The Stress of Online Learning (Distance Learning)


Stress is recognized today as impacting both quality and length of life (Kiecolt-Glaser, McGuire, Robles, & Glaser, 2002). Stress was defined by Hans Seyle (1936) as the unspecified physiological response to aversive stimuli. The stress of learning is not yet understood. If stress impacts physical and emotional well-being, and lifelong learning is needed to survive in the information age, then a study of the stress of learning may impact both nursing and educational practice.

Learning stress can create a number of long-term physiological and performance complications. Stress reduces immune function, making people vulnerable to disease. Studies indicate stress hormone levels can be predictive of relationship problems and chronic disease. Reducing stress could avoid colds, flu, and mild depressive symptoms, which complicate student relationships and achievements, thus increasing stress (Glaser, Robles, Malarkey, Sheridan & Kiecolt-Glaser, 2004). Stress also blocks learning by limiting perceptions, thinking, and memory capabilities during performance, triggering higher levels of stress during later performance events (Sapolsky, 1998). The inability to think or remember concepts, procedures, and methods during patient encounters can threaten lives.

Until recently, researchers found measuring learner stress under normal everyday conditions challenging. Self-report methods are unreliable due to the filtering that occurs between the experience and the report (National Space Biomedical Research Institute, 2003; Razavi, 2001). Measurement of physiological function during events is considered the most reliable reading of lived experience. The measurement of cortisol and cardiopulmonary variables provides a picture of stress as it occurs.

Increased glucocorticoids impair memory function (Kilpatrick & Cahill, 2003), problem solving, and spatial recognition memory (Rozzendaal, de Quervain, Ferry, Setlow, & McGaugh, 2001; McEwen & Seeman, 2003). Cortisol, a glucocorticoid, is regulated by the adrenocorticotrophic hormone (ACTH) (Carrasco & VandeKar, 2003) and plays a major role in stress (Char-ney, 2004), learning, and memory processes (Bremner et al., 2004) that underlie behavioral adaptation (Gesing et al., 2001). Salivary cortisol reacts within 10-20 minutes of a stressor.

Cardiac and respiratory systems instantaneously respond to perceptions (Sapolsky, 1998). Until recently, heart and respiratory variables could only be measured periodically, providing a snapshot of physiological fluctuations which can miss peak and valley responses. Recent technology advances make it possible to include biophysical measures in educational design models.

Nursing school is considered a stressful education experience (Hughes et al., 2003). Students experience stress-related ailments like ulcers and autoimmune disease (Heath, Macera & Nieman, 1992; Reid, Mackinnon & Drummond, 2001). This motivates colleges to provide stress intervention programs (Pitts, 2000). After graduation, nursing is considered a high-stress profession where role strain and burnout are frequent occurrences. The complex job requires making life and death decisions under pressure, and few rewards counter the job demands (Siegrist, 2000). Although nursing is an extremely stressful profession, the measurement of educational stress is an important topic for most professions.

Learning stress needs to be addressed as part of the educational experience. This article presents a framework that permits the study of biophysical simultaneously with environmental and instructional variables.

A report of a feasibility study asking if technology can detect differences in stress levels of online students in different settings is also presented.

Learning Allostasis Model

The Learning Allostatis Model (see Figure 1) was designed to better understand learning stress and is based on the work of Bruce McEwen. The term allostasis means “maintaining stability (or homeostasis) through change”; it was introduced by Sterling and Eyer (1988) to describe how the cardiovascular system adjusts to resting and active states. McEwen developed the Allostatic Load Model to explain how physical, environmental, and psychosocial elements combine in a cascading cause-and-effect process that accumulates over time and alters the body’s natural mediators producing disease processes. Experience, genetics, and behavior all contribute to the body’s reaction under aversive conditions (McEwen, 1998). The Learning Allostasis Model attempts to explain individual student stress reactions to learning events.

The learning allostatic model posits individuals begin learning events affected by biological, environmental, and major life events. The learner’s approach to the event includes protective behaviors that may either help or hinder achievement. Individual characteristics impacting the learning event include: genetic makeup, milieu, life experience, learning experience, and learning strategies. Instructional design is given similar weight to individual characteristics. Instructional design variables include pedagogy, facilitation, and institutional support.

Figure 1. Allostatic learning model

Allostatic learning model

The model holds promise for new approaches to educational research and instructional design. Use of the model permits control and analysis of biophysical, psychosocial, and instructional design variables in one study. The model encourages use of clinical data as well as self-reports in the same study.

Methods of studying stress

Until recently, measurement of cortisol or continuous heart and respiratory function demanded laboratory conditions. Blood draws, urine samples, and connection to machines during rest periods prevented measurement of the normally lived experience. Educational stress variables were not studied because of the laboratory limitations. The invention of salivary sampling enabled field study collection of cortisol, and the development of ambulatory monitoring devices permitted continuous cardiopulmonary tracking.

Salivary cortisol measures function of the glucocorticoid system (Stewart, 2000). Cortisol activates the fight or flight bodily reactions. Salivary samples are reliable, inexpensive, and possible during in vivo study. Since the heart and lungs are instantaneously responsive to aversive stimuli, continuous measurement can provide a “moving” picture of experience. In past educational research, heart and respiratory rates were measured periodically so that events between measurement periods were missed. Periodic measurement makes it impossible to find the peaks and valleys of stress or to detect abnormal events. The invention of ambulatory measurement devices permits noninvasive continuous cardiopulmonary measurement. Laboratory conditions are not needed and classroom data is researchable. In addition, the complex isolated distance education data becomes possible to track.

The lightweight LifeShirt® from Vivometrics, Inc. is worn under one’s clothing. The vest collects objective physiologic data through various sensors, including RIP bands that measure pulmonary function (tidal volume, respiratory rate, etc.), electrical activity of the myocardium via a 3-lead electrocardiogram, and activity/posture via a tri-axial accelerometer. The personal-digital-assistant-size computer permits participants to enter subjective information in a diary. The computer can also cue the participant. The device is slipped into a pocket, worn on a belt, or hung over a shoulder. The vest measures 30 cardiopulmonary variables and can be used as a platform for other tools like electroencephalograms. The descriptive statistical software accompanying the vest allows researchers to diagnose problems, conduct trends, view findings in graphic or numeric formats, and gather descriptive statistics from a few seconds to 14 hours.

The LifeShirt® and salivary cortisol samples provided the primary data in a recent study of online learning. The research question asked, “Can technology detect differences in the learning stress of online learners in two settings?” Personal information, self-report measures, and instructional design variables are reported elsewhere.


Thirty-nine students studied “how to author a database” using an online module adapted from a Microsoft Access tutorial. Junior year-semester one nursing students were randomly assigned to one of two sections. Students in the isolated section completed the study alone in a room with an Internet-connected computer. The classroom section participants studied in a computer lab with more than two companions and an instructor who answered questions. All students provided salivary cortisol samples prior to the study, one hour into the learning experience, and 30 minutes after completion of an online quiz. Participants wore ambulatory monitoring devices throughout the experience. Five-minute relaxation periods began and ended the study period.

All physiological measures were taken while students sat at their computers. Five-minute samples of the continuous variables provided the means used. Heart rate and pulmonary function were sampled for pre- and post-learning phases, and at the one hour “peak stress” period. Statistics for “maximum” and “minimum” as well as an “overall” mean for each period were compared.

salivary cortisol Results

All students experienced a salivary cortisol rise, indicating a stress increase at some time during the study. Since the diurnal expectation posits a steady decrease in cortisol levels as the day progresses, any rise is an indication of stress. Significant differences for the amount of change from times “pre” to “peak” and for “peak” to “post” were t-2.66 with a p value of <.01, indicating a rise and then a fall in cortisol levels for the whole class. No significant between-group differences were observed.

cardiopulmonary Results

A comparison of group means using the t-test noted no difference between the isolation and classroom groups for the pre-study heart rate mean, indicating students in both groups were similar as they entered the study. A slight difference between group heart rate mean during the “peak stress” period was p-.07 (t-1.842).

Repeated measures for the “overall” heart rate means (pre, peak, post) indicated a significant difference (f-12.712, p-.00) across time, but only a trend (f-2.72, p- .073) toward difference between groups. A mean of the maximum heart rate for the five-minute periods was taken, as an elevation in heart rate indicates reaction to aversive stimuli. The difference between the means of isolated and classroom groups’ “maximum” heart rate during the “peak stress” period was significant (t-8.96, p-.00). A minimum heart rate mean for each five-minute period was also compared. The minimum heart rate mean was p-.05 (t-1.991). The findings indicate learning stress differences were greatest during the study period for the two groups. The post period means were compared for the two settings and no significant differences were found for the overall, maximum, and minimum means. A comparison between groups for maximum heart rate means found (f-10.65, p-.00) over time and (f-9.01, p-<.01) for groups, indicating that although everyone experienced a jump in heart rate, isolated students experienced higher peak heart rates than their peers.

Respiratory rates differed over time. One measure of stress is the sigh. Sigh volumes, or the amount of inspiration that exceeded by 20% normal means, were measured. The sigh volume means significantly differed between groups (t-2.82, p-0.01). The whole sample differed in the sigh volume maximums from pre-study and peak stress periods (t-2.46, p-0.02), and then again between peak stress and post study (t-2.38, p-0.02), indicating study periods had more sighs than pre and post periods.

The tidal volume, or amount of inspiration of regular breathing, demonstrated similar findings to sigh volumes over time. Maximum rates (t-9.67, p-.00) and minimum rates (t- 5.68, p-.00) were significant. The only difference between groups was during the beginning relaxation period (t- 2.99, p-0.01), indicating students in isolation may not feel as relaxed as those in classroom situations.

The number of breaths per minute indicated people breathe differently during learning (t-4.68, p-.00). There was also a difference in breaths per minute between the isolation and classroom groups. The variation may indicate students in isolation were not as relaxed as those in the group situation (t-2.06, p-0.05).

Respiratory rates demonstrated similar changes over time (t- 9.05, p-.00). Differences between the two groups before the learning period were p-0.00 (t- 2.99). This again may indicate students in the isolation group were not as relaxed as the classroom group. Other respiratory measures provided similar findings with differences between pre and post levels of the two groups and differences over time.


Findings suggest that students studying online experience stress variable fluctuations. Students underwent a rise and fall of cortisol over the learning experience, indicating stress. No cortisol between group differences may indicate a lack of test sensitivity or too small a data collection period, since cardiopulmonary variables detected differences. A larger sample or a longer study period may produce other findings.Analysis of the continuous cardiopulmonary variables detected differences in over time and between groups’ means.

Measuring stress with the LifeShirt® and periodic salivary cortisol appears feasible. Participants stated that the LifeShirt® felt comfortable and did not interfere with learning. All measurements were taken while students remained seated. Variable differences occurred even though body movement was controlled. Salivary cortisol changes occurred, indicating stress. Findings suggest educational questions using biophysical data are now possible. Using the Learning Allostatic Model can contribute to research designs that use a whole-person approach.

Combining subjective self-report with objective physiological information in one study reveals the student learning process. Study of the student process as it occurs can give insight into distance education, which has proven “invisible” until now. Why satisfaction and achievement outcomes differ from one setting to another can also be tested.

Differences found between online isolation and classroom groups suggest that the learning environment influences student experience. Further study may indicate what types of supports are most efficacious for isolated students. The physiological response of distance learning may be different from face-to-face learning. Data suggest that all students experienced elevations in heart and respiratory rates, indicative of stress reactions. Continuous measurement of cardiopulmonary variables reveals significant differences between the pre, post, and learning periods. Amounts of “normal” fluctuations during study have not yet been determined. Research to determine levels of eustress and distress are needed. Qualitative feeling journals collected throughout the study period paralleled biophysical findings. All students reported some portions of the module confusing, frustrating, or upsetting.

Group cardiopulmonary data may indicate test sensitivity to different online conditions. Students in isolation experienced higher respiratory rates before the learning than their peers. The reasons for heightened respiratory rates should be examined in future research. Future researchers may establish “normal levels” of biophysical fluctuations that can “diagnose” individual and product problems. Educators may find the nursing process framework adaptable to learning situations in which teachers “nurse” students through their learning process.

Stress impacts the learning process according to researchers. Since the feasibility of gathering biophysical variables during lived experience is now possible, environments that impact learning such as distance education, simulations, and virtual learning can be compared. The immediate and continuous measurement of biophysical reactions can aid in understanding student process and filtering of psychological reactions.

Future discussion of the costs and benefits of delivery model differences could include biophysical measurements since research indicates stress impacts health. Once the discussion begins, researchers can then explore coping strategies for learning stress. More understanding of individual responses to learning stress are needed from different age groups, cultures, situations, and disciplines. Nurse researchers can now explore the impacts of stress during lifelong learning, patient education, and health promotion situations.

key terms

Cortisol: A stress hormone that activates the “flight or fight” syndrome.

Learning Allostatic Load: The accumulation of perceptions and decisions that develop into self-esteem and protective strategies that may or may not be beneficial to future learning events.

Learning Allostatis: The constant changes an individual makes to maintain homeostatis. The actions are in response to learning and life events that impact knowledge and performance.

Learning Allostatis Model: A whole-person model describing learner characteristics and relationships to stress and learning.

Stress: The physical and psychological reaction to aversive stimuli.

Next post:

Previous post: