Contagious Flow

Abstract

The current study examined undergraduate student understanding of, and interest in, course material as potential antecedents to student experiences of flow within a classroom setting. In addition, the social, informative, and contagious nature of flow were examined, as was the influence of being in flow during classroom coverage of material on subsequent quiz performance. Data from 14 students in an intensive course were collected over 15 days. All students provided ratings of their mood at the start of each class and ratings of their experienced flow, their interest in and understanding of material, and perceptions of their classmates’ and the instructor’s flow at the end of each class. In addition, the instructor provided ratings of her flow and perceptions of the class at the start and end of each class. Finally, students completed daily quizzes over the previous day’s material. Results revealed that, controlling for prior mood, understanding of and interest in material were related to daily reports of flow experiences. In addition, evidence for the social validation and contagion effects of flow emerged. Contrary to expectations, flow during knowledge acquisition was not related to subsequent quiz performance over the material. Practical and theoretical implications are discussed.

Keywords

flow performance contagion education teaching

When individuals are so deeply engaged and immersed in an activity that they seemingly forget about everything but the activity itself, they are said to be in a state of flow (Csikszentmihalyi, 1990). While in flow, individuals experience a sense of control, an intense level of concentration, a loss of self-consciousness, and a feeling that time has been altered (Csikszentmihalyi, 1975, 2000; Jackson & Csikszentmihalyi, 1999). Researchers have noted that a key precursor to the experience of flow is a balance between task demands and skill levels of individuals, such that as one becomes more proficient with a particular task, the challenge must also be increased. As Csikszentmihalyi and LeFevre (1989) noted, “when both challenges and skill are high, the person is not only enjoying the moment, but is also stretching his or her capabilities with the likelihood of learning new skills and of increasing self-esteem and personal complexity” (p. 816). In this manner, the concept of flow is inherently relevant to learning and particularly important within educational settings.

Research findings regarding flow within learning contexts have demonstrated that flow is associated with heightened creativity (Perry, 1999; Sawyer, 1992), persistence in studies (Nakamura, 1988), and overall learning and academic performance (Csikszentmihalyi, Rathunde, & Whalen, 1993; Heine, 1996; Nakamura, 1988). In addition, there is evidence that flow is related to teaching effectiveness (Csikszentmihalyi, 1996) and that flow within the classroom can crossover from one individual to others (e.g., from teacher to students; Bakker, 2005).

The understanding of flow within educational settings, however, is limited by the methodological approaches to date. For example, most research on the correlates of flow has been cross-sectional and therefore incapable of establishing the causal nature of the relationships between the potential antecedents and consequences of flow. In addition, cross-sectional studies must rely on measures assessing recalled flow as opposed to direct measures of flow at the times of the activities. As such, the measures are prone to biases in that flow could be confused with other positive states (e.g., enjoyment of a class, liking of a teacher). Indeed, Larson and Csikszentmihalyi (1983) suggest that flow would be best measured through experience sampling methodology, a methodology that assesses the thoughts, feelings, and behaviors associated with flow as they occur in the natural environment. By using such an approach, issues of retroactive assessment would be moot. Furthermore, recent research has demonstrated daily fluctuations in individuals’ experiences of flow (Fullagar & Kelloway, 2009). As such, the need to assess flow beyond a single point in time is imperative in gaining an accurate depiction of antecedents and consequences of flow, as well as crossover effects, within learning contexts.

The purpose of this study is threefold. First, we sought to examine two likely antecedents to student flow: (a) student understanding of the material covered in the class and (b) interest in what was being taught in the class. Given the need for challenge–skill balance for flow to occur, there must be a level of cognitive competence (i.e., understanding of the material) that is in line with the material being presented. Similarly, because flow is a form of intrinsic motivation that “is so enjoyable that people will do it even at great cost, for the sheer sake of doing it” (Csikszentmihalyi, 1990, p. 4), it is reasonable to believe that interest in the material is necessary for individuals to experience flow. Second, we aimed to examine the social validating and contagious nature of flow. Specifically, we explored the extent to which individuals are more likely to experience flow if (a) they believe others in class are in a state of flow (i.e., a social validation effect) and (b) their instructor is in a state of flow (i.e., a contagion effect). Finally, our third purpose was to examine the influence of being in flow during classroom coverage of material on subsequent recall and recognition of the material.

Understanding of Material and Flow

One of the most consistent findings within the flow literature is the need for a balance between perceived task demands and skill levels for flow to occur (Csikszentmihalyi, 1990, 1997; Nakamura & Csikszentmihalyi, 2009). In short, when the perceived demands of the task exceed one’s skills and abilities, individuals are apt to be frustrated or anxious, and unable to get into a state of flow. Conversely, when skills and abilities exceed perceived task demands, rather than experience flow, individuals are likely to be overly relaxed or apathetic (Csikszentmihalyi, 1975; Csikszentmihalyi et al., 1993). It is only when skills and abilities are at a moderate to high level, and balanced, that individuals can experience flow (Keller & Bless, 2008).

Within an educational context, challenge–skill balance can be viewed as hinging around cognitive competence, or the ability to understand and comprehend concepts. If material is too confusing, or beyond student capabilities, the student is not likely to be intrinsically motivated (Niemiec & Ryan, 2009). When in flow, actions feel automatic and effortless. In the classroom, the actions would be the mindful interaction with the material and with others in the context of learning. If information being presented is overly challenging, students will have to exert more effort to determine how the information fits in with what they already know and identify how best to store the information. Although heightened concentration is a key part of flow (cf. Hektner & Asakawa, 2000), having to concentrate too much without the reward of grasping the concept is only likely to lead to frustration or anxiety, not flow (Ahmed, van der Werf, Minnaert, & Kuyper, 2010). However, when material is presented in a way that matches the skill level of the students, encoding and storing information for later retrieval becomes much easier. Thus, learning is likely to feel effortless when material is viewed as understandable (Schweinle, Meyer, & Turner, 2006).

The relationship between the understanding of material and flow can also be examined through the lens of control. Just as a challenge–skill balance is necessary for flow to occur (Nakamura & Csikszentmihalyi, 2009), another important ingredient for flow is a sense of control (Chen, 2000; Csikszentmihalyi, 2000; Jackson & Csikszentmihalyi, 1999). Within the classroom, a high understanding of material should presumably make a student feel in control of the class, able to meet the requirements of the course. In this manner, students with a greater perceived understanding of concepts should be more likely to experience flow within a classroom setting.

Hypothesis 1: Understanding of material is positively related to flow.

Interest and Flow

Interest in the material is another critical component needed for students to experience flow in the classroom. One important precondition for flow is that there is clear purpose for engaging in the activity (Guo & Poole, 2009). In an academic setting, a student’s interest in a topic provides a sense of purpose for learning about it (Shernoff & Csikzentmihalyi, 2009). Moreover, “cognitive curiosity and intrinsic interest . . . are the initial conditions of both control and focus of attention” (Ryu & Parsons, 2012, p. 710). Given the role of control and heightened concentration in the experience of flow (Csikszentmihalyi, 2000; Hektner & Asakawa, 2000), interest in material is clearly important. Interest in the material directs focus and makes students want to learn more about what is being presented, which increases the likelihood that they will experience flow in the classroom (Shernoff & Csikzentmihalyi, 2009). Indeed, scholars have shown that students report higher levels of engagement when they find material both challenging and relevant (Shernoff, Csikzentmihalyi, Schneider, & Shernoff, 2003).

Although interest is a critical component for an individual to experience flow, it seems likely that the relationship is reciprocal in that experiencing flow can also lead to an increased interest in an activity. Flow experiences are so enjoyable that individuals will seek out those activities for no other reason than to experience that feeling again (Csikzentmihalyi, Abuhamdeh, & Nakamura, 2005). There is some neuroscientific evidence that suggests individuals gain pleasure from seeking out new and challenging information (Biederman & Vessel, 2006). Processing new and interesting information is associated with the production of endomorphines that are responsible for the experience of pleasure (Pert & Snyder, 1973). Furthermore, this research suggests that more novel and complex activities elicit more pleasure than simple tasks (Biederman & Vessel, 2006). This research would suggest that engagement in complex and challenging tasks is physiologically pleasurable, and that individuals would be motivated to seek out more opportunities to gain new information about a subject. In line with this, it is not surprising that several flow researchers believe individuals play an active role in trying to build experiences into their lives that have been associated with flow in the past (Delle Fave & Massimini, 2005; Delle Fave, Massimini, & Bassi, 2011). Said differently, just as interest in material may direct focus and thereby increase the likelihood that they will experience flow in the classroom (Shernoff & Csikzentmihalyi, 2009), experiencing flow in an activity may help to foster subsequent interest in the activity as well.

Hypothesis 2: Interest in material is positively related to flow.

Social Experience of Flow

Flow is typically discussed as a purely individualistic phenomenon, separate from the experiences of others in their environment. Nevertheless, there is reason to believe that, when in a group setting such as the classroom, the experience of flow is a social phenomenon in which the presence of others is used to gauge one’s own flow experience. Recent research (Walker, 2010) has indicated that individuals experience more enjoyment when flow is experienced as a social rather than a solitary activity. This research suggests that collective flow entails a high level of attention and focus on other group members and less awareness of self (Walker, 2010).

Independent of the research on collective flow, two frameworks guide the notion that individuals’ flow experiences are likely influenced by those around them. First, social comparison theory (e.g., Festinger, 1954) suggests that individuals turn to others in their environment for cues about how to think and feel. Furthermore, when comparing their own views with others, individuals are likely to look to similar others for comparison, as information from like individuals is informative and salient for self-evaluation (Festinger, 1954; Tesser, 1988). For example, in an experiment on the crossover of burnout among soldiers, Bakker, Westman, and Schaufeli (2007) found that soldiers were greatly influenced by attitudes of their peers (i.e., fellow soldiers), even more so than they were by attitudes of those in higher status positions (i.e., squadron leaders). Thus, within the classroom, students are likely to ascertain whether their classmates appear to be actively engaged and immersed in the content to dictate their own flow experiences.

A second framework that further suggests that flow may be socially constructed is the principle of social validation (Cialdini, 2009). This broadly states that individuals often look to the behavior of others when deciding how to behave across situations, especially ambiguous contexts. Highlighting the importance of social norms, the principle of social validation suggests that individuals seek social evidence that their own attitudes and behaviors are appropriate. Thus, when others appear to be similar in their beliefs and actions, people view their own attitudes and beliefs as being appropriate (e.g., Guadagno, Muscanell, Rice, & Roberts, 2013).

Although individuals are influenced by those who are similar to them (e.g., their peers), this is not to say that they are not affected by those who are less similar, but still in the vicinity of the individual. For example, although Bakker et al. (2007) found that soldiers were more influenced by the attitudes of fellow soldiers, they were still influenced by their squadron leaders as well. In summary, and in line with social comparison theory (e.g., Festinger, 1954), the principle of social validation (Cialdini, 2009), and prior work on collective flow (Walker, 2010), it appears that individuals look to the experiences of others when determining their own attitudes and beliefs about an event. Within the classroom, we would expect individuals to look both to their peers and the instructor for clues on their own flow experiences, as all would be involved in the learning context.

Hypothesis 3a: Student flow and perceptions of class flow are positively correlated.

Hypothesis 3b: Student flow and perceptions of instructor flow are positively correlated.

Flow Contagion

In addition to the social experience of flow in which one’s perceptions of their own flow may be influenced by their perceptions of others’ flow, it is also possible that there is a contagious element of flow such that an individual’s flow is influenced by another’s actual experience of flow. The way in which instructor and student experiences of flow reciprocally influence each other can be explained by both crossover effects (Westman, 2011) and emotional contagion theory (Hatfield, Cacioppo, & Rapson, 1994). While much of the research on both of these theories has been predominantly focused on negative emotional crossover and contagion (e.g., burnout, Bakker, Demerouti, & Schaufeli, 2003; stress, Westman, 2011), underlying premises of the theories also apply to positive emotional states such as that of flow. Crossover is a “dyadic, inter-individual transmission of stress or strain that occurs within a particular domain such as the workplace or the family” (Bakker, Demerouti, & Schaufeli, 2005, p. 662). Although the focus within this definition is on the negative emotional states of stress and strain, the process could certainly apply to positive emotions and emotional states. Indeed, researchers have found that positive states crossover between individuals, including feelings of general positive emotions (Westman, Shadach, & Keinan, 2013) and vigor (Westman, Etzion, & Chen, 2009).

Emotional contagion theory (Hatfield et al., 1994) also suggests that instructor and student flow will interrelate. Emotional contagion is “the tendency to automatically mimic and synchronize facial expressions, vocalizations, postures and movements with those of another person and, consequently, to converge emotionally” (Hatfield et al., 1994, p. 5). This contagion effect has been demonstrated both experimentally (e.g., Barsade, 2002) as well as in field settings (e.g., Bakker, van Emmerik, & Euwema, 2006; Totterdell, Kellett, Teuchmann, & Briner, 1998). Within the classroom, a reciprocal induction effect can occur when the instructor and students observe and react to each others’ emotional displays. There is also some evidence that indicates that in collective or social flow situations, a process of emotional communication and contagion occurs within the group (Ryu & Parsons, 2012; Walker, 2010).

In an early examination of the contagious nature of flow, Bakker (2005) found support for the crossover of flow from music teachers to their students, providing one of the first demonstrations in a field setting that positive experiences in general, and flow specifically, could cross over from one individual to another. This study, however, was limited in that it used a cross-sectional design involving a select few students from a class, chosen by the teacher. Such a selection strategy could have resulted in a biased sample in that the teachers may have chosen the more engaged students as opposed to the typical student. Furthermore, Bakker relied on a retroactive assessment of flow—that is, recalled flow as opposed to a direct of measure of flow at the time of the activity. Consequently, the measure was potentially more prone to biases, such as flow being confused with enjoyment of the class or liking of the teacher. As a result of some of these limitations, Bakker noted that it would be “interesting and relevant to examine this phenomenon in other teacher–student relationships” (p. 40). We do so in the current study, while addressing some of the limitations inherent in Bakker’s study by incorporating both instructor and student perspectives at multiple time periods and using all students in a class rather than a select few.

Despite the limitations of Bakker (2005), his findings align with crossover and contagion theories and help guide our expectations. Specifically, we propose the following hypothesis.

Hypothesis 4: Instructor and student flow are positively correlated.

Flow and Subsequent Academic Performance

Flow has been shown to be a positive state that acts according to the broaden-and-build theory of positive emotion (Fredrickson, 1998, 2001; Fullagar, van Ittersum, & Knight, 2012). This theory proposes that positive emotions broaden our awareness, cognitive thought processes, and action repertoires and enable us to function more optimally. Flow has been shown, for example, to induce positive emotions and facilitate higher levels of performance in video game situations (Fullagar et al., 2012). Within educational setting, this optimal functioning would likely translate into heightened academic performance.

There is some research that suggests flow is associated with optimal performance (Nakamura & Csikszenmihalyi, 2005). For example, in the workplace, flow has been associated with (a) required in-role behaviors that fulfill organizational goals and (b) discretionary extra-role behaviors that promote effective organizational functioning (Demerouti, 2006; Eisenberger, Jones, Stinglhamber, Shanock, & Randall, 2005).¹ Among athletes, the experience of flow has been related to self, coach, and objective reports of performance (Bakker, Oerlemans, Demerouti, Slot, & Ali, 2011). Finally, studies indicate that flow is a predictor of academic performance, particularly in one’s talent area (Csikszentmihalyi et al., 1993; Wong & Csikszentmihalyi, 1991).

There is additional reason to suggest that student flow should be related to academic performance. As Ryu and Parsons (2012) noted, “The person in the optimal flow experience becomes absorbed in the learning activity and is more intensively aware of his or her own mental processes, thereby enhancing relevant mental activities, such as remembering, thinking, feeling, and making decisions” (p. 710). In addition, research has revealed relationships between constructs conceptually similar to flow are related to learning and academic performance. For example, positive affect yields beneficial effects for cognitive processes (e.g., Stephanou, 2011). Similarly, task enjoyment is a strong predictor of students’ grades, irrespective of ability levels (Schiefele & Csikszentmihalyi, 1994). Given that flow is a state of positive affect defined in large part by an intense level of concentration and enjoyment (Csikszentmihalyi, 1990), it stands to reason that there is a positive relationship between flow and academic performance.

Several studies have supported the notion that flow is related to academic performance (Mendelson, 2007; Schüler, 2007; Shernoff & Schmidt, 2008) and performance in training situations (Choi, Kim, & Kim, 2007). These studies, however, are limited in that they were cross-sectional, thus limiting the extent to which causal statements can be made between flow and academic performance. Rather than flow leading to heightened grades, for example, it could be that doing well academically leads individuals to retrospectively report (perhaps erroneously) heightened feelings of flow.

In the current study, we examine the relationship between flow during a lecture on subsequent performance on a quiz over the lecture material. Through our study design, we establish a lagged situation in which ratings of flow are always gathered immediately following the lecture and the quizzes are always over only the material covered during those specific time periods. As such, we are better able to evaluate the causal nature of flow of and academic performance, despite it still being correlational (cf., Xanthopoulou, Bakker, & Ilies, 2012). Based on the previous research presented, we hypothesized the following.

Hypothesis 5: Flow during classroom coverage of material is positively related to subsequent quiz performance over the material.

Finally, we have argued that understanding of, and interest in, the material are associated with the experience of flow in the classroom (see Hypotheses 1 and 2). However, it could also be argued that flow facilitates an understanding of the material and increases interest in learning more about the topic, such that subsequent performance on tests improves. Consequently, we tested whether interest in the material covered in class, and an understanding of that material, independently mediated the relationship between flow and quiz performance.

Hypothesis 6a: Understanding of the material covered in class mediates the relationship between flow and subsequent quiz performance over the material.

Hypothesis 6b: Interest in the material covered in class mediates the relationship between flow and subsequent quiz performance over the material.

Method

Participants and Procedure

Participants were the instructor responsible for and students enrolled in an introductory survey course in large university in a Midwestern state in the United States. The course took place during 4 weeks in June 2012. Classes were held Monday through Friday, and each class period lasted 2 hours. Three classes (of the 20 days) were cancelled due to conference travel by the instructor. These days were planned absences and, as such, the course was designed around them (i.e., no material was “lost” as a result of cancelling the classes). The course was primarily lecture-based, and due to the intensive nature of the course, each class covered essentially one chapter from the textbook. The instructor used PowerPoint, and students were provided with the slides ahead of time to facilitate note taking. In the class period following each lecture, a quiz over the previous day’s material was given (e.g., a lecture on Wednesday covering Chapter 4 material would be followed on Thursday with a quiz over that chapter). After the quiz, which students were given 20 minutes to complete, the next chapter’s lecture would begin.

A total of 14 students (100% of the students enrolled in the course) participated in the study, of which 7 were female. We present information within the section describing our analytic strategy that demonstrates this sample size was sufficient for the current study’s research questions. The average age of participants was 20.58 years (SD = 1.83). Most respondents (78.57%) were Caucasian. The average self-reported GPA (on a 1-4 scale) was 2.9 (SD = 0.68). Participants were informed that their participation was voluntary, responses were confidential, and their decision to participate (or not) would not affect their grades. To ensure confidentiality, all students created their own five-digit code to use when completing all measures throughout the course. On a separate sheet collected at the end of the course, students indicated what their code was, in order to link their daily surveys to their quiz scores.

Immediately prior to each class, students completed one preclass survey that assessed their mood at the start of class. These mood sheets were available as students entered the class and were submitted in a pile at the front of class. After each class, students completed a postclass survey that assessed their views regarding the class, the instructor, and their feelings of flow during the class period. Students placed the completed sheets in the same pile as the presurveys, which were then transferred to an envelope to be given to a separate researcher to enter the data. The instructor did not view the data so as to not influence grading or subsequent class preparation and instruction. Students were aware that the instructor was not going to see the individual data sheets, further ensuring the confidentiality of their responses. Finally, the instructor also completed a pre- and postclass survey daily to assess similar characteristics.

Data from the instructor and students were collected for 15 days. According to Reis and Wheeler (1991), a “2-week record-keeping period is assumed to represent a stable and generalizable estimate of social life” (p. 287). In addition, previous research using similar methodology (i.e., daily surveys) have used 2-week time periods (e.g., Culbertson, Mills, & Fullagar, 2012; Ilies et al., 2007; Judge, Ilies, & Scott, 2006). Thus, 15 days should be sufficient to indicate daily variations (i.e., state fluctuations) as well as daily invariance (i.e., trait characteristics).The mean daily survey response was 14.14 days (SD = 1.03).

Measures

Student Understanding and Interest

After each class period, students completed a brief survey that included items to assess their understanding and interest in the material presented that day. Understanding of the material was assessed with the item, “I clearly understood the material.” Interest in the material was assessed with the item, “The material presented in class was interesting.” Response options ranged from 1 (Strongly disagree) to 5 (Strongly agree). We chose to use single-item measures in order to keep the daily survey from being overly cumbersome for participants. This decision should not have resulted in sacrificed measurement quality as single-item measures have been shown to have good convergent and divergent validity as well as high correlations with multiple-item scales (Abdel-Khalek, 2006; Nagy, 2002).

Student and Instructor Self-Reported Flow

At the completion of each class period, the instructor and the students completed an assessment of their flow during that day’s class. Participants responded to nine items that corresponded to the nine dimensions of flow (challenge–skill balance, action–awareness merging, clear goals, clear feedback, concentration, sense of control, loss of self-consciousness, transformation of time, autoletic experience; Jackson & Csikszentmihalyi, 1999). Sample items include, “I was completely focused” and “I was not worried about what others may be thinking of me.” Response options ranged from 1 (Strongly disagree) to 5 (Strongly agree). Across days, the average Cronbach’s alpha for the flow measure was .82 (minimum α = .73, maximum α = .93).

Student Perceptions of Class and Instructor Flow

Two items were included in the daily postclass survey that assessed student perceptions of class and instructor flow. Class flow was assessed with the item, “Students in the class seemed to be ‘switched on.’” Instructor flow was assessed with the item, “The instructor seemed to be in the ‘zone.’” Again, response options ranged from 1 (Strongly disagree) to 5 (Strongly agree).

Student Academic Performance

Academic performance was measured through daily quizzes over material from the prior class period. Specifically, each class covered material from a specific topic area. Students were then evaluated on that material the following class period with a relatively short quiz (eight multiple-choice questions plus two short answer questions to assess recall of the material and interpretive skills, respectively). Quizzes were distributed at the start of the class period. Once all quizzes were submitted or 20 minutes had elapsed (whichever came first), the quizzes were reviewed and new material for the day was presented. Most students completed the quizzes within 10 minutes. Students were aware that their lowest quiz score would be dropped. Performance on the quizzes accounted for 57% of a student’s final grade (220 of 385 points) in the course.

Control Variables

Using an experience sampling methodology, Fullagar and Kelloway (2009) found that over a 15-week period, flow was consistently associated with momentary positive mood within individuals. As such, we controlled for preclass mood in all analyses. To assess general mood, the instructor and students were asked to rate their mood before the class on a Visual Analog Mood Scale (VAMS; Stern, Arruda, Hooper, Wolfer, & Money, 1997). VAMS have been used extensively in medical research to assess internal mood states and have been shown to be valid (see House, Arruda, Andrasik, & Grazzi, 2012). The scale varied from a value of 0 (Worst—depicted by a frowning iconic face), to 5 (Neutral—depicted by an iconic face with a straight line for a mouth), to 10 (Best—depicted by a smiling iconic face). Given its simplicity and intuitive nature, VAMS is useful when administering mood measures frequently and where time is of a premium, as was the case in the current study.

Analytic Strategy

Our data had a multilevel structure, consisting of multiple classroom observations nested within observations at the course level. Such multilevel data structures are best analyzed using multilevel random coefficient modeling (MCRM) techniques, such as hierarchical linear modeling (HLM; Raudenbush & Bryk, 2002). The HLM approach is an extension of multiple regression that allows investigation of the relationships between variables at multiple levels of analysis. We used a longitudinal approach in analyzing our data in that we took a series of repeated measures for each student over the duration of the course. In our research, observations for each student and the instructor were nested within each class session. It is now believed that MCRM, which uses a maximum likelihood approach to estimate coefficients, is the most appropriate way to analyze such multilevel data structures, advantageous over the traditional ordinary least squares (OLS) method (Nezlek, 2003, 2012).

One issue of concern with our analytic strategy was with the relatively small sample size. Often, it is presumed that HLM cannot be conducted with fewer than 50 participants (Maas & Hox, 2005). That said, the issue is more complicated with within-subjects designs such as ours, and power analyses do not appear to exist in such cases to help determine the appropriate sample size. As such, we ran a series of fivefold cross-validations on models of varying complexity in order to determine whether we were overfitting our data (i.e., we had too many parameters relative to our number of observations). Evidence of overfitting would suggest that we had too few participants for the analyses we conducted, while the absence of such evidence would suggest that our sample size was sufficient.

The fivefold cross-validation is a specific form of a k-fold cross-validation in which we randomly divided our data into five roughly equal sections. The k-fold cross-validation technique is an effective technique for cross-validating results, having been used successfully within direct marketing and neural network research (Cui, Wong, Zhang, & Li, 2008; Thierne, Song, & Calantone, 2000). The k-fold cross-validation technique is advantageous over other means of cross-validating data (e.g., bootstrapping) because of its lower cost of computing power and the fact that it does not result in an overlap of data (Cui et al., 2008). Furthermore, our chosen method of randomly dividing our data is preferable to systematic elimination (e.g., dividing 15 days of testing into the first three, second three, and so on or, dropping three subjects from each) because with systematic elimination, an independent variable may be correlated with the variable used for segmentation.

With this fivefold cross validation, we tested our most complex model, in which we had three predictors (“prior mood,” “understanding,” and “interest”) compared with a more complex model (in which we allowed the “understanding” slope to vary across subjects) and three simpler models: “understanding” with “prior mood,” “understanding,” with “interest,” and only “understanding.” When we examined our most complex model with the more complex model (with the added slope effect), R² improved but R²-cross-validated decreased, thus suggesting overfitting for the more complex model. Furthermore, we examined the Akaike information criterion (AIC) and Bayesian information criterion (BIC), which are used as further criteria to ensure that overfitting is not a concern (Burnham & Anderson, 2004). Both AIC and BIC introduce a penalty term for the number of parameters in the model. BIC is a stricter criterion than AIC, with the penalty term being larger in BIC than in AIC. AIC was almost identical for the two models and BIC was worse for the more flexible model. Thus, overall, the most complex model—one that is more complex than what we actually do in our study—suggests overfitting. When we examined each of the simpler models, both R² and R²-cross-validated decreased, but only slightly (about a 2% drop in both R² values when we dropped “interest,” about 2% to 2.5% in both when we dropped “prior mood” but retained interest, and a 4% drop in both when we dropped “interest” and “prior mood”).

Taken together, the results of these analyses provide solid evidence that there is no serious overfitting occurring for our most complex analyses (i.e., when we include all three predictors in a single model). When using AIC and BIC as the criteria for establishing overfitting, AIC favors the model containing all three, whereas BIC favors the model that omits “interest.” Given that BIC is more conservative, with a larger penalty term, this is not surprising.

In sum, it seems there is solid evidence that we are not overfitting our model, and thus our sample size, though limited, does not appear to be a substantial problem given its ability to cross-validate, as evidenced by the 5-fold cross-validations with our most complex model. This said, we urge readers to keep the modest sample size in mind. We return to this issue in the discussion.

Results

We used HLM (Version 6.06; Raudenbush, Bryk, & Congdon, 2000) to generate a series of random-coefficient models to test our research hypotheses. In these Level 1 analyses, the class sessions and the student responses to the daily surveys were the units of analysis. Means, standard deviations, and inter-correlations across class session averages are presented in Table 1.

Table 1.

Means (M), Standard Deviations (SD), and Intercorrelations Between Students (N = 14) and Within Students^a (N = 198) for all Study Variables.

Level 1 variables^b	M	SD	1	2	3	4	5	6	7	8
1. Understanding of material	4.04	0.59		.65*	.03	.75**	.07	.81**	.10	.41
2. Interest in material	4.09	0.66	.47**		.33	.59*	.20	.30	.30	.06
3. Preclass mood	6.71	1.50	.09	.12		.28	.57*	.39	.91**	−.15
4. Self-reported flow	3.94	0.46	.64**	.52**	.28**		.30	.45	.33	.12
5. Perceived class flow	3.69	0.88	.28**	.30**	.24**	.37**		.23	.52	−.17
6. Perceived instructor flow	4.28	0.56	.54**	.30**	.06	.40**	.17		−.25	.47
7. Prequiz mood^c	6.59	1.58	.08	.08	.14	.09	.19*	−.06		.02
8. Quiz total score^c	15.35	3.67	.16	.04	−.09	.02	.01	.24**	−.02

Within-individual correlations are below the main diagonal and between-individual are above.

Daily class session variables were averaged in order to calculate the correlations between individuals.

There were only 12 quizzes and 10 instances when students missed a quiz (N = 158).

p < .05 (two-tailed). **p < .01 (two-tailed).

Before testing the hypotheses, systematic within- and between-individual variance in all six study variables was assessed in analysis of variance (null) models. Results revealed significant between-individual variance for all the study variables (see Table 2). However, in all instances, most of the total variance was attributable to within-individual rather than between-individual variation. Although these variance estimates include variance attributable to measurement error, the results do suggest that understanding of class material, interest in class, preclass mood, perceptions of both the class’ and the instructor’s flow, and self-reported measures of flow are largely state dependent and vary considerably from class session to class session. Self-reported flow exhibited the largest between individual variance (45%).

Table 2.

Estimates of Between- and Within-Class Variation for all Study Variables.

Null model
Level 1: Y = β_0j + σ²	σ^2a	τ₀₀^b	χ^2c(13)	ICC^d
Level 2: β_0j + τ₀₀
Understanding of material	0.25	.10	89.43*	.29
Interest in material	0.34	.11	73.16*	.24
Preclass mood	1.64	.64	86.04*	.28
Self-reported flow	0.12	.10	161.1*	.45
Perceived class flow	0.56	.22	85.03*	.28
Perceived instructor flow	0.22	.10	103.15*	.31

σ² is the Level 1 residual, but since there are no predictors in the null model it represents the within-class variance.

τ₀₀ is the Level 2 residual and represents the between-class variance.

χ² indicates whether τ₀₀ (between-class variance) is significantly different from zero.

The ICC (intraclass correlation) assesses the degree of between-group variance in the dependent measure. ICC = τ/(τ + σ²).

p < .001.

Our first two hypotheses were that flow in class would be positively related to understanding of the material (Hypothesis 1) and interest in the material (Hypothesis 2). We partialled out the effects of mood prior to class by entering this variable first in the random-regression coefficients model. Following Hofmann and Gavin (1998), Level 1 variables were person-centered, where the individual mean of the variable was subtracted from each individual’s class score. The rationale for using individual-mean centering to test our first hypotheses was that we wanted the variance in the intercept term to represent the adjusted between-individual variance in flow after controlling for the effects of prior mood. Results of the random-regression coefficients model (see Table 3) indicated that mood prior to class was significantly related to flow in class: γ₁₀ = 0.09, t(13) = 3.51, p = .004. Although this effect was relatively weak (explaining only 12% of the within-individual variance in flow), its significance justified the control of prior mood in subsequent models.

Table 3.

Parameter Estimates and Variance Components Testing Hypotheses 1 and 2 (Level 1, N = 198; Level 2, N = 14).

Model equations	γ₀₀	γ₁₀	γ₂₀	γ₃₀	σ²	τ₀₀	τ₁₁	τ₂₂	τ₃₃
Random-coefficient regression models^a
Step 1
Level 1: Flow_ij = β_0j + β_1j(Prior Mood_ij) + r_ij
Level 2: β_0j = γ₀₀ + U_0j
Level 2: β_1j = γ₁₀ + U_1j
	3.93**	0.09**	—	—	.11	.10	.00	—	—
Step 2
Level 1: Flow_ij = β_0j + β_1j(Prior Mood_ij) + β_2j(Understanding_ij) + r_ij
Level 2: β_0j = γ₀₀ + U_0j
Level 2: β_1j = γ₁₀ + U_1j
	3.93**	0.06*	0.40**	—	.07	.10	.00	.03	—
Steps 3 and 4
Level 1: Flow_ij = β_0j+ β_1j(Prior Mood_ij) + β_2j(Understanding_ij) + β_3j(Interest_ij) + r_ij
Level 2: β_0j = γ₀₀ + U_0j
Level 2: β_1j = γ₁₀ + U_1j
	3.93**	0.06*	0.34**	0.15*	.06	.10	.00	.02	.01

Note. γ₀₀ is the mean of the intercepts across individuals; γ₁₀ and γ₂₀ are the mean of the slopes across individuals; σ²(r_ij) is the Level 1 residual variance; τ₀₀ (U_0j) is the variance in intercepts.

p < .05. **p < .01.

We then tested the effect of understanding on flow (controlling for mood prior to class) by entering understanding of material into the above random-regression coefficients model (see Table 3). Our results indicated that understanding still had a significant association with flow after controlling for prior mood: γ₂₀ = 0.40, t(13) = 6.08, p < .001. The direction of the regression coefficient shows that higher levels of understanding are associated with greater flow. The magnitude of this relationship was calculated by comparing the within-individual variance in flow in this model with the variance explained by prior mood alone. Understanding of the material accounted for an additional 40% of the within-individual variance in flow, over and above that accounted for by prior mood. Therefore, Hypothesis 1 was supported.

Next, we tested whether interest in the material covered in class was associated with flow, after controlling for prior mood and understanding. Consequently, interest in the material was entered into random-regression coefficients model after the effects of prior mood and understanding had been partialled out. Results showed that interest was significantly associated with flow: γ₃₀ = 0.16, t(13) = 3.68, p = .003 (see Table 3). Interest explained 15% of the within-individual variance in flow (R² = .15). Thus, Hypothesis 2 was supported.

Hypothesis 3 aimed at establishing whether flow is a contagious and socially interactive construct by establishing if students’ self-reported flow was associated with their perceptions of class and instructor flow. To test this hypothesis, we ran a series of random-regression coefficient models (see Table 4). Results indicated a significant relationship between self-reported flow and perceived class flow, γ₁₀ = 0.17, t(13) = 3.57, p = .004, and perceived instructor flow, γ₁₀ = 0.25, t(13) = 3.75, p = .003. Perceived class flow (R² = .20) and instructor flow (R² = .21) explained 20% and 21%, respectively, of the within-individual variation in flow. These results support Hypotheses 3a and 3b.

Table 4.

Parameter Estimates and Variance Components Testing Hypotheses 3 and 4 (Level 1, N = 198; Level 2, N = 14).

Model equations	γ₀₀	γ₁₀	σ²	τ₀₀	τ₁₁
Random-coefficient regression models^a
Flow and perceived class flow
Level 1: Flow_ij = β_0j + β_1j(Class Flow_ij) + r_ij
Level 2: β_0j = γ₀₀ + U_0j
Level 2: β_1j = γ₁₀ + U_1j
	3.93**	0.17*	.10	.10	.02
Flow and perceived instructor flow
Level 1: Flow_ij = β_0j + β_1j(Instructor Flow_ij) + r_ij
Level 2: β_0j = γ₀₀ + U_0j
Level 2: β_1j = γ₁₀ + U_1j
	3.93**	0.25*	.09	.10	.03

p < .05. **p < .01.

We also sought to determine if the class instructor’s flow was related to the students’ flow. After each class, the instructor completed a flow scale, which was analyzed as a Level 2 variable. Our results indicated that instructor flow was not significantly associated with student flow: γ₁₀ = 0.03, t(12) = 0.27, p = .810. Thus, Hypothesis 4 was not supported.

Hypothesis 5 proposed that flow during the lecture would be associated with subsequent quiz performance. Again, we ran a series of random-regression coefficient models where quiz performance was regressed onto the student’s flow levels in the class prior, controlling for mood prior to taking the quiz. Our results indicated that flow did not predict quiz performance: γ₂₀ = 0.02, t(13) = 0.02, p = .98. Consequently, Hypothesis 5 was not supported.

Hypotheses 6a and 6b suggested that the relationship between flow and quiz performance would be mediated by an understanding of the material covered in class and an interest in that material. As there was no relationship between flow and performance, one of the basic assumptions underlying mediation effects was violated (i.e., that there should be a significant zero-order relationship between the predictor and outcome variable; Baron & Kenny, 1986). Consequently neither hypothesis was supported.

Discussion

According to Csikszentmihalyi (1990), flow exists when “people are so involved in an activity that nothing else seems to matter at the time; the experience is so enjoyable that people will do it even at great cost, for the sheer sake of doing it” (p. 4). Most of the research on flow has focused on voluntary leisure and sporting activities. Given the concerns that students all too often lack intrinsic motivation and interest in classroom material (e.g., Debnath, Tandon, & Pointer, 2007; McEvoy, 2011), one might expect that the classroom is not a place where students would be if it were not a requirement. However, research has shown that flow is also apparent in work-related activities (e.g., Csikszentmihalyi, 1975; Csikszentmihalyi & LeFevre, 1989; Demerouti, 2006; Fullagar & Kelloway, 2009; Nielsen & Cleal, 2010). The subjective experience has been found to be remarkably consistent across work and play, suggesting that it is the quality of the experience that is important and not the nature of the activity (Csikszentmihalyi & LeFevre, 1989). Indeed, our findings suggest students are able to experience flow in academic work and within the context (and confines) of a lecture-based summer course.

We examined the impact of perceived understanding of the course material and an interest in the subject matter on flow. Previous research on flow has identified three task-related antecedents of flow that facilitate optimal experience when performing the task: (a) there must be a balance between the perceived challenges of the task and the skills necessary to perform it, (b) the activity should have clear and specific goals, and (c) there should be some form of feedback concerning performance on the task (Nakamura & Csikszentmihalyi, 2009). As noted, these elements are inherent in the task itself. The current research indicated two individual difference variables that may serve as preconditions to the experience of flow. Our results revealed that both perceived understanding of, and interest in, course material were related to individuals’ experiences of flow during class. We have argued that the relationship between understanding and flow may be explained by the fact that students with greater cognitive competence may feel more in control and more likely to be able to meet the demands intrinsic to an academic subject and achieve the critical balance between challenges and skills that is necessary to experience flow (Csikszentmihalyi, 1990, 1997). In terms of interest in the material being taught, our results would confirm the research that has indicated that students who are interested are more likely to be engaged and experience flow (Shernoff & Czikszentmihalyi, 2009).

In addition, we wanted to ascertain whether there was a contagious component to flow. Specifically, we asked whether flow was affected by student perceptions of their classmates’ and teacher’s flow. We found that there appears to be a process of flow contagion, whereby student perceptions of their classmates’ flow and their instructor’s flow are associated with their perceptions of their own experiences of flow. It would appear that positive emotional states, as perceived in others, have a contagious component and that flow may indeed crossover from one person to another in collective situations. Both the sources that we investigated (instructor and students) had similar effect sizes explaining approximately 20% of the variance in self-reported flow. It may be that engaged teachers and students in the classroom create a learning climate that is more conducive to flow and that generates greater interest in the subject matter and facilitates understanding.

In terms of the null relationship between flow and performance on the material covered while in flow, it is first worthy of note that, in the current study, the quizzes always followed the material the very next day. So, unlike other class situations that are common in which there are multiple days between the presentation of material and the testing over it, there was minimal time delay between the presentation of material and the testing over the material. In this manner, the situation was optimal in terms of being able to determine a link between flow during knowledge acquisition and subsequent knowledge recall and recognition. The only stronger situation would have been if the quiz immediately followed the class, which is highly atypical in educational settings and arguably not indicative of true learning (which incorporates a time element; for a review, see De Houwer, Barnes-Holmes, & Moors, 2013). This situation would not be unheard of, however, within organizational training programs, where tests of learning frequently immediately follow the training programs. This all said, we offer three possible explanations for the null relationship between flow and performance. First, as Ullén et al. (2012) found that the tendency to experience flow (i.e., flow proneness) is related to personality, and not cognitive ability, noting that flow may “be a state of subjectively effortless attention that . . . has different underlying mechanisms from attention during mental effort” (p. 171). Along these lines, Konradt and Sulz (2001) found that flow was related to increased concentration and motivation, but not to actual learning performance. As such, it could be that the energy and effort that is expended while in flow is counter to what is needed for the level of information processing that is necessary to perform well academically.

A second explanation is the simple fact that, being state-like (Fullagar & Kelloway, 2009), the experience of flow dissipates. As Debus, Sonnentag, Deutsch, and Nussbeck (2014) noted, “Flow is considered to be a fragile state and a short-term peak experience.” In line with this, we conducted post hoc analyses to examine whether there were meaningful trends in within-person changes in flow over the duration of our study. Not surprisingly, and in line with evidence that suggests that flow is a short-term and fragile state that is primarily responsive to daily, momentary activities (Bakker, 2005, 2008; Ceja & Novarro, 2011; Debus et al., 2014; Nielsen & Cleal, 2010), we did not find any statistically significant effects in this regard.² Without lasting effects, individuals may not feel the same motivation after class to study for quizzes. Given that most information is forgotten shortly after it is presented and will be lost without adequate rehearsal (e.g., Ebbinghaus, 1885/1964), students who are not motivated to study may be ill-prepared for a test of recognition and recall, regardless of the amount of flow experienced during knowledge acquisition.

Finally, related to the preceding explanation, being in flow may lull students into a false sense of security regarding the material, leading them to purposefully choose not to study. For example, prior research has demonstrated that flow is related to student perceptions of learning (Guo, Klein, Ro, & Rossin, 2007; Rossin, Ro, Klein, & Guo, 2009). Thus, motivation notwithstanding, being in flow may make students feel as if rehearsal of material is simply not necessary. More research is needed to test these possibilities.

So what does it mean that flow in class appears to be unrelated to performance on the material covered during that class? Does it mean that instructors need not be concerned with whether their students experience flow in the class? We think the answer is a clear “no” given previous research on flow. For example, prior work has shown that experiencing flow in an activity is related to subsequent interest in that activity (Shernoff & Hoogstra, 2001). By inciting flow in students, instructors could create the desire for students to seek out additional information or pursue knowledge in areas they had not considered before. This said, we should note that we conducted post hoc tests to determine whether experienced flow influenced interest in the material presented in the class the following day.³ Results of the random-regression coefficient model indicated that flow was not significantly related to subsequent interest in the activity: γ₁₀ = 0.19, t(182) = 1.63, p > .05. We argue that this nonsignificant effect may be attributable to two factors. First, the nature of the material covered in each class varied considerably from day to day such that flow may have been generated by entirely different activities (material) on each day. Second, as we have noted above, flow appears to be a short-term, momentary state (Bakker, 2005, 2008; Ceja & Novarro, 2011; Debus et al., 2014; Nielsen & Cleal, 2010). As such, unlike activities in which the same exact activity is repeated and can therefore generate subsequent interest (e.g., rock climbing, playing a musical instrument), the current classroom did not present such activities. Nevertheless, it is possible that students may have heightened interest in the particular content covered; this simply was not the structure of the intense nature of the course.

Another reason that we would argue instructors should care about enhancing flow despite the lack of its relationship with quiz performance involves the mere fact that flow reflects an intense state of enjoyment. That is, student enjoyment may be enough reason to try to increase student flow in a classroom. Not only would heightened enjoyment likely lead to greater interest in material and attendance in class, but researchers have found that students who experience flow tend to report greater satisfaction with the course overall (Guo et al., 2007; Rossin et al., 2009). Indeed, instructors may obtain higher teaching evaluations from a class that tends to be in flow.

Implications and Future Research Directions

There are several practical implications that follow from this study. First, instructors should be cognizant of the fact that understanding and interest are key to students’ experience of flow within the classroom. Specifically, when students reported that they understood and were interested in the material for the day, they also reported greater levels of flow. Thus, the results of our study suggest that presenting information in interesting and clear ways are key to developing intrinsic motivation in a topic. Furthermore, whereby we found no meaningful trends in the experience of flow over the duration of the course (per post hoc analyses), it appears that the extent to which an instructor can facilitate an understanding of and interest in material will relate to that day’s engagement in the material, and not necessarily for subsequent days. That is, students’ experiences of flow appeared to be largely state dependent, varying considerably from one class period to another. As such, although this suggests that instructors should strive to be “on their game” in terms of helping students understand and digest the material for each and every topic, it also suggests that an instructor may be able to “bounce back” from a poorly explained or uninteresting topic and make later topics meaningful.

Second, and also in line with our finding that understanding of material is related to the experience of flow, students should be reminded to seek clarity when they do not understand material as well as engage in preclass preparation (e.g., reading the chapter before class), or they should not expect to feel engaged in the topic area. In this manner, they should take ownership over their class enjoyment. This said, it is important to keep in mind the challenge–skill balance that is key to experiencing flow. If students are expected to prepare before class in order to better understand the material, instructors must be careful not to simply regurgitate the information during the class period. As we noted earlier, when skills and abilities exceed perceived task demands, individuals are likely to experience boredom rather than flow (Csikszentmihalyi, 1975; Csikszentmihalyi et al., 1993). Thus, a student who is prepared can quickly be overprepared if the instructor does not teach accordingly. Given that skills and abilities must be at a moderate to high level, and balanced, for flow to occur (Keller & Bless, 2008), it is important for instructors to encourage (or require) a certain level of preclass preparation and then use the class time in ways that further stimulate and expand on the material.

Our study heeds several researchers’ calls. First, Bakker (2005) suggested future researchers should examine the contagion effect of flow in other teacher–student relationships, which we did. In addition, Bakker, Westman, and van Emmerik (2009) called for research examining crossover theory using within-subjects designs because, unlike cross-sectional designs, they allow for temporal sequencing of events and, “since there is relatively little time between the actual experience (e.g., work engagement) and the reporting on this experience, the retrospective bias is reduced and the validity and reliability are increased” (p. 214). We examined actual reported flow immediately following class. In this manner, we did not have problems with retrospective bias that is common in flow research, particularly research involving cross-sectional data. Finally, Bakker et al. (2009) also noted that future researchers “should include objective measures of cognitive tasks alongside traditional self-report measures, under the rationale that negative crossover can be indicated by decreased cognitive performance and positive crossover can be indicated by enhanced cognitive performance” (p. 214). Although we did not find that flow led to enhanced cognitive performance, we examined the relationship between flow and academic performance using objective measures (i.e., quiz performance) alongside self-report measures (i.e., flow) as suggested.

Future researchers may wish to examine under what circumstances the relationships examined herein may be further heightened or diminished. For example, although Csikszentmihalyi (1975) indicated that the “clearest sign of flow is the merging of action and awareness” (p. 38), researchers have demonstrated that individuals with low self-regulatory skills (Keller & Bless, 2008) and weak internal locus of control (Keller & Blomann, 2008) may be less likely to experience flow even if their skill level matches the task demands. Similarly, results from Eisenberger et al.’s (2005) research would suggest that students who are high in achievement orientation would be more likely to find challenging material interesting and therefore have an increased likelihood of experiencing flow in the classroom. Thus, it would appear that other individual differences that have a direct impact on students’ propensity to experience flow. Indeed, flow proneness has been shown to be associated with certain personality factors, such as neuroticism and conscientiousness, but unrelated to intelligence (Ullén et al., 2012). Future researchers might want to examine other factors that influence the experience of flow and flow proneness within classroom settings.

Another fruitful avenue for future research lies with potential interaction effects between a student’s perceptions of the class’ flow and their understanding of the material being taught. Social comparison theory suggests that individuals turn to others in their environment for cues on how to act and feel (Festinger, 1954). While this theory explained the positive effects we found in the current study for the relationship between perceptions of class flow and self-reported flow, the relationship likely depends on how students perceive their ability, or understanding, in relation to the class. For example, a recent study found that there is a negative effect on students’ views of their academic self-concepts when they compare themselves to others with high academic achievement (Marsh, Trautwein, Lüdtke, & Köller, 2008). Therefore, it is possible that students may experience high levels of flow when they perceive the class to be in flow and they understand the material. However, it is possible that students may experience frustration when they perceive the class to be in flow, but they do not understand the material being taught. Additional research is needed in order to better understand these possibilities.

Strengths and Limitations

The current study has several strengths that are worth highlighting. One strength of our study involved the participants. Namely, by having all students in the class participate, concerns about a biased sample were of little concern with the current study. In addition, we were able to collect data from numerous sources to examine our hypotheses. Specifically, rather than being limited by data obtained from a single source (e.g., the students), we collected data from both students and the professor, as well as an objective indicator of classroom performance (i.e., the quiz scores).

Another clear strength of our study was our use of an event-based experience sampling method in order to study the experience of flow in the classroom. We were primarily interested in intra-individual processes and understanding whether flow was consistently associated with certain individual characteristics (interest in, and understanding of, the class material), perceptions (of instructor’s and classmates’ flow), and behaviors (performance on class quizzes) across a series of events. As such, our approach was an ideal one to answer the questions at hand. In addition, such an approach has some distinct advantages compared with the typical cross-sectional methods that have been used in the literature. For example, as Riediger (2009) noted, “In most questionnaires or interviews, respondents have to rely on partial recall and inference strategies when asked to report on their past behavior or experiences. There is ample empirical evidence that this results in retrospective memory biases and aggregation effects that impair the validity of the information assessed, sometimes profoundly so” (p. 3). In short, our approach allowed for an immediacy of measurement that helped limit retrospective memory biases. Furthermore, experience sampling allows for responses to occur within the natural environment, in this case the classroom. As such, we were able to study experiences within their ecological context (Schwarz, 2007), thereby enhancing the validity of our assessment. Finally, our study design also allowed us to evaluate the causal nature of flow of and academic performance, despite it still being correlational (cf., Xanthopoulou et al., 2012).

Despite the strengths of the current study, it is not without limitations, one of which is its limited external validity. We studied a sample of 14 students from the same class, with a single instructor. As noted earlier, although this sample size appeared to be sufficient for the parameters being examined, it remains modest. As such, the extent to which the findings will generalize to other classroom settings with other instructors is not known. One limitation of our study is the use of single-item measures to measure some constructs. However, it must be pointed out that single-item measures have been shown to (a) demonstrate good convergent and divergent validity as well as high correlations with multiple-item scales (Abdel-Khalek, 2006; Nagy, 2002), (b) indicate satisfactory test–retest reliability, and (c) be ideal when time is of a premium, as in the current study. Furthermore, single-item measures have been argued to be appropriate when the construct being measured is concrete rather than abstract and has high face validity (Rossiter, 2002). We would classify both understanding and interest as concrete constructs. Nonetheless, such single-item scales may not have adequately captured the multidimensional complexity of these variables by assuming their unidimensionality (Nagy, 2002). This may explain why we did not find significant correlations between understanding of, and interest in, the material covered in the class and subsequent quiz performance. It is recommended that future researchers use multi-item measures to ascertain if the findings of the current study are reliable.

Conclusion

One aim of this study was to identify antecedents to the flow experience in the classroom. Our findings suggest that both an understanding of and interest in the material are important facilitators of flow. A second aim was to ascertain whether flow was contagious and to determine the relationship between perceptions of other’s flow and one’s own self-report of flow. We found there was contagion in the classroom, whereby perceptions of classmates’ and instructor’s flow affected individual optimal experience. Our research has pushed past the usual focus of crossover research (e.g., negative emotions of partners in long-term relationships; Bakker et al., 2009) and indicated a crossover effect of positive experiences among students in a classroom.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Abdel-Khalek

A. M.

(2006). Measuring happiness with a single-item scale. Social Behavior and Personality, 34, 139-150.

Ahmed

van der Werf

Minnaert

Kuyper

(2010). Students’ daily emotions in the classroom: Intra-individual variability and appraisal correlates. British Journal of Educational Psychology, 80, 583-597.

Bakker

A. B.

(2005). Flow among music teachers and their students: The crossover of peak experiences. Journal of Vocational Behavior, 66, 26-44.

Bakker

A. B.

(2008). The work-related flow inventory: Construction and initial validation of the WOLF. Journal of Vocational Behavior, 72, 400-414. doi:10.1016/j.jvb.2007.11.007

Bakker

A. B.

Demerouti

Schaufeli

W. B.

(2003). The socially induced burnout model. Hauppauge, NY: Nova Science.

Bakker

A. B.

Demerouti

Schaufeli

W. B.

(2005). Crossover of burnout and work engagement among working couples. Human Relations, 58, 661-689.

Bakker

A. B.

Oerlemans

Demerouti

Slot

B. B.

Ali

D. K.

(2011) Flow and performance: A study among talented Dutch soccer players. Psychology of Sport and Exercise, 12, 442-450.

Bakker

A. B.

van Emmerik

Euwema

M. C.

(2006). Crossover of burnout and engagement in work teams. Work & Occupations, 33, 464-489.

Bakker

A. B.

Westman

Schaufeli

W. B.

(2007). Crossover of burnout: An experimental design. European Journal of Work and Organizational Psychology, 16, 220-239.

10.

Bakker

A. B.

Westman

van Emmerik

(2009). Advancements in crossover theory. Journal of Managerial Psychology, 24, 206-219.

11.

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations, Journal of Personality and Social Psychology, 51, 1173-1182.

12.

Barsade

(2002). The ripple effect, emotional contagion and its influence on group behavior. Administrative Science Quarterly, 47, 644-677.

13.

Biederman

Vessel

E. A.

(2006). Perceptual pleasure and the brain. American Scientist, 94, 249-255.

14.

Burnham

K. P.

Anderson

D. R.

(2004). Multimodal inference: Understanding AIC and BIC in model selection. Sociological Methods & Research, 33, 261-304.

15.

Ceja

Navarro

(2011). Dynamic patterns of flow in the workplace: Characterizing within-individual variability using a complexity science approach. Journal of Organizational Behavior, 32, 627-651.

16.

Chen

(2000). Exploring web users’ on-line optimal flow experiences (Unpublished doctoral dissertation). Syracuse University, Syracuse, NY.

17.

Choi

D. H.

Kim

S. H.

(2007). ERP training with a web-based electronic learning system. The flow theory perspective. International Journal of Human-Computer Studies, 65, 223-243.

18.

Cialdini

R. B.

(2009). Influence, science and practice. New York, NY: William Morrow.

19.

Csikszentmihalyi

(1975). Beyond boredom and anxiety, Experiencing flow in work and play. San Francisco, CA: Jossey-Bass.

20.

Csikszentmihalyi

(1990). Flow: The psychology of optimal experience. New York, NY: HarperCollins.

21.

Csikszentmihalyi

(1996). Creativity: Flow and the psychology of discovery and invention. New York, NY: HarperCollins.

22.

Csikszentmihalyi

(1997). Finding flow: The psychology of engagement with everyday life. New York, NY: Basic Books.

23.

Csikszentmihalyi

(2000). Flow. New York, NY: Oxford University Press.

24.

Csikszentmihalyi

Abuhamdeh

Nakamura

(2005). Flow. In Elliot

A. J.

Dweck

C. S.

(Eds.), Handbook of competence and motivation (pp. 598-608). New York, NY: Guilford Press.

25.

Csikszentmihalyi

LeFevre

(1989). Optimal experience in work and leisure. Journal of Personality and Social Psychology, 56, 815-822.

26.

Csikszentmihalyi

Rathunde

Whalen

(1993). Talented teenagers: The roots of success and failure. New York, NY: Cambridge University Press.

27.

Cui

Wong

M. L.

Zhang

(2008). Model selection for direct marketing: Performance criteria and validation methods. Marketing, Intelligence, & Planning, 26, 275-292.

28.

Culbertson

S. S.

Mills

M. J.

Fullagar

C. J.

(2012). Work engagement and work-family facilitation: Making homes happier through positive affective spillover. Human Relations, 65, 1151-1173.

29.

De Houwer

Barnes-Holmes

Moors

(2013). What is learning? On the nature and merits of a functional definition of learning. Psychonomic Bulletin & Review, 20, 631-642.

30.

Debnath

S. C.

Tandon

Pointer

L. V.

(2007). Designing business school courses to promote student motivation: An application of the job characteristics model. Journal of Management Education, 31, 812-831.

31.

Debus

M. E.

Sonnentag

Deutsch

Nussbeck

F. W.

(2014). Making flow happen: The effects of being recovered on work-related flow between and within days. Journal of Applied Psychology, 99, 713-722. doi:10.1037/a0035881

32.

Delle Fave

Massimini

(2005). The investigation of optimal experience and apathy: Developmental and psychosocial implications. European Psychologist, 10, 264-274.

33.

Delle Fave

Massimini

Bassi

(2011). Psychological selection and optimal experience across cultures: Social empowerment through personal growth. New York, NY: Springer Science + Business Media.

34.

Demerouti

(2006). Job characteristics, flow, and performance: The moderating role of conscientiousness. Journal of Occupational Health Psychology, 11, 266-280.

35.

Ebbinghaus

(1964). Memory: A contribution to experimental psychology ( Ruger

H. A.

Bussenius

C. E.

, Trans.). Oxford, England: Dover. (Original work published 1885)

36.

Eisenberger

Jones

J. R.

Stinglhamber

Shanock

Randall

A. T.

(2005). Flow experiences at work, for high need achievers alone? Journal of Organizational Behavior, 26, 755-775.

37.

Festinger

(1954). A theory of social comparison processes. Human Relations, 7, 117-140.

38.

Fredrickson

B. L.

(1998). What good are positive emotions ? Review of General Psychology, 2, 300-319.

39.

Fredrickson

B. L.

(2001). The role of positive emotions in positive psychology: The broaden and build theory. American Psychologist, 56, 218-226.

40.

Fullagar

C. J.

Kelloway

E. K.

(2009). Flow at work: An experience sampling approach. Journal of Occupational and Organizational Psychology, 82, 595-615.

41.

Fullagar

C. J.

van Ittersum

K. W.

Knight

P. A.

(2012, April). Flow and well-being: An experimental approach. Paper presented at the 27th annual conference of the Society for Industrial and Organizational Psychology. San Diego, CA.

42.

Guadagno

R. E.

Muscanell

N. L.

Rice

L. M.

Roberts

(2013). Social influence online: The impact of social validation and likability on compliance. Psychology of Popular Media Culture, 2, 51-60.

43.

Guo

Klein

Rossin

(2007). The impact of flow on learning outcomes in a graduate-level information management course. Journal of Global Business Issues, 1(2), 31-39.

44.

Guo

Y. M.

Poole

M. S.

(2009) Antecedents of flow in online shopping: A test of alternative models. Information Systems Journal, 19, 369-390.

45.

Hatfield

Cacioppo

J. T.

Rapson

R. L.

(1994). Emotional contagion. New York, NY: Cambridge University Press.

46.

Heine

(1996). Flow and achievement in mathematics (Unpublished doctoral dissertation). University of Chicago, Chicago, IL.

47.

Hektner

Asakawa

(2000). Learning to like challenges. In Csikszentmihalyi

Schneider

(Eds.), Becoming adult: How teenagers prepare for the world of work (pp. 95-112). New York, NY: Basic Books.

48.

House

E. D.

Arruda

J. E.

Andrasik

Grazzi

(2012). The reliability and validity of the visual analog moods scales in non-English-speaking pain patients. Pain Practice, 12, 626-632.

49.

Hofmann

D. A.

Gavin

M. B.

(1998). Centering decisions in hierarchical linear models, Implications for research in organizations. Journal of Management, 24, 623-641.

50.

Ilies

Schwind

K. M.

Wagner

D. T.

Johnson

M. D.

DeRue

D. S.

Ilgen

D. R.

(2007). When can employees have a family life? The effects of daily workload and affect on work-family conflict and social behaviors at home. Journal of Applied Psychology, 92, 1368-1379.

51.

Jackson

S. A.

Csikszentmihalyi

(1999). Flow in sports: The keys to optimal experiences and performances. Champaign, IL, Human Kinetics.

52.

Judge

T. A.

Ilies

Scott

B. A.

(2006). Work-family conflict and emotions: Effects at work and at home. Personnel Psychology, 59, 779-814.

53.

Keller

Bless

(2008). Flow and regulatory compatibility: An experimental approach to the flow model of intrinsic motivation. Personality and Social Psychology Bulletin, 34, 196-209.

54.

Keller

Blomann

(2008). Locus of control and the flow experience: An experimental analysis. European Journal of Personality, 22, 589-607.

55.

Konradt

Sulz

(2001). The experience of flow in interacting with a hypermedia learning environment. Journal of Educational Multimedia and Hypermedia, 10, 69-84.

56.

Larson

Csikszentmihalyi

(1983). The experience sampling method. New Directions for Methodology of Social and Behavioral Science, 15, 41-56.

57.

Maas

C. J. M.

Hox

J. J.

(2005). Sufficient sample sizes for multilevel modeling. Methodology, 1, 86-92.

58.

Marsh

H. W.

Trautwein

Lüdtke

Köller

(2008). Social comparison and big-fish–little-pond effects on self-concept and other self-belief constructs: Role of generalized and specific others. Journal of Educational Psychology, 100, 510-524.

59.

McEvoy

G. M.

(2011). Increasing intrinsic motivation to learn in organizational behavior classes. Journal of Management Education, 35, 468-503.

60.

Mendelson

(2007). The functional mediation of flow between achievement anxiety, academic procrastination, and academic performance (Unpublished doctoral dissertation). Fordham University, New York, NY.

61.

Nagy

M. S.

(2002). Using a single-item approach to measure job satisfaction. Journal of Occupational and Organizational Psychology, 75, 77-86.

62.

Nakamura

(1988). Optimal experience and the uses of talent. In Csikszentmihalyi

Csikszentmihalyi

(Eds.), Optimal experience, Psychological studies of flow in consciousness (pp. 319-326). New York, NY: Cambridge University Press.

63.

Nakamura

Csikszentmihalyi

(2005). Engagement in a profession: The case of undergraduate teaching. Daedalus: Journal of the American Academy of Arts and Sciences, 134(3), 60-67.

64.

Nakamura

Csikszentmihalyi

(2009). Flow theory and research. In Snyder

C. R.

Lopez

S. J.

(Eds.), The Oxford handbook of positive psychology (2nd ed., pp. 195-206). New York, NY: Oxford University Press.

65.

Nezlek

J. B.

(2003). Using multilevel random coefficient modeling to analyze social interaction diary data. Journal of Social and Personal Relationships, 20, 437-468.

66.

Nezlek

J. B.

(2012). Multilevel modeling for psychologists. In Cooper

(Ed.), Handbook of research methods in psychology (pp. 219-241). Washington, DC, American Psychological Association.

67.

Nielsen

Cleal

(2010). Predicting flow at work: Investigating the activities and job characteristics that predict flow states at work. Journal of Occupational Health Psychology, 15, 180-190. doi:10.1037/a0018893

68.

Niemiec

C. P.

Ryan

R. M.

(2009). Autonomy, competence, and relatedness in the classroom: Applying self-determination theory to educational practice. Theory and Research in Education, 7, 133-144.

69.

Perry

S. K.

(1999). Writing in flow. Cincinnati, OH: Writer’s Digest Books.

70.

Pert

C. B.

Snyder

(1973). Opiate receptors, Demonstration in nervous tissue. Science, 179, 1011-1014.

71.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models (2nd ed.). Thousand Oaks, CA, Sage.

72.

Raudenbush

Bryk

Congdon

(2000). HLM 6 Hierarchical linear and nonlinear modeling. Skokie, IL, Scientific Software International.

73.

Reis

H. T.

Wheeler

(1991). Studying social interaction with the Rochester Interaction Record. In Zanna

M. P.

(Ed.), Advances in experimental social psychology (Vol. 24, pp. 270-318). San Diego, CA: Academic Press.

74.

Riediger

(2009). Experience sampling (Working Paper No. 62). Retrieved http://library.mpib-berlin.mpg.de/ft/mr/MR_Experience_2009.pdf

75.

Rossin

Klein

Guo

(2009). The effects of flow on learning outcomes in an online information management course. Journal of Information Systems Education, 20, 87-98.

76.

Rossiter

J. R.

(2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19, 305-335.

77.

Ryu

Parsons

(2012). Risky business or sharing the load? Social flow in collaborative mobile learning. Computers & Education, 58, 707-720.

78.

Sawyer

(1992). Improvisational creativity, An analysis of jazz performance. Creativity Research Journal, 5, 253-263.

79.

Schiefele

Csikszentmihalyi

(1994). Interest and the quality of experience in classrooms. European Journal of Psychology of Education, 9, 251-269.

80.

Schüler

(2007). Arousal of flow experience in a learning setting and its effects on exam performance and affect. German Journal of Educational Psychology, 21, 217-227.

81.

Schwarz

(2007). Retrospective and concurrent self-reports: The rationale for real-time data capture. In Stone

A. A.

Shiffman

Atienza

A. A.

Nebeling

, (Eds.), The science of real-time data capture: Self-reports in health research (pp. 11-26). New York, NY: Oxford University Press.

82.

Schweinle

Meyer

D. K.

Turner

J. C.

(2006). Striking the right balance: Students’ motivation and affect in elementary mathematics. Journal of Educational Research, 99, 271-293.

83.

Shernoff

D. J.

Csikzentmihalyi

(2009). Flow in schools, Cultivating engaged learners and optimal learning environments. In Gilman

R. G.

Huebner

E. S.

Furlong

M. J.

(Eds.), Handbook of positive psychology in schools (pp. 131-145). New York, NY: Routledge.

84.

Shernoff

D. J.

Csikszentmihalyi

Schneider

Shernoff

E. S.

(2003). Student engagement in high school classrooms from the perspective of flow theory. School Psychology Quarterly, 18, 158-176.

85.

Shernoff

D. J.

Hoogstra

(2001). Continuing motivation beyond the high school classroom. San Francisco, CA: Jossey-Bass.

86.

Shernoff

D. J.

Schmidt

J. A.

(2008). Further evidence of an engagement-achievement paradox among U.S. high school students. Journal of Youth and Adolescence, 37, 564-580.

87.

Stephanou

(2011). Students’ classroom emotions: Socio-cognitive antecedents and school performance. Electronic Journal of Research in Educational Psychology, 9, 5-47.

88.

Stern

R. A.

Arruda

J. E.

Hooper

C. R.

Wolfer

G. D.

Money

C. E.

(1997). Visual analog mood scales to measure internal mood state in neurologically impaired patients, description and initial evidence. Aphasiology, 11, 59-71.

89.

Tesser

(1988). Toward a self-evaluation maintenance model of social behavior. San Diego, CA: Academic Press.

90.

Thierne

R. J.

Song

Calantone

R. J.

(2000). Artificial neural network decision support systems for new product development project selection. Journal of Marketing Research, 37, 499-507.

91.

Totterdell

Kellett

Teuchmann

Briner

R. B.

(1998). Evidence of mood linkage in work groups. Journal of Personality and Social Psychology, 74, 1504-1515.

92.

Ullén

de Manzano

Ö.

Almeida

Magnusson

P. K. E.

Pedersen

N. L.

Nakamura

. . . Madison

(2012). Proneness for psychological flow in everyday life: Associations with personality and intelligence. Personality and Individual Differences, 52, 167-172.

93.

Walker

C. J.

(2010). Experiencing flow: Is doing it together better than doing it alone? Journal of Positive Psychology, 5, 3-11.

94.

Westman

(2011). The impact of stress on the individual, the dyad, and the team. Stress and Health, 27, 177-180.

95.

Westman

Etzion

Chen

(2009). Crossover of positive experiences from business travelers to their spouses. Journal of Managerial Psychology, 24, 269-284.

96.

Westman

Shadach

Keinan

(2013). The crossover of positive and negative emotions: The role of state empathy. International Journal of Stress Management, 20, 116-133.

97.

Wong

M. M.

Csikszentmihalyi

(1991). Motivation and academic achievement: The effects of personality traits and the quality of experience. Journal of Personality, 59, 539-574.

98.

Xanthopoulou

Bakker

A. B.

Ilies

(2012). Everyday working life: Explaining within-person fluctuations in employee well-being. Human Relations, 65, 1051-1069.