230

ROBERT R. HOFFMAN

Zeier, H. (1998). Psychophysiological stress research. (this issue)

Zsambok, C. & Klein, G. (Eds.) (1997). Naturalistic decision making. Mahwah, NJ:

Erlbaum.

Psychophysiological stress research

HANSZEIER

This paper gives an overview of physiological stress respollS:es! and shows methods for measuring such responses in field studies. It discusses the available techniques for assessing endocrinological and immune functions, ~diovascular functions, physical activity, electrodermal activity, muscle activity, respiration, and:, the size of the eye pupil. Furthermore, methods for collecting psychological data are depicted. As examples for psychophysiological stress research under natural conditions, a showmaster's heart rate responses during a TV show are reported, as well as psychophysiological effects of work demands in air traffic controllers and in simultaneous interpreting. The latter example shows that mental overload in simultaneous interpreting may change the attitude to the job: It is taken less seriously and a certain carelessness sets in. This might be an effective self-defense mechanismagainst mental overload caused by situations such as increased time on task, .extremely fast speasees, unIntelligible speakers, and.long working hours. It seems to occur rather automatically, without being noticed by the performing interpreter. Especially ~increa.~ time on task, interpreters' own judgement of output quality, wheilier tbe)'llEe novices or experts, seems to become extremely unreliable.

1. Introduction

Stress is a commonly used term for a wide area of problems associated with physiological, psychological and social human activities. Although the lay public appears to have a fair understanding of the concept of stress, there is no generally accepted scientific definition. Stress consists of the psychophysiological processes caused by a perceived threator danger. From a p~ychological point of view the phenomenon has hV0 components: (1) the experseece of a threatening and strenuous situation, and (2) the uncertainty vtne~ one is able to cope withfuis situation.

~~ringVol'. :l!{m),.l%rr. 231-'2!'@ © JohaBenjamins PUblishlng Co.

232

HANS ZEIER

At· the workplace, most of the stress is either due to social factors such as interactions with colleagues, superiors and subordinates, or due to workload. The latter is especially likely to become a stress factor if a subject cannot control workload and cannot foresee critical situations. Typical professions in this regard are those of air traffic controllers, doctors and nurses in intensive care units, or simultaneous interpreters.

The term stress is generally used to denote three different aspects of the complex phenomenon of stress: (1) The stimuli producing stress reactions, sometimes also called stressors or objective stress; (2) the physiological and behavioral stress -i'eactions themselves; and (3) the various intervening processes. Stress-producing stimuli and stress reactions can be observed, while the intervening processes can only be assessed indirectly by measuring stress reactions or questioning the subject being exposed to stress. The intervening processes decide if an individual perceives a given situation as stressful or not. This stress experience is also called SUbjective stress or strain. As it is an individual experience, it is very subjective. What is stressful for one person may not be so for another, and even what is stressful for a specific person on one occasion may not be so on another.

2. Two separate response systems for active and passive stress

reactions -

With respect to the behavioral stress reactions one can distinguish between active and passive coping behavior. In accordance with the kind of behavior chosen, active coping behavior consists of the fight or flight response. Both alternatives evolved over time and are associated with physical activation. By stimulating the sympathetic nervous system the body is instantaneously prepared either to fight or to flee. Today's threats and strains can in most cases no longer be managed by physical activity. However, they still activate the sympathetic nervous system as in our evolutionary past. Therefore, physical activity, as e.g, jogging, is a good means of reducing stress-induced overactivation of body functions.

Situations which engender in us the fear of losing control usually elicit some active coping behavior. However, if one loses control and does not know which active behavior to select for managing a stressful situation, active coping behaviors will get suppressed and passive coping occurs. This re-

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

233

spouse may range from avoidance behavior, resignation, feelings of inferiority and lack of self-confidence to severe depression.

Active and passive coping behaviors are associated with two separate physiological response systems: The sympatho-adrenomedullary system and the hypothalarnic-pituitary-adrenocortical system. Given their pioneering work regarding the functions of these neuronal mechanisms, the sympathoadrenomedullary system is also called Cannon stress-axis (Cannon, 1932), and the hypothalarnic-pituitary-adrenocortical system Selye stress-axis (Selye, 1950). Stressors have an impact on the sensory organs. They transmit sensory information over afferent neuronal pathways to the central nervous system. The conscious analysis of this information is performed in the association cortex of the brain, while emotional aspects are evaluated by the so called limbic system. In stressful situations this system elicits physiological stress reactions.

The Cannon stress-axis stimulates the sympathetic nervous system via the hypothalamus; this leads to an activation of the adrenal medulla and the secretion of the catecholamines epinephrine and norepinephrine (Goldstein, 1987). As a result, blood pressure, heart rate, respiration cycle, metabolic rate, muscle activity and electrodermal activity increase. On the ofu!erhamL saliva secretion and peripheral body temperature - e.g, the temperature of the fingers and hands _ decrease, leading to feelings of dry mouth aadcold hands. Furthermore, energy stores are tapped and stored sugars and lipids are released into the bloodstream. This physiological response is called ergotopic response (Hess, 1948), as it prepares the organism to perform physical work. Chronic stress may lead to increased blood pressure, arteriosclerosis and even to coronary heart 'disease. On the behavioral level, chronic stress leads to overactivity, nervousness, irritation, motor restlessness, aggressiveness, continuous readiness to fight and muscle tension.

The Selye-stress-axis leads from the hypothalamus VIa fue pituitary gland to the adrenal cortex. It stimulates the secretion of theadirenocortical hormones, in particular the stress hormone cortisol. Pro[o;_nged mental and emotional stress causes disturbances in mineral metabolism. Fm;fuermore, it inhibits secretion of growth and sex hormones which may~seimpotence, and impair immune functions (Ader, Felton & Cohen, 1991; Van RGOd, Bogaards, Goulmy & van HOl!lweling.en, 19'93). The:str:e,ss;-iDON£ed,effec~ on immune functions explain why undercondltioills of ooderdl'rmmcsnessollie is more likely te catch. a cold or flu, which sometimes breaks <Jutjnst atfue moment when a prolonged period 'Of stress comes, to' an ood. In skessful

234

HANS ZEIER

situations, the withdrawal into passivity, and even paralysis, can be accompanied by an activation of the parasympathic nervous system. The consequences are vegetative stress symptoms such as an upset stomach, vomiting or diarrhea, and in the long run ulcers and other gastrointestinal diseases.

The two stress-axes do not function independently of each other, they may also be activated simultaneously. However, in most cases one of the two stress-axes takes the lead and dominates over the other. The output systems do not directly induce the corresponding active and passive coping behaviors, but favor them in the sense of behavioral preferences. To a certain extent, these dispositions can be changed, but also reinforced, by rational control.

Recent studies have shown that sympathetic activation is not necessarily a negative event, particularly when a stressful episode is followed by rapid recovery. Sympathetic activation can indicate successful as well as unsuccessful coping, or in Selye's (1974) terminology, stress or distress. On the other hand, the pituitary-adrenocortical axis appears to react in a more differentiated manner. It is mainly activated by distress. According to Frankenhaeuser, Lundberg & Forsman (1980), and Dienstbier (1989), passive coping is associated with pituitary-adrenocortical arousal, while successful coping seems to suppress the cortisol response. However, this dissociation might depend on a high level of personal control in task performance (Frankenhaeuser et al., 1980) or be more typical of chronic rather than short term stress, because periods of excitement, engagement, and involvement are reported to correlate with brisk increases in cortisol levels (Singer, 1974).

3. Physiological measurements in stress research

In the field of psychophysiological stress research physiological responses are measured, in order to assess subjective stress or strain associated with a particular situation. Alterations in the time course of physiological functions are used as an indicator of emotional and mental processes. As mind and body form a unity, every physiological function may be influenced in a certain way by some psychological process. Therefore, a wide variety of physiological functions can be used for this purpose. However, some physiological functions are much easier to record than others, and therefore are more frequently used. Continuous recordings from free moving subjects can be obtained by ambulatory monitoring, using either biotelemetry or portable recording de-

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

235

vices. For stress research, observations in natural settings are more appropriate than laboratory experiments, since simulated stress is quite different from real stress. The following overview, which merely provides examples rather than being complete, shows some examples of physiological functions used in

stress research.

3.1 Endocrinological and immune functions

Parameters representing a wide variety of endocrinological and immune functions can be assessed by analyzing blood samples. These techniques are very well established and routinely used in clinical medicine. However, taking blood samples has to be considered as a stress inducing event. Therefore, it is not very suitable for stress research. AB an alternative, urine or saliva samples can be taken. Many substances circulating in the blood are also present in these fluids, although some of them only in very low conce~trations. Urine samples are sometimes used for assessing the catecholammes epinephrine and norepinephrine. However, collecting urine samples might also be rather unpleasant. The least problematic assessment procedure for stress research is the collection of saliva samples. Saliva samples can be collected through a simple procedure that has a subject chew a small cotton roll for a minute or two. Today's analytical techniques are now sensitive enough to reliably determine even very low concentrations of a varie~ of substances present in saliva. Parameters most frequently used are cortisol (Kirschbaum & Hellhammer, 1989), dehyciroepiandrosterone,testosterone, catecholamines and immunoglobulin A (Kugler, Hess & Haake, 1992;

Hennig, 1994).

3.2 Cardiovascular functions

Recordings of cardiovascular functions are frequently used in psychophysiological stress research since stressful situations have a considerable impact on them. Continuous recordings of heart rate can be easily obtained ,villi a cardiotachometer. Such a device provides beat-to-beat measures of the heart rate, usually by recording the intervals between R"wa¥esofthyele~ogram. Measuring blood pressme is more complicated, altbOi£il!gn·· there exist portable, fully automatic devices that IDfiate anann-cWIaccommg to a preprogrammed schedule and record both systolic 'and diastolic psessese, Some

236

HANS ZEIER

approximation of peripheral blood flow can be obtained by using plethysmography or measuring skin temperature.

3.3 Physical activity

Body movements and physical activity can be assessed quantitatively using mechanical motion detectors, usually acceleration sensors. Such actometers transform body movements into an electrical analog signal. The output of even very simpledevices shows significant correlations with metabolic parameters such as o5tygen uptake (Avons, Garthwaite, Davies, Murgatroyd, & James, 1988). When monitoring heart rate, actometers are routinely used in order to assess heart rate increases which are not caused by physical activity. Furthermore, subjects can be requested to answer questions regarding behavior and emotional state by feedback signals of emotionally induced increases of heart rate (Myrtek, Brugner, Fichtler, Konig, Muller, Foerster, & Heppner, 1988). Other uses of actometers include the assessment of motor restlessness

,

patterns of physical activity or sleep quality.

3.4 Electrodermal activity

Sympathetic activation of the sweat glands reduces the electrical resistance of the skin. If one places two electrodes on a skin surface rich in sweat glands, e.g. on the palm or ventral surface of the digits, and applies a small constant current through them, alterations in galvanic skin resistance or skin conductance can be measured. This parameter responds very sensitively to sensory or emotional stimulation (Boucsein, 1992). This technique is frequently used in psychophysiological research and much easier to apply than direct measurement of sweat activity. The latter could be carried out by counting wet sweat glands on a circumscribed surface area of the skin, or by measuring vapor output or total weight loss.

3.5 Muscle activity

cd,·

~;\f'

In psychophysiological research, interest in muscle activity focuses on skel-

etal muscle tension, as it is maintained via involuntary mechanisms. Muscle tension is the resting state of muscular contraction. It consists of a so called isometric contraction, since the length of the entire muscle remains un-

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

237

changed. In a state of complete relaxation, there is virtually no muscular contraction, and, consequently, no tension. In a stress situation, muscle tension increases in order to prepare the body to perform the movements necessary for fighting or fleeing. A limb starts moving, as soon as the tension of an associated muscle exceeds the mechanical resistance caused by gravity and tension of antagonistic muscles. In such situations, muscle length decreases, and an isotonic contraction occurs. Isometric and isotonic muscle contractions are both the result of summated contractions of a certain number of single muscle fibers. The more muscle fibers contract simultaneously, the bizzer the resulting muscle tension. Contraction of a muscle fiber is .. initiated

eo...... .

by an electrical current, called action potential, flowing over-its membfime. A

small portion of thiselectricaI activity spreads from the rnesele to the skin and can be recorded with surface electrodes to produce an electromyogram (EMG). If many muscle fibers contract simultaneously, the electrical potentials may be quite large. In order to get an indicator of muscle tension,the potentials picked up with two surface electrodes placed over the muscle.gro~p to be evaluated, are first amplified. In a next step, the output of fue amplifier IS rectified, so that the positive and negative deflections of the current wave appear in one direction. This signal is then fed into an integrator which accumulates the deflections over a short period of time. The result is a numeric measurement of total electrical activity as a function of time, usually calibrated in microvolts per second. Although this integrated EMG is not a direct measure of either muscle tension or muscular contraction, it gives a fairly good estimation of tension in skeletal muscles. EspeciaHym the range of isometric contraction, the relationship between EMGamjJ1iltlldes and muscle tension is quite linear. Therefore, recordiag . integrated EMG has become the most popular technique for measuring muscle tension iapsychophysiological research.

3.6 Respiration

There are several possibilities for monitoring the respiratory cycle. The easiest way is to put a flexible strain gauge belt around the chest or upper abdomen. Respuatory movements fuem alter the lefigtb6ftb~ !JelL. The resulting alrerations in electrical resistance can be meastm:d wi:ID an eleetronic device ... In this way, the respiratory movemeatsare tr:anlSforinedmtoan electrical analog. signal.

238

HANs ZEIER

3.7 Pupillometries

Psychological influences, such as positive and negative emotions, perceptual processes and mental activity, have very consistent effects upon the size of the eye pupil. Alterations in pupil size can be recorded with a small TV camera, fed into a computer and automatically analyzed.

4. Psychological measurements in stress research

In stress research, psychological data can be collected either by an independent expert who is observing and recording the behavior of the subjects to be investigated, or through self-report and self-assessment. Furthermore, in many situations objective data can be collected by using some performance measurement, such as workoutput or workload. Behavioral assessment by an observer has its limits with respect to the number of subjects to be simultaneously observed and the observer may interfere with the investigated situation, especially if one has to intervene in subjects' privacy. However, a minimal amount of observational data and objective measurements should always be collected since they are far less biased than self-reported data. One method for simultaneously observing a larger number of subjects is the multimoment analysis. With this technique repeated brief observations are carried out according to a predetermined sampling schedule.

The simplest means of collecting self-reported data are questionnaires and interviews. Standardized questionnaires exist for the assessment of control variables such as personality traits, intelligence, or feelings of anxiety and other emotions. For recording stressful encounters, memory problems may arise if there is a large time lag between the stressful events and their report in a questionnaire or interview. Furthermore, verbal reports about past experiences are frequently general self-descriptions of reaction tendencies, representing characteristics of the subject's cognitive self-representation, rather than of his or her behavior (Ericson & Simon, 1980). Therefore, self-reports about stressful events and stress-related behaviors should be made at the moment when these events occur, or, in order to avoid situational interference and behavioral reactivity, shortly thereafter. This can be achieved by way of keeping systematic diaries. Recently, sophisticated assessment procedures that rely on modern data acquisition technologies have become aVai\~ple. Using

PSYCHOPHYSIOLOGICAL STRESS RESEARCH

239

small pocket computers, psychological data can be self-recorded and accurately timed under real life conditions (Buse & Pawlik, 1996; Perrez & Reicherts, 1996). Furthermore, with acoustic beep signals which occur at certain times or when a physiological measurement exceeds a threshold, such recording systems can prompt the subject to respond to a questionnaire asking about activities, mood, coping behavior, or other aspects of the actualsimation.

The importance, for psychological assessments, of taking verbal as wen as nonverbal behavior into account is supported by physiological findings indicating that verbal information is predominantly processed in the left brain hemisphere and nonverbal in the right hemisphere. The complementarity of functions of the two hemispheres of the human brain is an efficient arrangement because each can independently bring to bear its own abilities to develop and fashion the neuronal input (Zeier, 1989). In general, the left hemisphere is specialized in language and in attending to fine imaginative details in all descriptions and reactions, i.e., it is analytic and sequential. Furthermore, the left hemisphere can also carry out arithmetic and other computer-like operations. With its capacity for logical tbinkmgand d:e:tailed analysis, it can detect linear relations between causes and effects· in isolated subsystems. The right hemisphere, on the other hand •. pefiiOI:IIlSS}'Bthetic functions. It is specialized in spatial, pictorial and coherentperceptioH. It bas. practically no verbal, but only musical abilities. In contrast to the abstract thinking of the left hemisphere, the thinking of the right hemispbere is associative and intuitive. In integrating detailed information into pictures and symbols it achieves holistic and synthetic thinking. With this capability it is better able to comprehend complex relationships, reciprocal causalities and holistic values than the left hemisphere which is specialized in anaIyzingfine details.

5. Examples of psychophysiological stress research

5.1 Heart rate response ala TV showmaster

Figure 1 shows the recordings of heart rate and pbysicalacti¥i~ of a TV showmaster at work. During rehearsal the subject was.~l'foonj,g .saIDe activities. as during the live. transmission, Ramel)!introduc~gi~edifferegt blocks of tD:e TVshow~ The upper recording trace depitpts ph:y,:sillcal activity . It

240

160
140
120
~
e 100
.....
...
0;-
...
::: 80
60
40
20
0
0 HANS ZEIER

Rehearsal

Life transmission

20

40

60

80

100

120

140

160

Time (minutes)

Figure l'l!eart rat~ and physical activity of a TV showmaster at work. Note the marked Incr~ase In heart rate occuring during a life transmission, but not during the previous rehearsal.

clearly prove.S' that t~~ observed increases in heart rate are induced by stress and not physical activity, During rehearsal, only few alterations in heart rate are observed. On the other han~, during the live transmission, which represents ~ real challenge, marked increases in heart rate occurred. A detailed analysis showed that the peaks appeared whenever the subject was announcrng a new show block. The last peak coincides with his last appearance on scre~n. Thereafter, heart rate quickly returns to the pretransmission level proving that the subject is well adapted to his job. '

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

241

5.2 Psychophysiological effects of work demands in air traffic controllers

The professional activity ofair traffic controllers (ATC) was investigated in a larger field study, during. which physiological and psychological data were collected (Zeier, 1994; Zeier, Brauchli, & Joller-Jemelka, 1996). Each subject completed two question sessions and two working sessions, once during a low traffic period and the second time during a period of high traffic. The working sessions lasted about 100 minutes and were always the first work period of the subject'S usual working day. For the question sessions standardized questionnaires were used. Different aspects of subjective and objective workload were recorded for all the working sessions investigated .

The questionnaire parameters chosen were psychological stress symptoms, such as complaints about physical discomfort, depression, negative self-communication, work-increased anxiety, coping behavior and subjective ratings and interpretations of various aspects of the work situation. The mean values measured were within the normal range, and only - a few significant differences were found between the two question sessions. However, about 10 to 15 percent of the ATCs showed elevated values forpsychologica1 stress symptoms, to an extent that indicated that they might have serious stress problems at work and/or in their private life.

In order to assess stress-induced alterations in concentrations of cortisol and immunoglobulin A, saliva samples were taken before and after each of the two working sessions. Additional saliva samples were taken before and

Table 1. Concentrations .of salivary cortisol and immunoglobulin A of air traffic controllers performing a working session. Postsession values are adjusted for session duration (mean ± SD).

N Before After
First working session
Cortisol (nmol/I) 126 10.5 .± 7.:6 H3 ± 't4
Immunoglobulin A (mg/eIl) 126 3.0 ±1.6 4.1 ± 2.2
Second working session
Cortisol (nniolll:) 126 8.7 ± 6.2 11.9 ± 504
Immunoglobulin A (mg/dl) 126 2.6 ± 1.4 3.9 ± 2.2
Control situation
Cortisol (mnoYI) 10 12 .. 1 ± 5.6 6:9/·± 2~1
Immunoglobulin A (mg/dl) 10 3.4 ± 1.9 2.9·± 1.3 242

HANSZEIER

after a routinely scheduled lecture about a future European air traffic control system, which provided the control situation. This lecture started at 08: 15 a.m. and the time interval between the two saliva samples was 100 minutes, i.e. the same as the average length of the working sessions investigated. The data of the assayed saliva samples are summarized in Table 1. Concentration of cortisol increased significantly during the working session, but decreased during the control situation. The latter is due to the well known circadian variation in cortisol concentration, showing a peak followed by a marked decline during the first hours of the morning (Vining, McGingley, Maksvytis, & Ho, 1983; Kirsehbaum & Hellhammer, 1989). In contrast to the expected immunosuppressive effects, the working sessions caused a marked increase in the concentration of salivary immunoglobulin A (slgA), This increase, however, was not correlated with the salivary cortisol response nor with the amount of actual or perceived workload. It seems to be determined by a different mechanism than the cortisol response, since the latter was significantly correlated with actual and perceived workload. Furthermore, the cortisol response was more pronounced in the second sessions. This fits well with the fact that the workload was slightly higher in the second than in the first sessions. Therefore, salivary cortisol may be a valid indicator of work demands. A comparison with psychological data revealed a negative correlation between the cortisol response and some stress factors such as alcohol consumption, coping by drinking, drug consumption, difficulty to relax from work and decreased, bounceback flexibility. This finding could be a consequence of some particular burnout behavior (Freudenberger & Richelson, 1980; Jones, 1982). Burnout is not only characterized by the above mentioned factors, but also by a certain indifference to and carelessness on the job. It may be assumed that indifference and carelessness reduce the psychological and physiological stress reactions, thus leading to a lower cortisol response. Although this coping strategy may be helpful to an overloaded ATe, it could compromise air traffic safety.

Air traffic controllers usually exhibit strong professional pride. As reported elsewhere (Zeier, 1992), the subjects of the present study were well satisfied with their work, although they were dissatisfied with the management, and they showed high positive scores on mood scales of an adjective check list measuring their emotional response to the work situation. This positive emotional engagement might explain the increase in sIgA. Therefore, measuring this physiological response seems to be a valuable tool for differ-

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

243

entiating between positive and negative stress effects (eustress and distr~ss according to Selye, 1974) or between successful and unsuccessful adaptation (Levine & Ursin, 1991), or coping (Monat& Lazarus, 1991; Perrez & Reicherts, 1992; Gaillard, 1993) with situational demands.

5.3 Psychophysiological effects of work demands in simultaneous interpreting

Simultaneous interpreting is a: highly complex activity (Gerver, 19174; Moser, 1978; Paradis, 1994) that requires the interpreter to simultaneously listen, anal;ze~ comprehend, translate, edit and reproduce j:nreal~timea speaker's utterance. All interpreters work with at least. two, man)' \V1ththree or, more languages that must be constantly main:amed at a high.l~ve1 of proficl~ncy. During a regular 30-minute turn, working from an original speaker whose speaking speed is between 100 and 130 words per ~ute, consider:dmore or less comfortable depending on the source language involved, an mterpn:ter processes and delivers final copy of an average of 3000 to 3900 words:bemg equivalent to approx. 12 to 15 type-written double-sp~ced page~.Wlthfa;t speakers, speaking at a rate of 135 to 180 words per mmute, themterpr~ter s output can increase to 4050 - 5400 words, or 16 to ~2 pages per 30-m:nute turn. As a comparison, the output of translators working as permanents m an international organization or in a company translation service varies between 3 and 10 pages a day. (Moser-Mercer,Kiinzli, &Korac,19.~8) .•...

Over time interpreters have developed workllgcondiuofllS deslgned to alleviate fatigue and help to ensure high quality of outputove{'thei~mse of a normal working day, normal work week and over the course of.lherrc:areer. These include a recommendation that simultaneous; inte'EpFeteIS not wock: alone for longer than approximately 40 minutes.fora single speee:h.~r sbo~ meeting and that they take turns roughly every 30 minutes when wor~~allday meetings. While it is generally accepted ~atmino!. stress fa~ilitates memory performance, because it raises the baseline level of arousal, mt~~se stress produces anxiety and creates overarousal which hinders cOgnIuv,e

functioning ineludinz memory performance (Searleman & HeIflIl.an.n, .1994 . .).'

, b. ri

Cardiovascular activity measures (Klonowicz, 1994)pOintIO;)JI~te;~il[}C~u

increased arousal in simuLtaneous interpreting,prodn?lligbt~. pressure changes th?fInMmicthose leading to the deVielopmentofessentiia[b~rteosion. (Moser.ccMerceret al., 1"998)

244

HANS ZEIER

In a preliminary study (Moser-Mercer, 1998) psychophysiological stress reactions and performance of interpreters were investigated. Five subjetcs, four females and one male, volunteered to participate in the study. All of them were experienced conference interpreters, with a minimum of twelve and a maximum of 25 years of professional experience in the booth. They had English as their native language, and German as their second or third language. Four speeches by German politicians were chosen as realistic source texts. They covered highly topical political and economic issues and had been delivered between August 1995 and September 1995. The issues had appeared in the daily press, both in German and English papers. All four texts were spoken by a native speaker of German, at a rate of 120 words per minute, onto audio casette tapes in the sound-proof recording studio of the Ecole de Traduction et d'Interpretation of the University of Geneva. They were checked for accuracy and speed by an independant German speaking judge. The experiment was carried out in the professional interpreting facilities of the Ecole de Traduction et d'Interpretation. The size of the interpreting booths, sound quality, air quality, visibility and working surfaces all conform to ISO-norms. The interpreter's output was recorded on tape. Interpreters participated in the experiment on a day and time chosen by themselves in accordance with their availabilities. Evening sessions were excluded. Subjects were given information in advance of the experiment, the titles of the speeches together with the dates and the names of the original speakers. In order to be able to situate the input material, they were informed about the rationale of the experiment and the fact that they had to collect saliva samples in order to determine stress hormones. Subjects were asked to budget at least 3 hours for the experiment. On site, they were familiarized with the installation, given water and drinking cups for their convenience, as well as written instructions on the saliva tests, and were told that they should stop working whenever they felt that their quality of output no longer met their expectations. The experiment started with a first saliva test. Then, the first speech was played and the interpreters started to work. The first break came after 30 minutes, lasting about 2 minutes for performing the second saliva test and changing the tapes. The next break came at roughly 60 minutes, again lasting about 2 minutes for performing the third saliva test and changing the tapes. No subject worked long enough for a third break. All cassette tapes of the interpreters' output were transcribed and analyzed for quality according to a rating scale adapted for this purpose from similar rating

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

245

scales used by researchers in the field of interpreting (Gerver, 1974; Barik, 1971).

Test tubes (salivettes, Sarsted, Germany) marked with the personal code-

number of each subject and containing a small sterile cotton roll were provided before the test session, after 30 minutes and after 60 minutes. Subjects were instructed to remove the test tube stopper, take theroll out of the test tube, put it in their mouth and chew it slightly for exactly two minutes. Then they had to put the saliva filled cotton roll back into the test tube, seal the ~be wi~ the stopper and return it to the investigator. The test tubes were immediately shipped to the Swiss Federal Institute of Technology (Depar_tment of Behavioral Sciences) in order to be frozen until laboratory analysis of all samples could be performed within a single run. Instructions for taking the saliva samples were handed out in written form. The subjects w~re instructed ~ot to eat during the experiment, and to rinse their mouth well WIth water 10 minutes before the first saliva collection. Smoking was prohibited, but water, sweets and chewing-gum were allowed. For laboratory analysis, test tubes were thawed_ and centrifu zed in order to extract the saliva from the cotton. The volume of saliva in each vial was recorded. Concentrations of cortisol and immunoglobulin A were determined by laboratory personal who had no knowledge of the

. experimental conditions of the analyzed samples, Concentrations of free cortisol were determined using the RIA kit Cortisol Coat-A-Counl: from Diagnostic Products and a gamma counter from Canberra"Packard. using the RIA-CALC and 4PL programs. Concentrations of immunoglobulin A were assayed by the single radial immunodiffusion (RID) method. For this p~dure, saliva samples and three human reference sera of known conce~lIait10nS were deposited by micropipette into cylindrical wells in an agar pIau: impregnated with mono specific goat serum containing antibodies to htmlan rmrmmoglobulin A (LC-Partigen-IgA, Behring Diagnostika, M~~ur.g a.~., Ge~y) and incubated at room temperature for 24 hours. Precipitation nng diamet~r formed in agar was measured with an ocular micrometer. 1}le immunoglobulin A concentrations were determined by regression equations derived from the square of the ring diameters of the reference sera against dl:eir concentEations

(Mancini, Carbonara, & Heremans, 1965). . .. . .

The data of the assayed saliva samples are su.mmarn;ed.lnTaa~~ 2. Due to the small number of subjects investigated, the data do notreadh<<mlySiatistic:a1 significance. However, they show some remarkable trendswlEcb are.~ line with the above mentioned findings in air traffic CQRtrofiers:. Secretion of

246

HANS ZEIER

Table 2. Concentrations of salivary cortisol and immunoglobulin A in interpreters performing a session of simultaneous interpreting.

N

Before

After 30 min

After 60 min

Cortisol (nmolJ!) Immunoglobulin A (mg/dl)

5 5

10.6 5.0

12.5 10.5

9.0 8.7

cortisol and immunoglobulin A tend to rise between initial measurements and those taken at 30q,minutes, probably caused by a considerable work load and emotional engagement, and then decrease with further time on task. It can be assumed that, since none of the interpreters had ever participated in such a study before, the baseline value taken at 0 minutes most likely reflects some elevated level of stress. A further increase in stress and considerable emotional engagement can be deduced from increased values at 30 minutes. The decrease in cortisol and immunoglobulin A during the second 30 minutes may be due to reduced eagerness on the job. This would very well correlate with considerable anecdotal evidence from professional interpreters who will readily admit that, if put in a situation where they have to exceed their personal limits, they just could not care less after a certain time. This assumption is supported by the finding that performance quality showed a marked impairment during the second 30 minutes. Especially in the category of most serious errors, i.e. meaning errors, almost twice as many errors were made during the second 30 minutes as during the first 30 minutes. Considering that each meaning error, no matter how minor, distorts the message, the considerable increase during the second 30 minutes on task represents a significant decline in output quality. Interestingly, the interpreters themselves did not realize this performance deficit, since they continued on task at least till the second break took place. This same lack of awareness of deterioration in quality can readily be seen in less experienced interpreters as well: students are usually unaware of how poor their performance can become with increasing time on task. With students, the point of diminishing returns appears usually after 10 to 15 minutes, as they have not yet developed the type of efficient processing strategies professionals have. Increased automation of sub-processes leads to more efficient utilization of cognitive resources and novices usually still have a long way to go (Moser-Mercer & Ktinzli, 1995). One must conclude that the professional interpreters included in this analysis

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

247

decided to stop only once their fatigue level had reached such proportions that they felt unable to continue. This demonstrates that interpreters' own judgement of output quality, whether they are novices or experts, is extremely unreliable after increased time on task.

Analogous to the above mentioned burnout behavior of some air traffic controllers (Zeier, 1994), mental overload in simultaneous interpreting may change the attitude to the job: It is taken less seriously and a ce~ carel~sness sets in. This might be an effective self-defense mechanism agarnst mental overload caused by situations as increased time on task, extremely fast speakers, unintelligible speakers, and long working hours '. It s~ems to occur rather automatically, without being noticed by the performrng interpreter,

Department of Behavioral Science, Swiss Federal Institute of Technology, ZUrich/Switzerland

References

Ader, R., Felten, D.L., & Cohen, N. (Eds.) (1991). Psychoneuroimmwwlogy (2nd 00.).

San Diego, Ca.: Academic Press.

Avons, P., Garthwaite, P., Davies, H.L., Murgatroyd, P.R -: , & James.' W.~.T. \198:8).

Approaches to estimating physical activity in the community: Cal?r:memc v~datiO; of acrometers and heart rate monitoring. European Journal of Clinical Nutrition, 4_,

185-196.. . d

Bacik, H. (1971). A description of various types of omissions, additlons,an errors

encountered in simultaneous interpretation. MEIA, 16, 199-210.

Boucsein, W. (1992). ELectrodermal activity .. New York: Plenum Press.,

Buse, L., & Pawlik, K. (1996). Ambulatory behavioral assessment and: ,m-fie1d,pelf~~ mance testing. In J. Fahrenberg, & M. Myrtek (Eds.), AmbulatO'ry asse:ssrnent (pp. _.9

50). SeattletW A: Hogrefe & Huber Publishers.

Cannon, W.B. (1932). The Wisdom of the Body. New York: Norton... . - _

Dienstbier, R. A. (1989). Arousal and physiological toughness: Implicanoas for mental

physical health. Psychological Review, 96, 84-100.- . . .' . ~

Ericson, K.A., & Simon, H.A. (1980). Verbal reports as data. Psydwloglcal ReVI£!II', 8/,

215-251. .

Frankenhaeuser, M, Lundberg, U., & Forsman, L. (1980) .. Dissociatlo~ re:ween sympa-

thetic-adrenal and pituitary-adrenal responses to an achievemeet srtuation character-

ized '0)" high contmllability: Comparison between T)-pe; Pi.. ,··.amd 1i;}'PC'. B ,males ;;m,d

females. Bioiogical. Psychology, 10, 79-91. . e : _ •.

FreudeIibergcr, R.J., & Ricaelson, G. (1980). Bum-oilL. The lurgih. CtJ.5,(lJj Jng!h aC.II:tel'f-

ment. Gardea City, N.Y.: Anchor Press.

248

HANS ZEIER

Gaillard, AW.K. (1993). Comparing the concept of mental load and stress. Ergonomics

36, 991-1005. '

Gerver, D. (1974). The effects of noise on the performance of simultaneous interpreters:

Accuracy of performance. Acta Psychologica, 38, 159-167.

Goldst~i~: ~.S. ~1~87). Stress-induced activation of the sympathetic nervous system.

Bailliere s Clinical Endocrinology and Metabolism, 1, 253-278.

Henni~, J. (1?94). D~e psychobiologische Bedeutung des sekretorischen lmmunglobulin A tm. Speichel. Munster/New York: Waxmann.

Hess, W.R. (1948). Die funktionelle Organisation des vegetativen Nervensystems. Basel:

Schwabe. .

Jones, J.W. (Ed.) (19"82). The burnout syndrome. Park Ridge, Ill.: London House Press. Kirschbaum, -C:,'1§'& ~ellhammer, D.H. (1989). Salivary cortisol in psychobiological

research: An overview, Neuropsychobiology, 22, 150-169. .;1\"

Klonowicz, T. (1994). Putting one's heart into simultaneousinterpretition. In: S: Lambert & B. Moser-Mercer (Eds.). Bridging the gap: Empirical research in simultaneous

interpretation (pp. 213c224). Amsterdam: John Benjarnins. '

Kugler, .J., Hess, M.,.& Haake, D. (1992). Secretion of salivary immunoglobulin A in relation to age, saliva flow, mood states, secretion of albumin, cortisol, and catecholamines in saliva. Journal of Clinical Immunology, 12, 45-49.

Levine, S., & Ursin, H. (1991). What is stress? In M.R. Bown, G.P. Koob, & L.c. Rivier (Eds.), Stress, neurobiology, and neuroendocrinology (pp. 3-21). New York: Dekker.

Mancini, G., Carbonara; A.O., & Heremans, J.F. (1965). Immunochemical quantitation of antigens by single radial immunodiffusion. Immunochemistry, 2, 235-254.

Monat, A, & Lazarus, R.S. (Eds.). (1991). Stress and coping: An anthology (3nd ed.).

New York: Columbia University Press ..

Moser, .B ". (1978). Simultaneous interpretation: A hypothetical model and its practical application. In D. Gever & H.W. Sinaiko (Eds.), Language, interpretation and communication (pp. 353-368). New York: Plenum Press.

Mos~r-Mercer, B., & Kunzli, A (1995). Human strategies for translation and interpretanon. In L. Dreschler-Fisher & S. Pribbenow (Eds.), KJ-95 Activities: Workshops, posters, demos (pp. 304-306) .. Bonn: Gesellschaft fur Informatik.

Moser-Mercer, B., Kunzli, A, & Korac, M. (1998). Prolonged turns in interpreting: '. Effects on quality, physiological and psychological stress. Interpreting, 3(1): 47-64:

Myrtek, M., Brugner, G., Fichtler,A., Konig.K; Muller, W., Foerster, F., & Heppner, V. (1988). Detection of em cit ion ally induced EEG changes and their behavioral correlates: A new method for ambulatory monitoring. European Heart Journal, 9 (Supple-

ment N), 55-60. .

Paradis, M, (1994). Towards a neurolinguistic theory of simultaneous translation: The framework. International IournalofPsycholingistics, 10, 319-335.

Perrez,M., & Reicherts, M.(1992). Stress, coping, and health. Seattle: Hogrefe ec Huber

Publishers. . J:~

Perrez.M; & Reicherts, M. (19~6). A computer-assistedsclf-monitorjsg procedure for assessing stress-related behavior under real life conditions. In J. Fahrenberg & M. Myrtek (Eds.), Ambulatory assessment (pp. 51-67). SeattleIW A: Hogrefe & Huber Publishers.

PSYCHOPHYSIOLOGICAL STRESS REsEARCH

249

Searleman, A., & Herrmann, D. (1994). Memory from a broader perspective. New York:

McGraw-Hill.

Selye, H. (1950). Stress. Montreal: Acta Inc., Medical Publishers. Selye, H. (1974). Stress without. distress. Philadelphia: J.B. Lippincott.

Singer, M.T. (1974). Engagement-involvement: A central phenomenon in psychophysiological research. Psychosomatic Medicine: 36, 1-17.

Van Rood, Y.R, Bogaards, M., Goulmy, E., & van Houwelingen., H.C. (1993). The effects of stress and relaxation. on the in vitro immune response in man: A metaanalytic study. Journal of Behavioral Medicine, 16, 163-181.

Vining, RF., McGingley, RA., Maksvytis, J.1., & Ho, K.Y. (1983). Salivary cortisol: A better measure of adrenal cortical function than serum cortisol, Annals of Clinical

Biochemistry, 20, 329-335.

Zeier, H. (1989). Complementary functions of the two brain hemispheres: Comparisons with earlier conceptions and implications for indiviOOal and society, In:re:maJioNll

Journal of Neuroscience, 47, 175-179.

Zeier, H. (1992). Psychophysiologische StressjorschlUlg. Bern: PanlHanpt,

Zeier, H. (1994). Workload and psychophysiological stress reactions in air ·traffic control-

lers. Ergonomics, 37, 525-539.

Zeier, H., Brauchli, P., & Joller-Jemelka, H.I. (1996). Effects of work demands on immunoglobulin A and cortisol in air traffic controllers. Biological Psychology, 42,

413-420.