Using an Online-Based Mindfulness Intervention to Reduce Test Anxiety in Physics Students

Abstract

Using an 8-week mindfulness intervention, we investigated the mechanisms of mindfulness to address test anxiety in introductory physics II. Our goal was to explore the effectiveness of using an 8-week online mindfulness intervention to address student test anxiety. We used self-report measures to assess participants on mindfulness, emotional regulation, attentional regulation, and physics quiz scores at three-time points—time 1 (pre), time 2 (mid), and time 3. Participants in the mindfulness condition received 8 weeks of mindfulness training while the control group was used as a baseline for comparison with no mindfulness training. The 8-week online-based mindfulness intervention resulted in non-significant changes in test anxiety and quiz scores for the treatment compared to the control. Additionally, results from multiple regression modeling suggest that emotional regulation positively predicted mindfulness at time 3 (T3) of data collection, whereas attentional regulation negatively predicted mindfulness at T3. These findings have both theoretical and practical implications. Our findings suggest more research is needed to investigate the usefulness of mindfulness-based interventions in STEM-related settings.

Mindfulness and Test Anxiety

Problem Statement

Experiencing anxiety is a normal emotional experience. Test anxiety occurs when students experience high levels of stress, nervousness, or apprehension related to testing situations that are accompanied by a combination of physical, behavioral, and affective symptoms (Salend, 2010). Examples of physical symptoms from test anxiety include nausea and rapid heartbeat. Examples of behavioral symptoms from test anxiety include difficulty maintaining attention, pretending to be ill, or missing an exam entirely. Examples of affective symptoms of test anxiety include a lack of motivation and making negative comparisons to other students. Chronic test anxiety is associated with poor educational outcomes, reduced self-esteem, reduced effort, and increased apathy to complete schoolwork (Huberty, 2009).

Students at all levels of education are subject to assessment and that does not appear to be changing any time soon. While test anxiety is a major problem affecting many students, the good news is that there are ways to combat test anxiety. Teaching students the skills of mindfulness could help address this problem. The goal of this intervention study is to explore the theoretical relationship between mindfulness and test anxiety. This will be accomplished by creating and implementing an 8-week mindfulness-based intervention that will be primarily delivered in an online delivery format. The goal of this mindfulness-based intervention is to empower college students to be successful in academics by reducing their experience of test anxiety. The research will fill a gap in the research literature on mindfulness by using an online-based mindfulness intervention to specifically target college student’s test anxiety. Previous research by Cho and colleagues (Cho et al., 2016) used a single-week mindfulness self-guided intervention to reduce test anxiety with positive results; however, they only worked with students who registered high test anxiety and did not measure mindfulness directly, but instead measured the construct of reappraisal. We start by defining the constructs relevant to the proposal including mindfulness and test anxiety.

What Is Mindfulness?

To begin, Bishop and colleagues (Bishop et al., 2004) define mindfulness as a two-component construct including self-regulation of attention toward the present and an open curiosity toward what happens in the present. Being mindful means directing attention on the immediate cognitive experience while maintaining an orientation of openness, curiosity, and acceptance to experiences arising in the present (Bishop et al, 2004).

The first component of Bishop and colleague’s (Bishop et al., 2004) definition of mindfulness is attention regulation in the present. When an individual intentionally directs their attention to the present without trying to change that moment, their present focus allows them to slow down to recognize their default cognition in action. Research by Flook and colleagues (Flook et al., 2010) showed that mindfulness-based interventions (MBI) had the strongest impact for students who scored lowest on executive function assessments at mindfulness training. This suggests that students who have the most difficulty maintaining their attention would benefit the most from MBIs. This is important in educational contexts because students who can maintain their attention on an instructional task are more likely to be successful (Posner & Patoine, 2009).

The second component of Bishop and colleague’s (Bishop et al., 2004) definition of mindfulness is emotional regulation, specifically being open to all emotional experiences with the awareness that emotions are not permanent. Research indicates that emotional regulation may be an important subcomponent when operationalizing mindfulness (Schonert-Reichl & Lawlor, 2010; Heppner et al., 2015). Noticing changes in emotions through mindfulness makes it easier for individuals to regulate their emotions. For example, individuals practicing mindfulness may initially experience an unpleasant emotion like anger. Practicing mindfulness allows individuals to be aware that they are angry in a way that creates distance from the emotion. Research by Maloney and colleagues (Maloney et al., 2016) shows that MBIs can teach students a variety of social-emotional skills useful for success in and outside of school. Furthermore, research by Schonert-Reichl and Lawlor (2010) suggests that MBIs can be an effective method for regulating academic emotions and generating positive activating emotions (Pekrun & Linnenbrink-Garcia, 2012). College students, as a specific population, may also benefit from practicing mindfulness (Chiang & Sumell, 2019; Ramsburg & Youmans, 2014; Sumell et al., 2021). This is important in classroom contexts because students are likely to encounter experiences that could be emotionally aversive like testing, but with mindfulness students can reappraise their initial emotional reaction into something more conducive for learning. Now that mindfulness is defined, we define the second variable of interest—test anxiety.

What Is Test Anxiety?

Test anxiety, defined as excessive anxiety when preparing for or taking an exam, goes above normal levels of anxiety (Cassady, 2004). Higher levels of test anxiety prior to an exam are negatively correlated with exam scores (Barrows et al., 2013; Chapell et al, 2005). Test anxiety is thought to be experienced by a third of college students (Huberty, 2009) but numbers may be higher for students with learning disabilities (Peleg, 2009), ethnic minorities (Carter et al., 2008), and female students (Gerwing et al, 2015). Test anxiety varies across student experience. Research by Dawood and colleagues (Dawood et al., 2016) found a negative correlation with test anxiety and academic level. This implies that more experienced students have figured out strategies to cope with test anxiety or perhaps many students with high test anxiety do not attend (or drop out of) college.

While we have not encountered literature that says physics courses elicit more test anxiety there is a reasonable case that introductory physics courses can produce high test anxiety. Dawood and colleagues (Dawood et al., 2016) found a negative correlation with test anxiety and students’ academic level so students in entry-level courses may be more prone to experiencing test anxiety. Introductory physics is required in science, technology, engineering, and math (STEM courses) related majors and serves as a gatekeeper for future STEM courses. This is problematic for STEM majors because physics courses are associated with a large amount of attrition (Sithole et al, 2017). Test anxiety may be an unnecessary barrier to future courses in STEM programs and career.

The Relationship Between Mindfulness and Test Anxiety

Mindfulness training could address the problem of test anxiety. Research by Cho and colleagues (Cho et al., 2016) used a 1-week self-taught mindfulness intervention to reduce test anxiety for college students. A common symptom of test anxiety is the inability to maintain attention during testing (Salend, 2010). Mindfulness could help students by making it easier to recognize when their attention has wandered to anxious thoughts (Lutz et al., 2008; Rahl et al, 2017). Individuals who are mindful will have a greater ability to regulate their attention to the present. Attentional regulation is defined as the ability to orient attention to external objects and to manage streams of thought (Posner & Dehaene, 1994). Mindfulness is like an anchor to remain grounded in the present. Mindfulness training may also benefit students by regulating intense emotions in testing situations. Given that emotional regulation is defined as the ability to exert control over the emotions an individual experience (Gross, 2015), individuals who are mindful may gain skills they can use when feeling anxious that can bring physiological relief and help to see the impermanent nature of emotions (Sayers et al., 2015; Chancey et al., 2021). Practicing mindfulness allows individuals to create distance from emotions as they happen resulting in decreased emotional distress (Kiken & Shook, 2012).

The aim of this study was twofold: (1) explore the theoretical relationship between the constructs of mindfulness and test anxiety and (2) explore the impact of an online mindfulness intervention on the test anxiety of college students enrolled in an introductory physics class. Investigating mindfulness as a concept is on the rise in psychological science literature but remains difficult to define.

Research Questions and Hypotheses

We developed the following research question to explore the theoretical relationship between mindfulness, attentional regulation, and emotional regulation:

• What is the predictive relationship between emotional regulation and attentional regulation on mindfulness?

Additionally, we developed the following research question to explore the influence of mindfulness training on test anxiety with college physics students:

• Will an internet-based mindfulness intervention facilitate a reduction in test anxiety more than a control?

Previous research shows that participating in MBIs can have positive effects on emotional regulation (Maloney et al., 2016) and attention toward the present (Jha et al., 2007). Thus, we hypothesize that both attentional and emotional regulation will serve as positive predictors of mindfulness for students in our study. We predict that students who practice mindfulness will have an easier time managing intense emotional experiences like test anxiety. Therefore, we hypothesize that students who are more mindful will be better able to cope with test anxiety in introductory physics because they have a set of skills to regulate those intense emotional experiences (Schonert-Reichl & Lawlor, 2010).

Method

Participants and Context

Participants were 153 undergraduate students from a large public university in the American Midwest recruited from a single section of introductory physics II. The first author of this study addressed the introductory physics class with a recruitment script encouraging participants to volunteer for the study. Participants were informed by their physics professor that up to 20 points of extra credit would be added to their grades for participating. Participants were offered two opportunities to participate in the study. Initially, all participants were asked to complete assessments on the variables of test anxiety, emotional regulation, attentional regulation, mindfulness, and quiz grades at three different times of the semester (Weeks 3,8,13) for 10 extra credit points in their physics class. Participants had the opportunity to earn an additional 10 extra credit points by participating in an 8-week mindfulness-based intervention intended to address test anxiety, for a total of 20 points. Both groups were given extra credit; however, those who chose to only complete the surveys were in our control condition and earned 10 points extra credit while those who chose to earn 20 points of extra credit were the quasi-experimental group. The extra credit points were put toward the 2400 total points offered in this class. There were no exclusion criteria for the intervention. Students from the physics II class were allowed to self-select if they wished to participate in the intervention portion of this study. A total of 60 participants were self-selected to be in the intervention portion of this study, with 93 students serving as the control group.

Instruments

Test Anxiety

To assess test anxiety, participants were given the cognitive test anxiety scale, a 27-item, four-point Likert scale instrument (Cassady & Johnson, 2002) at three different points in time during the 8 weeks of this study (T1 = week 0, T2 = week 4, T3 = week 8). The scale is anchored by (1) not at all typical of me and (4) very typical of me. Items were modified to reflect the context of a physics test. An example item was “My mind goes blank when I am pressured for an answer on a physics test.” The authors found evidence of concurrent validity by comparing the cognitive test anxiety scale to the measurement of reactions to tests. The two constructs were found to be significantly correlated to one another (Cassady & Johnson, 2002). We used Cronbach’s Alpha to evaluate the reliability of this scale at each time point, T1 = 0.94, T2 = 0.78, T3 = 0.92.

Emotional Regulation

To assess emotional regulation, participants were given the exam emotion regulation questionnaire, a six-item, 10-point Likert scale instrument (Levine et al., 2012). The items are anchored with (1) not at all and (10) extremely. Participants were asked to think about an upcoming exam and reflect on the type of strategies they would use to cope. An example item includes “I try to see the positive of this experience.” We used Cronbach’s Alpha to evaluate the reliability of this scale at each time point, T1 = 0.51, T2 = 0.57, T3 = 0.58.

Attentional Regulation

To assess attentional regulation, we used the self-regulation scale, a ten-item, four-point Likert scale assessment of self-regulation (Diehl et al., 2006). Items on this instrument are anchored by “Not at all true” (1) and “Completely true.” Items on this scale were modified to fit the context of a physics exam. An example item is “I can control my thoughts from distracting me from a physics exam.” The self-regulation scale was found to be a reliable instrument by Diehl and colleagues (Diehl et al., 2006) with Cronbach’s alpha scores ranging from 0.76 to 0.84 across three waves of assessment. We used Cronbach’s Alpha to evaluate the reliability of this scale at each time point, T1 = 0.85, T2 = 0.40, T3 = 0.86.

Mindfulness

To measure mindfulness, we used the Philadelphia Mindfulness Scale (PHLMS), a 20-item, five-point Likert scale instrument (Cardaciotto et al., 2008). Items on this instrument are anchored by “never” and “very often” with “sometimes” (3) in the center. The PHLMS assesses awareness and acceptance. An example item of awareness is “I am aware of what thoughts are passing through my mind.” An example item of acceptance is “I wish I could control my emotions more easily.” We used Cronbach’s Alpha to evaluate the reliability of this scale at each time point T1 = 0.81, T2 = 0.79, T3 = 0.83. In contrast to trait-based mindfulness measures, which conceptualize mindfulness as a relatively stable, enduring characteristic of a person, the PHLMS approaches the measurement of state mindfulness as a reflection on recent experience with mindfulness practice. Because we were interested in detecting changes in mindfulness over time, the PHLMS was the appropriate mindfulness measure for that context.

Physics Quiz Scores

We used quiz scores at T1, T2, and T3 to explore the impact of test anxiety on academic performance. The T1 quiz included twelve multiple-choice items about electric fields and dipoles. At T2, participants completed a group discussion problem on cyclotron motion and the mass spectrometer. The group discussion problem included a single physics problem that students were asked to solve in groups of three under the guidance of a course teaching assistant (TA). At T3, participants completed an in-class eight-item multiple choice assessment over Faraday’s Law and inductance completed on Zoom because the class transitioned to online due to the COVID-19 outbreak.

Pre-Data Manipulation Check

We asked participants at T1 if they would like to participate in the intervention portion of our study, if they had experience practicing mindfulness, and to evaluate their motivation to reduce test anxiety.

Intervention Manipulation Check

To assess the frequency of participation, we asked participants once a week using the Remind app how often they practiced mindfulness using the Insight timer app out of the last 7 days.

Procedures

Our 8-week online mindfulness intervention was based on an MBI used by Morledge and colleagues (Morledge et al., 2013) to target participant stress. The authors used themes, articles, and daily tips each week to help participants develop their own mindfulness practice. Our intervention targeted a specific psychological outcome in test anxiety (Lippmann et al, 2019) and included opportunities to practice mindfulness formally and informally (Germer, 2004). Researchers were given access by the instructor of a single section of the introductory physics course. In a brief announcement at the beginning of class, participants were informed by the first author that they could earn extra credit by participating and were given a link to sign up for a single in-person session to learn about mindfulness. We used T1 data to identify participants for the intervention. The intervention began with a focus on two types of mindfulness—focused attention and open-monitoring (Lutz et al., 2008). We read out loud a short children’s book on mindfulness titled “King Calm” (Sweet & Miles, 2016) to illustrate how to practice mindfulness in their everyday lives. Participants were shown how to download the insight timer and remind mobile apps. Participants were informed they would receive daily reminders, articles, and tips from the Remind app. Participants were shown how to use the Insight Timer app by practicing mindfulness for 5 min. This initial session lasted for approximately 1 h.

Remind notifications were sent to treatment participants five times per week for 8 weeks between T1 and T3 data collection. Participants were encouraged to practice mindfulness for at least 5 min daily, especially when strong emotions arise like test anxiety. On Monday’s participants were sent a short video with a theme for what to focus their mindfulness training for the week. Themes included (1) focused attention, (2) open monitoring, (3) mindful eating, (4) mindful walking, (5) mindful sitting, (6) mindful compassion, (7) mindful relaxing, and (8) mindful movement. Each theme showed participants formal and informal practices of mindfulness to use in their everyday lives (Germer, 2004). Participants were sent daily tips via Remind on how to practice mindfulness throughout the week that came from a series of short readings by prominent mindfulness instructor Thich Nhat Hanh (2016). Our intervention thus contained all four instructional design components, which are typically found in successful Internet-based mindfulness interventions (i.e., formal learning tasks, supportive information, part-task practice, and just-in-time information; Lippmann et al, 2019). On Friday’s participants were asked to self-report their participation from the previous week. In total, participants in our treatment condition were asked to participate for approximately 4 h, assuming they followed each of the Remind suggestions.

All participants were given a link to complete data collection materials at the midway point and conclusion of the intervention. Participants in the control condition did not receive mindfulness instruction, nor were they asked to download any mobile phone apps. The control group completed all T1, T2, and T3 measures as a baseline for comparison with the treatment group to isolate the effects of participating in a mindfulness intervention on test anxiety. Procedures are depicted in a flow chart below (See Fig. 1).

Fig. 1.

Flow chart of procedures

Results

Demographic Data

All data was analyzed using SPSS version 26.0. We chose to analyze T3 data because this data was collected after our mindfulness intervention concluded. Of the 153 students, 84 (54.9%) identified as male, 52 (33.9%) as female, and 17 (11.1%) did not specify their gender. The average participant age was 20.51 years. A total of 73 (45.6%) participants in this study self-reported as white/Caucasian, 22 (13.8%) identified as Asian, followed by nine (5.6%) students who identified as Hispanic/Latinx, eight (5%) students as belonging to more than one race/ethnicity, seven (4.3%) students identified as Black/Native American, seven (4.3%) students identified as Native American, and two students identified as Arabic. The rest of our participants did not disclose their race/ethnicity. The largest majors were mechanical engineering (10.6%), civil engineering (8.1%), chemical engineering (8.1%), and meteorology (7.5%).

Participants were also assessed on their previous experience practicing mindfulness. In this study, 62 (40.6%) responded with “Yes” while 91 participants (59.33%) responded with “No.” Additionally, participants were asked to evaluate their own motivation to reduce test anxiety on a five-point Likert scale with 1 representing “Not at all,” 3 representing an anchor point of “A moderate amount,” and 5 representing “A Great Deal.” Only 12 self-reported a score of 1, indicating that most of the participants have experienced or are currently experiencing test anxiety to some degree, thus providing evidence for the existence of test anxiety in our sample.

RQ 1

To answer our first research question—What is the predictive relationship between emotional regulation and attentional regulation on mindfulness?—we ran a multiple linear regression using T3 scores for attentional regulation and emotional regulation while using previous mindfulness experience as a covariate to predict T3 mindfulness scores. We chose to use T3 scores as the primary data for analysis to explore the degree to which our intervention influenced our results. Table 1 shows the means, standard deviations, skewness, and kurtosis for treatment and control on the cognitive test anxiety scale, exam emotion regulation questionnaire, self-regulation scale, Philadelphia Mindfulness Scale, and quiz scores at time one. All skewness and kurtosis values were less than or equal to an absolute value of three, indicating that the assumption normality can be used for the remainder of my analyses (Lomax & Hahs-Vaughn, 2012). Additionally, Table 2 shows the means, standard deviations, skewness, and kurtosis for treatment and control on the cognitive test anxiety scale, exam emotion regulation questionnaire, self-regulation scale, Philadelphia Mindfulness Scale, and quiz scores at time three. Correlations for each of the measured variables at T1 can be found in Table 3.

Table 1.

Descriptive Statistics for T1. Descriptive statistics for scores on the cognitive test anxiety scale, exam emotion regulation questionnaire, self-regulation scale, Philadelphia Mindfulness Scale, and quiz scores at time one

	Treatment (N = 60)				Control (N = 93)
	Mean	SD	Skewness	Kurtosis	Mean	SD	Skewness	Kurtosis
Test Anxiety	75.0333	16.60301	− .288	− .832	69.9333	17.0338	.031	− .826
Emotional Regulation (Total)	36.1667	7.70025	.489	− .133	35.8495	6.75172	.243	.174
Philadelphia Mindfulness Awareness	37.2931	5.68781	− .172	− .213	35.4382	5.21134	− .200	.649
Philadelphia Mindfulness Acceptance	35.5932	8.67441	− .245	− .647	33.9444	6.79451	.173	− .509
Attention Regulation	27.5333	5.29940	.094	− .370	28.4667	5.12278	− .248	− .311
Quiz Grades	82.6667	19.12197	− .757	− .608	77.0115	2.44214	− .485	− .991

Table 2.

Descriptive statistics for scores on the cognitive test anxiety scale, exam emotion regulation questionnaire, self-regulation scale, Philadelphia Mindfulness Scale, and quiz scores at T3

	Treatment (N  = 60)				Control (N = 93)
	Mean	SD	Skewness	Kurtosis	Mean	SD	Skewness	Kurtosis
Test Anxiety	68.65	14.55	0.253	− 0.759	70.34	14.80	− 0.44	− 0.729
Emotional Regulation (Total)	37.65	7.72	− .004	− 0.811	37.65	6.79	− 0.154	0.070
Philadelphia Mindfulness Awareness	36.87	6.07	0.159	− 0.534	35.24	5.74	− 0.116	− 0.394
Philadelphia Mindfulness Acceptance	35.32	7.93	0.024	− 0.994	33.50	6.55	0.313	0.159
Attention Regulation	28.08	5.43	0.374	.0520	28.25	5.00	− .053	0.117
Quiz Grades	82.39	15.50	− .0868	0.538	82.95	17.49	− 2.061	6.926

Table 3.

Correlation data from T1 data

	Test anxiety	Grades	Aware	Accept Emo Reg Atten Reg
Test anxiety	1	− .361**	.063	.430** − .077 − .684**
Grades	− .361**	1	.072	.060 − .020 .135
Awareness	.063	.072	1	.131 .105 .152
Acceptance	.430**	.060	.131	1 .105. − .482**
Emotional Reg	− .077	− .020	.105	.202 1 .159
Attentional Reg	− .684**	.135	.152	− .482** .159 1

** = p > .001 (2-tailed)

RQ 2

In addition to gaining a better understanding of the mechanisms of mindfulness in classrooms, we also explored the effectiveness of using an online mindfulness-based intervention to address the phenomena of test anxiety. We attempted to answer the following research question—Will an internet-based mindfulness intervention facilitate a reduction in test anxiety more than a control? Differences between the two conditions at T1 were investigated by using an independent sample t-test on each of the measures at T1. All T1 scores did not differ significantly between the two conditions, except for scores on the awareness scale of the Philadelphia mindfulness instrument where the treatment group scored significantly higher than the control. Overall, these results suggest that it can be assumed that there were no significant differences between conditions that may skew results except for the awareness variable.

The intervention portion of the study was to test (a) whether participants would exhibit significant change over time in test anxiety, emotional regulation, attentional regulation, and grades and (b) whether any observed changes might vary as a function of treatment condition. We performed a series of 3 (time) X 2 (treatment condition; tx group versus control group) repeated measures ANOVAs with each of the above-mentioned variables as outcome variables. Treatment participants were asked to self-report the number of days they practiced mindfulness each week, from zero to seven. The self-reported average number of practice days per week was 3.89 days, slightly more than half. This suggests that participants engaged in the intervention when prompted often.

Both conditions showed changes in test anxiety from T1 to T3. The treatment condition experienced a decrease in scores on the cognitive test anxiety assessment compared to the control group; however, the results were not significant. Additionally, both conditions experienced a significant change in their quiz grades from T1 to T3. The treatment condition experienced a large change from T1 (83.39) to T3 (90.21), but those results returned close to T1 averages at T3 (83.01). The control condition demonstrated an overall increase in assessment scores from T1 (77.67) to T3 (83.04). Score increases represent the percentage of correct answers on each quiz. However, because each quiz assessed different topics of physics, these gains do not necessarily indicate an increase in their absolute knowledge of introductory physics II content.

There was no statistically significant difference with the interaction with mindfulness training on quiz score grades from T1 to T3. The control group showed a modest decline in attentional regulation scores from T1 (28.39) to T3 (27.64) while the treatment group showed a modest increase in attentional regulation scores from T1 (27.80) to T3 (28.12). There were no significant differences in interaction between group, attentional regulation, and time.

Similar results occurred with emotional regulation scores. Both conditions showed a modest increase in emotional regulation scores with the control increasing from 35.78 (T1) to 36.42 (T3) while the treatment group increased in emotional regulation scores from 36.44 (T1) to 38.39 (T3). The group x emotional regulation x time interaction, however, was not statistically significant.

Multiple Regression Findings

Here, we report the results of multiple regression using T3 data. Results of a multiple regression indicated that there was a collective significant effect between emotional regulation at T3 and attentional regulation at T3. Both variables significantly predicted mindfulness at T3, (F (3, 492.119) = 5.849., p < 0.001, R^2 = 0.123). The individual predictors were examined further and indicated that emotional regulation (t = 3.029, p = 0.003) and attentional regulation (t =  − 3.016, p = 0.000) were significant predictors in the model. Previous mindfulness experience, however, was not a significant predictor of mindfulness at T3, (t =  − 1.469, p = 0.144). T3 Emotional regulation scores significantly predicted T3 mindfulness scores over and above attentional regulation at T3.

Discussion

Summary of Findings

Our goals were (1) to explore the predictive relationships between the constructs of emotional regulation and attentional regulation on mindfulness and (2) to examine the impact of an online MBI with introductory physics students. Results from multiple regression analyses suggest that the emotional regulation component of mindfulness was a strong, significant predictor mindfulness at T3. Attentional regulation, however, was a significant, negative predictor of mindfulness at T3. Additionally, results from repeated measures ANOVA revealed that our MBI did not have a significant effect on reducing test anxiety or improving quiz scores. The results of this study have theoretical and practical implications, which are discussed below.

Theoretical Implications

Our findings add to the literature on how mindfulness should be theoretically conceptualized when developing interventions. Bishop and colleagues (Bishop et al., 2004) define mindfulness in terms of self-regulation of attention toward the present and an openness toward emotional experiences happening in the present. We used a two-factor state-based instrument of mindfulness with the Philadelphia Mindfulness Scale (Cardaciotto et al., 2008) to explore the underlying mechanisms of mindfulness in educational settings. Results from multiple regression modeling showed that emotional regulation at T3 was a significant, positive predictor of mindfulness T3. These results are not surprising given previous research linking mindfulness with positive outcomes in emotional regulation (Sayers et al., 2015). Those who are more able to maintain positive or neutral emotions in the present could use these skills before, during, or after examination to be more mindful. Previous research has shown that mindfulness can have positive impacts on emotional experiences and well-being (Maloney et al., 2016; Kiken & Shook, 2012; Menezes & Bizarro, 2015).

What was surprising from our results, however, was that attentional regulation at T3 played a significant, but also negative role in the prediction of mindfulness at T3. We hypothesized that the variable of attentional regulation would have a positive impact on mindfulness because practicing mindfulness through focused attention meditation (Lutz et al., 2008) can aid students by making it easier to recognize when their attention has wandered away while also providing an anchor to come back to the present (Jankowski & Holas, 2020; Lutz et al., 2008; Rahl et al, 2017), thus potentially improving academic outcomes in the form of reduced test anxiety and increased quiz grades. Our negative findings of attentional regulation alone should not completely rule out attentional regulation acting as a mechanism in the experience of mindfulness. Perhaps in our intervention having students be aware of their attention wandering because of their test anxiety made students focus more on their test anxiety which could run counter to the experience of mindfulness.

Finally, our results have theoretical implications. As more mindfulness-based research is conducted in school settings there is a growing need for MBIs grounded in constructs related to educational outcomes. By implementing an MBI using the variables of emotional and attentional regulation, we were able to isolate which variables had a significant effect on experiencing mindfulness, which turned out to be emotional regulation but not attentional regulation.

Practical Implications

Our research builds on the growing literature on developing MBIs to address academic outcomes (Schonert-Reichl & Roeser, 2016; Diamond & Lee, 2011; Flook et al., 2015; Goodman et al., 2014; Benn et al., 2012) including college students (Chiang & Sumell, 2019). However, our findings suggest that educators should be discerning and careful when selecting MBIs to use in classroom settings. For example, our research draws into question the findings from previous research that used mindfulness to address test anxiety (Cho et al., 2016). The intervention we used to facilitate mindfulness was 8 weeks, while the intervention used by Cho and colleagues (2016) was only a single week in length. They did not measure mindfulness directly and only used participants who self-reported high levels of test anxiety. Our research tried to expand upon Cho and colleagues’ (2016) research by measuring mindfulness directly with the PHLMS and using a complete range of test anxiety. Our results revealed no difference between conditions. Given the modifications in our own research, it appears that the length of the intervention and/or the tasks used in the intervention were not sufficient to facilitate mindfulness as predicted. This suggests that researchers should be cautious in how they develop interventions interpret the results. It is possible that under the right conditions that an MBI could facilitate changes in test anxiety, but it is also possible that it could have no effect.

Results have important practical implications for how MBIs should be developed toward obtaining scientifically sound results. Researchers using MBIs to address academic outcomes should be certain that their expected outcomes are aligned with their conceptualization of mindfulness and how the intervention is structured to facilitate mindfulness. For example, when attempting to address the cognitive aspects of test anxiety, such as thoughts wandering during testing (Salend, 2010), a mindfulness intervention could be developed around attention training through practices like focused attention meditation for an intervention grounded in theory (Lutz et al., 2008). We attempted to ground our conceptualization of mindfulness by aligning expected outcomes with theoretically defensible intervention practices. For example, we used focused attention meditation to facilitate the experience of attentional regulation and open-monitoring meditation to facilitate emotional regulation. The experience of mindfulness regarding test anxiety seems to be strongly facilitated by the ability to regulate emotions in the present, while attentional regulation had very little predictive power.

Limitations

No research endeavor is without limitations. All data except for quiz scores were collected in the form of self-report measures. While self-report measures are a valuable tool and effective method for gathering participant data, there is a possibility that participants responded positively to the instruments out of social desirability or complying to the requests made within the study. Even with asking participants once per week to self-report on how often they participated in the intervention, we cannot be completely certain how often participants engaged in all the various elements of the mindfulness intervention. A possible outcome is that some participants participated fully in the intervention by reading each of the articles sent and complying with the prompts to practice mindfulness while other participants may have engaged only partially or did not engage at all.

Another limitation can be found in the timing of this study. Our study took place during the COVID-19 global pandemic. T1 data was collected prior to the pandemic, but T2 and T3 were collected during the pandemic. The university that we collected data from transitioned to online instruction between T1 and T2. The transition of classroom settings and the global pandemic likely had a large impact on student stress and even their experience of test anxiety. Our results suggest that there is a chance that all students in this study experienced a heightened degree of general stress that could render the intervention less effective. This was perhaps most evident in our data collection using attentional regulation scores from T1 to T2. Reliability scores using Cronbach’s alpha dropped from 0.85 to 0.40. This assessment was attempting to measure the construct of attentional regulation so it seems possible that students might have experienced a much more difficult time managing their own attention during the beginning of the global COVID-19 pandemic when certainty was much scarcer. T3 data was collected 4 weeks later when students were likely more acclimated with the transition to online learning. Attentional regulation scores at T3 were measured at 0.86, much closer to the value at T1.

An additional limitation from this study is that participants were given an assessment on different topics and in completely different formats of physics as the semester progressed, based on the COVID-19 outbreak. While the topics in each unit and assessment build upon each other, results could imply that improvement occurred from one assessment to another for other reasons beyond the scope of this intervention such as familiarity with the assessment style, changes in study and preparation habits, or maturation across time in the course. Additionally, at T2 participants were given an online discussion topic instead of a quiz because students could not meet on campus due to the COVID-19 outbreak. While the discussion topic was described by the classroom instructor as being more difficult than typical classroom assignments, it was completed in groups instead of individually, and is likely not comparable to the assessment given at T1 and T3. This makes evaluating the effectiveness of the mindfulness-based intervention on quiz grades difficult.

Furthermore, it should be noted that 40.6% of our participants self-reported previous mindfulness experience. This could have impacted the results by limiting the impact of the intervention for those who already had mindfulness training or elevated the mindfulness scores of participants in the control condition, making the differences between our conditions not statistically significant.

Finally, a limitation was found in the exam emotion regulation questionnaire used to assess emotional regulation. This instrument generated low levels of reliability using Cronbach’s alpha. These results are not surprising since we were unable to find a fit for emotional regulation in our theoretical models on mindfulness and test anxiety. Perhaps the exam emotion regulation questionnaire was not the most useful instrument to implement for assessing this construct.

Future Directions

There are several future directions that follow from the findings of our research. We will review and discuss two potential areas of future research exploring mindfulness and test anxiety including: (1) exploring the relationship between mindfulness training and test anxiety with other academic populations including adolescents and elementary students and (2) using the experience sampling method to gain a better understanding of when mindfulness training is most effective at reducing test anxiety.

This research explored the relationship between mindfulness and test anxiety with college students. However, test anxiety is experienced by students at all academic levels, not just in college (Salend, 2010). Based on the results, we hypothesize that it would be worth exploring the effectiveness of mindfulness-based interventions, especially interventions that can be delivered online that target younger student populations such as adolescents in middle and high school. Research by Lillard (2011) suggests that mindfulness-based interventions in school settings can be adapted to be appropriate for students of all age groups. We would like to continue this line of research but with students at different levels of educational experience including adolescence, who may benefit from mindfulness-based interventions given that this period of development is a time of storm and stress (Roeser & Pinela, 2014).

Another future direction is to explore other sampling methods for measuring the outcomes of mindfulness-based research that targets test anxiety as an outcome. This research attempted to capture participants’ experience of practicing mindfulness before training began, during training, and after training concluded. Research exploring mindfulness-based interventions might also benefit from assessing the experience of mindfulness closer to when participants are practicing the skills, they are taught during an MBI, including their assessment throughout their day-to-day lives during the intervention. The experience sampling method (ESM) offers some guidelines for how this could be accomplished. ESM has the potential to reveal more nuance about how participants are experiencing mindfulness in the moment throughout an intervention. For example, ESM was used successfully by Nakamura and Csikszentmihalyi (2014) to assess the construct of flow as this construct occurs in a participants’ everyday experience. This type of information could be very useful to future researchers exploring the impact of mindfulness-based interventions by examining when they might be most effective for reducing test anxiety and improving grades on academic assessments. The construct of mindfulness is ripe for investigation using alternative methods for gathering data such as ESM. For example, recent research by Tschacher and Lienhard (2021) used ESM data collection techniques to explore the relationship between affectivity and state mindfulness.

Contributions and Significance

Our research fills a gap in the literature by using an online mindfulness-based intervention that targeted test anxiety with introductory physics students. Research by Cho and colleagues (2016) used a self-taught mindfulness-based intervention to address test anxiety in college students; however, they only included participants with high self-reported test anxiety. Additionally, their intervention only lasted a week and did not measure mindfulness directly. Findings from the intervention portion of this study did not result in significant changes of test anxiety, suggesting that 8 weeks was potentially not enough time to facilitate the skills of mindfulness.

This research also adds to the existing literature by exploring the way in which mindfulness is conceptualized for the purposes of facilitating a mindfulness-based intervention. This research was guided by Bishop and colleagues’ (Bishop et al., 2004) two-factor construct of mindfulness being defined by (1) self-regulation of attention toward the present and (2) an open curiosity toward what happens in the present. We predicted that these two dimensions represented elements of attentional regulation and emotional regulation, respectively. Using multiple regression modeling, our results suggest that emotional regulation played a significant, positive role in predicting test anxiety and quiz grades, while attentional regulation was a negative predictor of mindfulness. Perhaps our intervention was able to facilitate the experience of emotional regulation, but not attentional regulation. This is important for future interventions using elements of mindfulness because future interventions should be developed in which the theoretical conceptualization of mindfulness and predicted outcomes are in line with one another and therefore can be justified.

Conclusions

Our goal was to explore the theoretical bounds of how mindfulness is conceptualized in educational settings by using an 8-week online mindfulness intervention with physics II students. The intervention did not generate significant differences between conditions; however, results suggest that emotional regulation was a significant, positive predictor of mindfulness. Conversely, attentional regulation was a negative predictor of mindfulness. This research adds to the literature on how future mindfulness interventions should be structured, implying that the intervention design and expected outcomes should be theoretically aligned to one another. Results also suggests that educators should be discerning and cautious using mindfulness-based interventions in classroom settings.

Author Contribution

John Chancey took the lead in writing this manuscript. All authors provided meaningful, critical feedback that helped shape the research, analysis, and manuscript.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available due to the data being associated with student grades, but are available from the corresponding author on reasonable request.

Declarations

Ethics Approval and Consent to Participate

Our research was given approval by the University of Oklahoma Institutional Review Board (https://compliance.ouhsc.edu/HRPP) that performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. All data collected and used for this manuscript were given informed consent by the participants.

Conflict of Interest

The authors declare no competing interests.