Haylitt Retief (English draft version), Amsterdam, 20 February 1974
Some methodological considerations
Like all basic science psychology is concerned with establishing functional relations among important variables. But in psychology nearly all variables have a high degree of abstraction. To the extend that a variable is abstract rather then concrete we speak of it as being a construct. Examples of important constructs in psychology are intelligence, anxiety, dive level, neuroticism.
Essentially constructs represents hypothesis that variety of behaviors will correlate with one another in studies of individual differences or will be similarly affected by experimental treatments.
It is important to realize that all theories in science mainly concern statements about constructs, rather then about specific observable variables.
Yet, it is only through observable variables that these statements can be verified. The observable variables are suggested measures of the construct and it must be kept in mind that they have to be validated for that purpose.
Specifying the domain of observables which relate to the construct is a major aspect of validation. Another important aspect is determining to what extend these observables are affected a like by similar experimental treatments. It must also be determined if the observables "act" as though they measure the construct. This is done by relating the observables to a variety of treatment variables in different experiments.
The assumption in these experiments is that the treatment variables truly represents different levels of the construct. If the observables act as though they measure the construct they must follow the treatment variables in each experiment in a predictable manner.
When a variety of measures have similar curves of relationship with a variety of treatment variables representing a construct it becomes meaningful to speak of them as measuring the construct.
The measures that most consistently behave as the majority of measures can be said to have the most construct validity.
Mental load.
In psychology "mental load" has the statues of a construct.
It seems useful to distinguish it from such construct as "stress" or workload in general, since these may also relate to purely physical load ( requiring bodily effort). Yet, there are many situations were physical load is entirely absent, but the organism is clearly under some kind of burden, as may be observed from a variety of bodily reactions.
Mental load is expected to be present when a talk requires information handling or when the situation entails phenomena such as social pressure, time pressure, conflict and danger.
The observables assumed to be related to the mental load construct include physiological measures. Among these, sinus arrhythmia has been claimed to be especially sensitive for load involving information handling ( ).
Other observables variables require some kind of task performance.
The binary choice task ( ) has been mentioned in this respect.
In the following pages a number of experiments will be described in which proposed measures for mental load are validated. To strengthen the assumption that the experimental treatments were finally different in mental load, rather extreme conditions had to be created. This is difficult to realize under laboratory conditions which resulted in doing the experiments in real life situations. An additional advantage of this was that the feasibility of applying the measurement technique in practice (which is their ultimate purpose) could be assessed at the same time.
First the experiments involving physiological variables will be described. Performance measures will be treated afterwards.
Validation of physiological measures.
Three physiological variables are investigated, viz. heart rate (HR), heart rate irregularity or sinus arrhythmia (SA') and respiratory rate (RR).
Sinus arrhythmia has been proposed as a variable with special sensitivity for mental load involving information handling ( ).
Hence one aspect of the experimental problem was to investigate if the dependent variables differed in sensitivity with regard to differentiating power between treatments. It was hypnotized that SA would be more sensitive to treatments involving information handling and HR to treatments involving tension (from danger), with information handling minimized.
It may be pointed out that physiological reactions to "stress" in general are very similar, irrespective of the type of stressor. They follow the pattern of the general adaptation syndrome (Selye). Some authors have demonstrated specific stressor related reactivity for certain physiological variables ( ) but the evidence is far from conclusion. This research is also complicated by the fact that most individuals have a personal physiological reaction pattern, which in itself may not be constant ( ).
The importance of developing stressor related measures is however generally agreed upon.
The experiments incorporating information handling are done in flight simulators. For comparison a experiment with parachute jumpers (attention, no information handling) is included.
Experiment 1.
The experiment was done in a DC-7 flight simulator (KLM). The effect of flying maneuvers varying only slightly in degrees of difficulty, on the depended measures was studied.
Material and methods.
The flying maneuvers in order of supposedly increasing difficulty were:
1) continuous straight and level flight on AUTO - PILOT;
2) continuous straight and level flight MANUAL.;
3) straight and level flight between TURNS;
4) rate one TURNS;
5) ground controlled approach.
"Rest phase" outside the simulator was used as reference. Only heart rate and sinus arrhythmia were measured.
Results
Both variables differentiated significantly between the flight phases ANOVAR F values were 7.12. for HR ( p < 0.001) and 3.69 for SA ( p <0.025 ). However, it was obvious from the material that these significancies were achieved by the extreme positions of the approach as the most difficult phase.
Values for the other phases were not in a predicted order, suggesting that the variables were not sufficiently sensitive to differentiate between phases that varied only little in degree of difficulty.
Experiment 2.
In the same simulator an experiment was done which incorporated more difficult flying procedure. A realistic flying training exercise was flown, consisting of the Rotterdam Instrument Overshoot Procedure with rejoining the ROT HOLDING PATTERN followed by standard approach on instruments (ILS).
Samples of the physiological variables were taken during four different flight configurations which are typical in terminal area instrument flying.
In order of supposedly increasing workload: 1.level flight (LF), one minute stretch with the aircraft steady on course towards the holding pattern; 2. holding (H0), sample starting overhead the beacon until leaving the beacon for ILS; 3. take-off (TO) one minute sample commencing at forward push of throttles; 4. approach, sample starting at ILS glide past needle first movement, until touchdown. A rest phase outside the simulator was used as reference (RP).
Results.
The relationship between the physiological variables and the flight phases was as predicted as seen by the figures 1 and 2
FIGURE 1
ANOVAR F values are 37.57 for HR and 18.78 for SA (both p < 0.001)
The stronger differentiating power of SA is best expressed by the W2 criterion ( ) which was 0.32 for SA compared by 0.18 for HR. In this experiment two pilots carried out nine flights in three different days.
Experiment 3.
The experiment was conducted at the Governmental Flying School ( RLS), with 14 air cadets, two sessions each.
Main differences with previous experiments: a simpler type of flight simulator was used( AT-10) beach craft configuration). The holding pattern was left out. Pilots were younger and less experienced. Respiratory rate (RR) was also measured.
Results.
The figures 3,4 and 5 shows the relationship between the physiological variables and the flight phases.
FIGURE 3. FIGURE 4.
Conclusions.
1. The 3 physiological variables differentiate significantly between the sample flight phases,
2. The differentiating power of RR is greatest followed closely by SA.
The differentiating power of HR is not so strong, though still significant (p < 0.005).
3. The change in the physiological variables from one phase to another is generally in the predicted direction.
Summary.
The 3 previous experiments were conducted to validate the sensitively of heart rate, heart rate irregularity, and respiratory rate as indicators of mental load. A type of mental load was selected which was assumed to constitute mainly of information handling. For this purpose experimenter controlled flight-simulator flights' were chosen. In these flights more and less difficult procedures were incorporated. It is concluded that when differences in mental load are very small ( such as between straight and level flight and simple turns- rate 1- ) differentiating is inadequate. When differences are more pronounced ( such as between level flight, holding and let down) differentiating is significant and generally in a predicted direction.
The highest discriminating power according to the W2 criterion is shown by respiratory rate, followed closely by heart rate regularity and to a lesser extend by heart rate itself.
Experiment 4.
Variation of heart rate irregularity, heart rate and respiratory rate as a function of mental load consisting largely of perception of danger.
In a previous research ( ) heart rate irregularity had been proposed as a measure for mental load caused primarily by information handling (binary choice tasks). In the experiments with flight simulators (mainly information handling) this proposal was verified and to some extend corroborative evidence was given.
However, it was also found that heart rate reacted significantly to the type of mental load used in these experiments.
When defining mental load distinction is made between mental load caused by information handling, emotion and tension ( ). Heart rate irregularity had been proposed as the typical measure for mental load by information handling.
It follows that when mental load is caused by "emotion" and/or tension, with information handling minimized, one should expect heart rate irregularity to be less reactive.
To verify this assumption an experiment was conducted were mental load consisted of various levels of "tension", information handling being minimal.
It was hypnothized that in this experiment heart rate would have a greater differentiating power between load levels than SA and RR, with the magnitude of W² as a criterion.
Material and methods.
To achieve differences in mental load created mainly by tension and with information handling minimal, parachute jumping was used as a stressor. The (dependent) physiological variables were HR SA, and RR ( on ground only ).
The independent variable had 4 levels, consisting of situations found in parachute jumping. The levels, in order of supposedly increasing mental load were : 1. rest value, relaxed at the ground at the end of the day; 2. rest value, relaxed on ground, before jumping; 3. registration in aircraft before subjects jump run; 4. registration during subjects jump run.
A subjective ratting of the 4 situations on a seven points scale was obtained after the final registration ( ).
A high degree of experimental control was achieved in this experiment, notably in standardization of sample periods ( e.g. same stimulus always preceding and terminating period, and irrelevant variables being similar in all samples).
Twelve subjects took part, average age 21.1 years, each was measured approximately on 3rd and 4th jump.
Results.
figure 6 shows HR and SA as a function of mental load. ANOVAR values are given in table 1.
Figure 7 shows the curve of the subjective scala.
FIGURE 6. FIGURE 7.
TABLE 1
'===================
F P W²
'-------------------------------
HF 123,95 0,001 0,71
SA 36,86 0,001 0,46
'===================
Table 2 gives ANOVAR values for comparison of ground values only ( e.g. before jump and after jump on ground ). This comparison includes RR.
TABLE 2.
'===================
F P W²
'-------------------------------
HF 34,32 0,001 0.20
SA 1,52 n.s.
RR 0,04 n.s.
'===================
P.M. correlation coefficient between subjective scale values and HF and SA are 0,97 and - 0,88 respectively. Only the HF correlation is significant (p < 0.025).
Conclusions.
The results show that HF is the most sensitive parameter for mental load caused by tension only. This conclusion is supported by 3 significant observations 1. the higher W² value ( Table1 ).
2. only parameter differentiating between ground sessions only ( Table 2 )
3. only parameter with significant correlation with subjective scale.
Comparison of the results with the flight simulator experiments shows reverse order, which was predicted on the assumption that sinus arrhythmia is especially sensitive for mental load involving information handling.
A summary of physiological validation experiments.
Our research, so far, suggest a selective sensitivity of heart rate for mental load caused predominantly by tension and of SA and RR for mental load involving information handling, such as cockpit-workload. All these variables tend to show some change in a similar direction with increasing mental load however and it is to early to consider them as definite criteria. The strong relationship between SA And RR in information handling mental load, could make the registration of RR redundant when heart rate is measured anyway.
(When supposed differences in mental load are small, differentiation with physiological measures in our present experiment has not been successful).
Binary choice performance and mental load.
To obtain an indication of the mental load implied in the performance of a task subjects may be asked to perform a secondary task at the same time. It is suggested that the performance on the secondary task is a measure for the so called spare mental capacity. This in turn is obviously related to the mental 'Load implied in the performance of the primary task.
Bourneman (1942,) Poulton (1950), Brown (1962), Haider (1962), Michon (1966), Polfe (1966,1969) and others have used this method. In the laboratory for Ergonomic Psychology research is done with the binary choice generator in this way. Subjects are required to perform the binary choice task as a secondary task. In some of the following experiments the primary task is not a task in the true sense of the word, but some special treatment which is also supposed to effect mental capacity The effect of the treatment on mental capacity is thus quantified . The unit of measurement is the binary choice response. Sample values are determined by subtracting incorrect responses multiplied by 2 from the number of correct responses. Since the task is a typical information handling task it was thought to be less abstract to refer to information handling capacity, rather than mental capacity.
Experiment 1
Radar controllers workload as a function of pilots workload.
Purpose of the experiment was to investigate 'he relationship between air traffic controllers workload and pilots workload. The experiment is of the explorative type, seeking to establish if the BCG task could be used in the radar controllers situation to measure spare information handling capacity.
The experiment was conducted at Schiphol airport (Amsterdam) under VMC at night.
Material and methods
A VMC aerodrome traffic circuit pattern was flown six times by an aircraft, chartered for the experiment.
The aircraft was under radar control with the pilot on instruments following the controllers instructions.
It was assumed that the radar controllers workload would be different in selected flight phases.
The flight phases can thus be regarded as the experimental treatments.
They included level flight . (LF), take off (TO) and approach (AP) in order of supposedly ; increasing workload for the radar controller (and the pilot). A reference phase with the aircraft on the ground was included.(R).
The dependent variable was the radar controllers BCG task performance. In this experiment the stimuli were high or low tones. The required response was to press a pedal under the controllers left foot when the tone was low and a pedal under his right foot when the tone was high.
Results.
Results are shown in figure 1.
The controllers BCG performance is clearly related to the flight phases in the predicted
manner.
Discussion.
The controllers performance with the aircraft in level flight is better then in the reference phase (R). There may be two reasons for this. Firstly it is pointed out that with the aircraft in level flight the controller has practically nothing to do ( their are no other aircraft in the circuit). Secondly, the reference phase was recorded before the experiment. It must be assumed that learning has the effect of improving BCG performance in the later stages.
It is felt that the experiment has demonstrated the possibility of using the BCG as a possible workload measure for air traffic controllers on the job.
Experiment 2
The experiment is of the explorative type seeking to establish if the BCG task could be executed during and if it gives an indication of the pilot space information handling capacity as a workload measure.
Material and Methods.
Two pilots flew the Rotterdam Overshoot Procedure (2 times each) in a DC-7 flight simulator (see experiment 2 of the physiological experiments).
The BCG was used in the oral mode, the tones emitting from the pilots headset.
Required responses were pressing one of two little buttons spaced approximately 1 cm apart on a small strip fixed on the left top side of the control column, close to the place were the column is normally held.
Samples of BCG performance are taken in 5 experimental conditions (flight phases) supposedly increasing in mental load. The conditions are similar to those in the second experiment validating Physiological variables and the characters on the horizontal axis have the same meaning.
RESULTS
Results are given in figure 2.
FIGURE 2
The general trend is in the predicted direction, except when comparing take off and approach.
Discussion.
It is felt that the possibility of using the BCG as a workload measure quantifying space information handling capacity during the flight task has been demonstrated. The curve of relationship between treatment and dependent variable is very similar to the trend of the physiological variables as in experiments 2 and 3. The discrepancy between the values for take-off and approach also occurred between the physiological experiments 2 and 3.
Experiment 3.
Purpose of the experiment was to investigate the effect of mental load, without information handling, on information handling capacity. It was also intended to demonstrate how a rather isolated situation is made accessible for psychometric measurement by using the BOG with LDRS unit. The experiment is hypotheses testing with regard to the effect of the mental treatment.
Material and methods
14 Firemen had to perform the BCG task on top of a 90ft. ladder. 4 Treatment conditions were distinguished viz.
1) after the experiment.
2) before starting the climb;
3) after resting 1 minute at 45 ft.;
4) after resting 1 minute at the top.
The mental load was considered to consists of tension caused by being more or less suspended in free space on top of the ladder. The order of increase is self explanatory.
The dependent variable was BCG performance with the LDSR-unit used on the ladder ( ).
RESULTS.
Results are given in figure 3.
It is obvious that the treatment has no effect on BCG performance. No further statistical analyses was considered necessary.
Discussion.
Going strictly by the experimental results it must be concluded that mental load caused by tension alone has no effect on information handling capacity.
The author is however, tempted to conclude that in this particular case the treatments were not providing effectual different intention.
The firemen climbed the ladder in a carefree routine like manner and the whole atmosphere was relaxed rather than tensed. It is the author's experience that in experiments where the tension is real (meaning an element of realistic danger is present) all individuals on the scene are tensed and not only the subjects. The condition is rather like a public holding !!! breath in a circus watching the flying trapeze. When the experimenter is breath holding himself, he has a real stressor.
Experiment 4.
Purpose of the experiment was to investigate the influence of mental load without information handling on information handling capacity.
It was also intended to demonstrate how the BCG with LDSR-unit could be used as a research tool in an important research field, notably the adaptation of man to submarine conditions with extremely high pressures. This type of research is done in wet recompression tanks where most conditions of extremely deep dives can be simulated.
In this field a great demand exists for on-line human performance evaluation tools, since performance is the ultimate criterion deciding the acceptability of working conditions ( when certain physiological values are not exceeded). Especially when cerebral malfunction may be suspected ( as is the case in hyper baric research), refind performance criteria give the earliest symptoms ( ).
Material and methods.
A wet recompression tank was used tested to an equivalent of 450 ft of water depth (Figure..... ).
cylindrical tank has a diameter of 11 ft. The lower part contains 10 ft of water, separated from the dry section by a metal grid with an opening through which the diver can descend.
In the dry part, the medical staff and responsible diving officer are seated on benches fixed against the cylinder wall.
They can walk around on the meta1 grid covering the wet part.
In the dry and wet part port 'wholes allow observation from outside.
The binary choice generator was placed outside the tank, in the tank control room. It was also connected to a pen-writer giving an immediate and continuous record of BCG performance, notably response interval times ( ).
Performance was also scored on electromagnetic tape.
Experimental treatments consisted of three pressure levels, viz. 0 ft., 210 ft., and 270 ft.
It was the first time in the Netherlands, human subjects were subjected to pressures in access of 150 ft. and it was decided that if BCG performance would be clearly deteriorated at the 210 ft-level no further increase in pressure would be made.
This discision would be made by the physician in the tank control room, who was operating the BCG.
The senior physician was, supervising inside the tank but allowance had to be made for a possible incapacitation at these high pressures with which we had no experience. Hence the responsibility shift at the 210 ft with the BCG performance as common sense criterion.
Three subjects performed the BCG task : 1) The senior diving officer, also responsible for the divers safety 2) A medical orderly; 3) The author.
Subjects 1 and 2 were in the dry compartment, subject 3 was submerged until after the 210 ft. samples had been taken and in the dry at the 270 ft level.
This was a safety measure since it was not known that each individuals reactions would be under a pressure of ten atmosphere (air).
RESULTS
Figure gives information handling capacity as a function of pressure level.
The individual performance traces are shown on a sheet at the end of this paper.
DISCUSSION
The results clearly show a performance deterioration with increasing pressure.
Yet it is also clear that the change is not dramatic, suggesting adequate information handling capacity at all levels. The results do not show that at the higher pressure levels subjects possibly had to concentrate much more on the task.
The authors subjective experience was at the 270 ft level he had to put an extreme effort into concentration, with a funnel vision, focused on the light stimulus.
The experiment has demonstrate the usefulness off using BCG with LDSR- unit for getting an on - line performance indication of a submerged diver to personnel outside the pressure area.
Experiment 5.
Purpose of the experiment was to investigate if the BCG with LDSR-unit could be used to measure the effect of mental load without information handling, on information handling capacity. The mental load consisted of a dangerous condition, created at 60 ft depth on the North sea bed. The experiment also served to demons-irate how an isolated situation was made accessible for psychometric measurement with the BCG with LDSR-unit.
Materials and methods.
A navy ship with thirteen frogman was taken out to sea for the experiment. In this experiment, strictly speaking, mental load of the tension type is only assumed to be truly present in the dangerous condition.(E2). This "dangerous condition" consisted of taking out the mouthpiece of the breathe apparatus at 60 ft, for a period of 30 seconds, subjects being relaxed at the bottom. The danger element is mainly that if a subject has difficulty in putting back the mouthpiece,(e.g. swallows water, starts coughing), he still has 60 ft of water above him, and he is already out of breath. Under the prevailing experimental conditions visibility was approximately 1 ft, resulting in total darkness at the bottom. Control for the dangerous condition-was a one minute sample, taken a short time before subject removed the mouthpiece ( El ) .
Conditions El - E2 replicated at ships deck as controls for the effects of submersion. Thus Cl was a sample in a relaxed position shortly before the dive and C2 under same conditions but with the mouthpiece removed and breath holding. In order of supposedly increasing load the conditions are ranked Cl - C2 - El - E2. Subjects were reasonably experienced frogman, average age 23 years. Each subject descended alone to the bottom were the LDSR-unit was placed. All other apparatus was situated at the ships deck.
Results.
Figure 5 shows the effect of experimental treatments on information handling capacity.
It is clear that information handling capacity decreases in the predicted direction.
Figures 6. E2 and C2 show an example of computer processing of data (response interval
times).
Figures 6. E2 and C2 show an example of computer processing of data. Left
plot is 30 seconds with mouthpiece removed at bottom (E2), next plot is 30 seconds breath
holding at ships deck (C2). Each interval time between responses is expressed as per
minute equivalent.
It is shown that this particular subject worked irregularly in the dangerous condition (control comparison on the right).
DISCUSSION.
The experiment shows that the BCG can be used for psychometric measurement in the submarine environment. It also shows an effect on information handling capacity in this environment, and more so in the critical condition. It should be pointed out however that, although protective gloves were worn under all conditions. water temperature ( + 5OC) may still have influenced manual dexterity. This could have influenced performance as measured, but not necessarily information handling capacity as intended.
Amsterdam, 27 February 1974.
Experiment 6.
Purpose of the experiment was to investigate if the BCG with LDSR-unit could be used to measure the effect of mental load with minimum information handling on information handling capacity.
The mental load consisted of a dangerous condition:" Free escape " from an underwater habitat at the bottom of a lake at 40 ft.
The experiment also served to demonstrate how a condition that was totally in accessible for psychological observation, let alone psychometric measurement, was made accessible for psychometric measurement with the BCG with LDSR-unit.
Material and methods.
Figures show underwater habitat used for the experiment. Shore facilities consisted of large army tents were the subject-divers were looked after (catering etc.) Medical facilities included standard emergency equipment, a recompression tank and two physicians for medical supervision.
The' electronic apparatus(BCG with cardiomod for electronic storage of performance) was placed in a boat moored at a buoy above the underwater habitat. A speed boat was available for transport between this boat and the landing side. Telecommunication was provided between the habitat, the boat with the apparatus, and the shore via a radio-van. The author conducted the experiment from inside the habitat and, as a senior diver, was responsible for the divers safety inside the habitat until the moment they actually left the habitat shaft for free escape. During "free escape" subjects were accompanied by a "body" who had been waiting outside the habitat.
Procedure. Two subjects would leave the shore (150 ft. from habitat) after a call from the habitat. They would swim to the buoy, collect the LDSR-unit from the boat and then descend to the habitat below. One subject entered the habitat took off his scuba-equipment and was instructed. After this a one minute practice with the LDSR -unit was given followed by one minute registration (CB). After this subject would descend in the habitat shaft, with the LDSR-unit in the left hand, when settled, he would commence free escape after clearance from the experimenter. He was instructed to operate the LDSR-unit during the actual period of free escape (F/E).
The condition of "free escape" is characterized by a change in pressure (100% in this case ) and of course the absence of external air. Subjects are constantly exhaling to compensate for the reducing pressure. Under the conditions of the experiment feasibility was approximately 5ft., meaning that the subjects were ascending "in the green" for the greater part of the trajectory.
Results.
Figure shows effect of free escape on information handling capacity, with registration
in the habitat as a reference.
Figure shows individual performance traces of 16 subjects.
This type of registration shows the occurrence of performance "blocks" (large response interval times).
Comparison of choice making capacity in the habitat with free escape shows a reduction of 78 % during free escape. Inter- individual differences were extreme, ranging from 8% to 100% reduction. It appeared that these differences were more or less related to the degree of experience of the various subjects.
Discussion.
The experiment has demonstrated that the BCG with LDSR-unit can be used to measure ( spare) information handling capacity during such complex conditions as free escape. The data provide a comparison between conditions and also between individuals.
To our knowledge this is the first time a method had been used to quantify an important aspect of cerebral functioning during this exercise.
Free escape has a distinct place in naval training programs of many countries, notably in relation with sub-marine personnel. The considerable inter-individual difference in performance found in this experiment indicate that the instrument could be used as a selection device or as a tool with which the degree of proficiency in this exercise could be established.
Binary choice performance and distraction stress.
In the preceding experiments it was demonstrated how the BCG could be used to measure the effect of experimental treatments(mainly differing in mental load) on information handling capacity.
Purpose was to validate (spare)information handling capacity as defined as a measure for mental load.
Refined measures for mental load could be useful in many situations. One possible important application would be to use them as a criterion when decisions have to be made with regard to optimal design of displays or control units. However, if differences are only small, the measures used would have to be extremely accurate, and a high degree of construct validity would be required. It may be a long time before this ideal state of affairs is reached with regard to measures for the mental load construct.
Decisions regarding optimal display or control unit designs, may also be based on performance on the relevant task. A problem with this strategy is that the subjects effort in performing the task is not measured. This may invalidate performance as a workload criterion.
One way of improving the state of affairs is to put the subject in a difficult condition. His "spare" mental capacity is absorbed. In this critical condition small differences in system design may now be mote clearly reflected in performance, favoring the better system.
Various techniques are used to make conditions critical. In aviation. the aircraft may be placed in an awkward position for approach and it is verified which system gives the best final approach parameters,
The binary choice generator can also be used to make conditions critical. This way of employing the binary choice generator has been called "the method of distraction stress" ( ). In these experiments the BCG is used in a paced condition. In paced condition the frequency with which tones or lights are presented is determined by the experimenter. If the subject fails to respond within the time depended on the pre-selected per minute frequency, an ed omission is registered and the next stimulus presents.
In the Laboratory for Ergonomic Psychology of the Organization for Health Research TNO experiments have been carried out in flight simulators to demonstrate the possibilities of the distraction stress method for problems of aircraft cockpit lay-out. In one experiment it was demonstrated that the use of "a flight director" resulted in superior flying under distraction stress. Without the use of distraction stress the superiority of the flight director could not be demonstrated (other forms of workload that are, normally present were minimized).
Experiment 7.
Purpose of the experiment was to demonstrate the use of distraction stress to improve the validity of performance as a workload criterion. In a DC-7 flight simulator instrument approaches with and without flight director system were compared. The superiority of the flight director system can be taken for granted and it could therefore be used as a criterion: flying performance should be better with flight director and this would be more obvious under distraction stress.
Material and methods.
Distraction stress. The BCG was used in a aural mode, paced condition (30 tones/min). Subject had to respond by pressing push buttons on the left topside of the control column. When an error or mission was made, a light on the dash-panel would illuminate. Only one error per minute was allowed. If more errors were made, subject had to overshoot.
Flight performance. The flight performance criterion was obtained by averaging the total variance of speed and optimal horizontal and vertical approach path together. This results in one figure for each condition. The larger the figure the poorer the flying performance.
Subject were 3 junior airline pilots. Each made 8 approaches, equally divided over two sessions.
Results.
Figure 4 shows flying performance with and without distraction stress under the two conditions as described.
It is shown that distraction stress causes a deterioration in performance both for flight with and without flight director. However for flights without the flight director the situation is much more pronounced without flight director the flight performance error value increases by 97% under distraction stress. With flight director this error value increases only by 36.5%. Also, performance without distraction stress is slightly better without flight director. Only with distraction stress the flight director proved to be superior. Since this in already known, this finding illustrate the significance of the method
Thus both findings support the hypothesis that under distraction stress the superiority of the more advanced system, as apparent from flying performance, is more obvious.
FIGURE 4.
= Distraction stress
FD = With flight director
NFD = Without flight director
return to: Start page - return to: International page
Copyright protected ©
