiences, and performance. Learning and Instruction, 39, 88–96. http://dx.doi.org/10.1016/j.learninstruc.2015.05.011

Touroutoglou, A. (2009). Η διακοπή της γνωστικής επεξεργασίας: Γνωστικές, μεταγνωστικές, και θυμικές επιπτώσεις [Interruption of cognitive processing: Cognitive, metacognitive, and affective effects] (Unpublished Ph.D. thesis). School of Psychology, Aristotle University of Thessaloniki, Greece.

Touroutoglou, A., & Efklides, A. (2010). Cognitive interruption as an object of metacognitive monitoring: Feeling of difficulty and surprise. In A. Efklides & P. Misailidi (Eds.), Trends and prospects in metacognition research (pp. 171–208). New York, NY: Springer.

Usher, E. L. (2009). Sources of middle school students’ self-efficacy in mathematics: A qualitative investigation. American Educational Research Journal, 46, 275–314. doi:10.3102/0002831208324517

Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92, 548–573. doi:10.1037/0033-295X.92.4.548

Weiner, B. (2014). The attribution approach to emotion and motivation: History, hypotheses, home runs, headaches/heartaches. Emotion Review, 6(4), 353–361. doi:10.1177/1754073914534502

Winkielman, P., & Cacioppo, J. T. (2001). Mind at ease puts a smile on the face: Psychophysiological evidence that processing facilitation elicits positive affect. Journal of Personality and Social Psychology, 81(6), 989–1000. doi:10.1037//0022-3514.81.6.989

Zimmerman, B. J., & Schunk, D. (Eds.). (1989). Self-regulated learning and academic achievement: Theory, research, and practice. New York, NY: Springer.