36.2 Review of Selected Ergonomic/Learning Environment Studies

36.2.1 Introduction

The aim of this review is to summarize what constitutes ergonomic research as applied to educational facilities, and thus is intended to be representative and not comprehensive. In making my selection of ergonomically representative studies, I have chosen (1) two user preference survey studies that related student responses on a rating scale-type questionnaire to the physical measurements of environmental and display features; (2) an experimental study of social interaction patterns with different classroom seating arrangements employing television as an observational tool; (3) an experimental study investigating the effect of photometric brightness contrast on student preference, attention, visual comfort, and fatigue; and (4) an experimental study on display legibility that explored qualitative differences between front- and rear-screen projection. The first study reported is one of my own, and the other four were doctoral dissertations that I supervised at Boston University between 1976-91. By providing such a sampler, it is hoped that the reader will acquire some insight into representational methodologies in educational ergonomic research, as well as an awareness of the substance of their findings, which I believe hold considerable significance today relative to learning environment design and utilization.

36.2.2 Environmental and Ergonomic Features in Educational Facilities: Two User Preference Studies (McVey, 1979; Bethune, 1991)

36.2.2.1. Rationale for the Studies. Every year, millions of dollars are spent on the construction and renovation of educational facilities that are often inadequate for both students and instructors. Students and faculty alike frequently complain that classrooms are too hot or too cold, that they have uncomfortable seating, or that the seating location does not permit clear and accurate viewing of the room's display systems. Additional complaints relate to the difficulty in hearing lectures because of excessive internal noise or because of excessive sound reverberation within the room. Other problems are more subtle and frequently result in complaints like "My eyes seem to hurt after a lecture," or "I just can't seem to concentrate for very long in that room," or "I'm just not comfortable in there. I don't like the room."

These classroom conditions and complaints point out the need to determine why such facilities fail to meet the objectives of facility planners and architects. One reason proposed for this failure is the source material for the guidelines used in the planning and design of educational facilities. Discussions with numerous architects indicate that most rely almost exclusively on architectural standards such as those found today in Ramsey and Sleeper's Architectural Graphic Standards (1988), or those published by BOCA (Building Officials and Code Administration International, 1990), and rarely adopt or refer to long-standing and widely available ergonomic standards and references such as contained in the publications of Woodson (198 1), Woodson and Conover (1973), Bennett (1977), Van Cott and Kinkade (1972), and Grandjean (1969, 1987).

To determine the relative efficacy of these two different sources of planning and design information, in either their earlier or current editions, the assessments of two college student populations, one from a large midwestern university, and the other from a large eastern university, were recorded approximately 20 years apart using two slightly different versions of the same questionnaire. In spite of the passage of time and changes in the student population and the schools' curricula and the increased sophistication of available educational technology, and although the statistical tools employed and the analysis procedures differed, the results of both studies were strikingly similar. Given the similarity of the research methodology employed in these two studies and their results, they will be presented here as a two-part case study.

36.2.2.2. Method (1973 Study). The first study conducted at a large midwestern university in 1973, and reported in 1979 (McVey, 1979), was basically a posttest- only comparison of the assessments of a static group of college freshmen, sophomores, and juniors (N = 214) who, during semester 1, were assigned to three popular lecture halls that had been constructed in accordance with guidelines found in standard architectural handbooks, with a comparable static group (N = 289) assigned to a lecture hall constructed in accordance with guidelines found in ergonomic handbooks and other guidelines developed from the author's own human factors literature search (McVey, 1969) and verified by in-house laboratory experiments. The user assessment instrument consisted of a Likert-type scale (questionnaire) made up of 59 measurement items related to specific interior environmental factors, 10 items that tested face validity, and 10 subject identification items. The 59 measurement items were divided into 10 categories: seating, desks, acoustics, audio systems, visual display systems, lighting, color and reflectance, and two "other" considerations.

36.2.2.3. Rationale for Employing a User Assessment Methodology. My reasons follow for employing a combination of questionnaire (with a rating scale) to solicit student assessments of their classroom environments, with an analysis of those findings in light of actual physical measurements taken in the environments being assessed:

One approach gaining popularity is the utilization of the users themselves as evaluators. Armed with instruments ranging from simple attitudinal scales and various modifications of the semantic differential originally developed by Osgood, Suci, and Tannenbaum (1957) to the more complex Guttman scales (Markus, 1974) and the various adaptations of the multitrait-multimethod model originally proposed by Campbell and Fiske (1959), researchers have evaluated such wide-ranging environmental settings as "school study areas" (Sommer, 1968), "landscaped offices" (Boyce, 1974), and "low and high use housing" (Francescato et al., 1975). Through these and other studies, strong justification for employing users as evaluators of their environments has emerged (Canter, 1975; Lee, 1973; Preiser, 1970; Wools, 1970), as has support for using the questionnaire to record this evaluation.

Canter (1970) notes:

Using a questionnaire is one stage towards getting the user to set up hypotheses about the effect of the physical environment and to explain his interaction with it... . The investigator does not pressure to understand or to hypothesize the nature of the mechanisms by which the subject deals with the physical environment, but rather to get the subject to show how satisfied he is with the functioning of the environment in which he is (p. 14).

Discussing the legitimacy of this measurement approach as opposed to the more traditional, physiological response recording methods, Sommer and Becker (197 1) state:

A psychologist can take the position that a check mark on a scale indicating dissatisfaction, particularly when the respondent has no incentive to falsify or distort his reply, is just as legitimate a basis for remedial action as a physiological measure.... Our results make clear that psychologists must deal with organisms or environments separately (p. 416).

Other researchers believe that while the questionnaire can reveal important information about the efficacy of an environment, much can be gained by using a design that also allows the researcher to relate the user's subjective assessment to specific causes, i.e., the physical variables inherent in that environment. For example, asking students to rate a desk's design in terms of "How well does it support accurate and comfortable note taking" will reflect on such physical variables as the size, height, and inclination of the desk being evaluated.

An approach along this line was used with success by the Building Performance Research Unit of the United Kingdom to assess school buildings, one major basis for comparison being a building performance profile called a psarchigraph (Patterson & Passini, 1974). A similar multi-method approach also has been used to evaluate acoustic experience in concert auditoria (Hawkes & Douglas, 1970). It is theorized that such an approach should make it possible for the researcher to relate subjective effects to physical causes, thus improving the predictability of subjective experiences from physical data. Such information, on the face of it, would seem to be of considerable value to architects and facility planners as they make decisions regarding the design of new construction or remodeling projects.

36.2.2.4. Validity, Reliability, and Analytical Measures (1973 Study). The following statistical data were obtained for each variable in each class: arithmetic mean, standard error of the mean, standard deviation, unbiased variance, coefficient of skewness, coefficient of kurtosis, and .05 confidence interval for the mean. One-way analysis of the variance also was conducted for each group across the two semesters and the results of that analysis expressed as analysis of variance tables. Other data included statistics relating to the validity (UWMACC Factor 2 Program) and stability of the questionnaire (Hoyt Reliability Index), individual student seat location, and anthropometric data regarding the student population taking part in the study.

36.2.2.5. Results and Discussion (1973 Study). In the 1973 study, the completion of the questionnaire took an average of 25 minutes, with some students taking as long as 35 minutes and others only 15 minutes. Examinations of the distributions indicated that the 5-point rating scale (5 = exceptionally good, 1 = unacceptable) yielded considerable variance, since every point on the scale was used. Results found that students gave statistically significantly higher ratings (p < .05) to Room 204, the lecture hall that had been designed and constructed in accordance with ergonomic recommendations, than they did for the three designed and constructed in accordance with published architectural standards. Overall mean student ratings for each of the major categories are shown in Table 36- 1.

When reporting the findings of this study in 1979, comparisons were made between the overall mean scores per category received by the ergonomically derived lecture hall and the overall mean scores received by the three architecturally derived lecture halls. These data supported the study's hypothesis.. However, because of space limitations in that publication, it was only possible to report on the overall student responses to the main environmental and display system design features in the rooms. In order to determine the student's individual responses to specific room characteristics, it was left to the reader to look at each of those responses while cross-checking those ratings with the room photographs and the 60-item physical descriptor developed for each room.

Given the importance of specificity in design, that study has now been revisited by its author in order to compile a set of empirically defensible lecture hall design guidelines. These generalizations follow.

• Given free choice in seat location and given the design features of an ergonomically derived room layout where viewers are located at distances between 2 and 6 times a projected image width and at lateral locations not greater than 60' off screen axis, students will rate all seat locations equally high.
• Maximum acceptable viewing distances are determined not so much by the apparent size of the screen but by the legibility of the materials (words, captions) displayed. Acceptable boundaries of the horizontal viewing sector are determined more by the ability of the display system (projector, lens, and screen) to deliver an adequately bright image than by the amount of trapezoidal distortion caused by oblique viewing.
• Sight lines with inclination angles greater than +250 (to the top of the display) and depression angles greater than -10' (to the bottom of the display) will be criticized by college students. Rear projection can be more effective than front projection when projection screen type, size, and location are carefully selected and coordinated with the room's viewing sector.

• Illumination levels of 30 foot-candles from either incandescent or warm-white fluorescent will satisfy a student's note-taking requirement. Cool-white fluorescent illumination will require 50 foot-candles in order to achieve as satisfactory a rating as the above-mentioned lighting relative to the same student activity.
• Matte-finished, warm-colored walls (off-white, parchment, buckskin) and light-colored furniture (brown, tan, cream) will produce a more visually comfortable environment than specular finished dark-colored walls, even when the illumination level of the latter is double that of the former.
• While incandescent downlighting is well received by most students, some of those who wear glasses will report distraction and discomfort. The degree of this response will be affected by ceiling height, fixture design and spacing, and the overall brightness in the room.

• College students are extremely sensitive to noise intrusion, either from outside student traffic or from inside mechanical, electrical, or heating and air-conditioning systems.
• An ambient noise level of NC10 will be perceived as being too low in terms of masking unwanted sounds, while ambient noise levels of NC30 and above will interfere with students' ability to hear clearly the unamplified speech of a lecturer.
• Amplified audio response systems will be well received by students in large lecture halls (4,000 SF +) but will not be as important in the smaller halls (1,200-2,000 SF) as will be room acoustics.

• In terms of physical comfort, ease of access and egress, a sense of personal space, and for book and coat storage, students will prefer separate swivel seats and fixed desks (counter) in a seating layout providing them with 7.5-8.0 SF over fixed nonswivel chairs with tablet arms in a layout providing 6.3 SF or less per station.
• In rating the following design features of chairs, students will give the highest ratings to the features shown in bold print: floor-seat pan height (17", 15", 16"), seat inclination (3°, 0°, 10°, 17°), back inclination (10°, 20°, 22°), lateral spacing (25-31", 20", 21 frontal spacing (48", 32", 28", 34").
• A linear span of 28"-31" of 18"-deep and inclined (15°) continuous counter writing surface will promote more accurate and comfortable note taking during lectures and visual presentations than will flat or moderately inclined (2°-7°) movable tablet arms offering between 81 and 116 SF of writing surface.

• When attending classes having the following thermal factors, students will prefer the combination of temperature, relative humidity, and air velocity as noted in bold print: Room A: (82'F, 15% RH, front: 0-75 fpm, rear: 0-280 fpm, ), Room B: (82'F, 29% RIJ, 0-75 fpm), Room C: (69.5'F, 47% RH, 0-25 fpm), and Room D: (770F, 29% RH, w/o AC: 0-10 fpm, ,w/ AC: 0-75).
• Students will complain of drafts when air velocities exceed 75 fpm.

36.2.2.6. Method (1991 Study). A second user assessment study-a doctoral dissertation by James Bethune employing a modified version of the questionnaire-used in the 1973 study was conducted in 1991 at a large eastern university to determine whether today's college students would also prefer educational facilities constructed on the basis of ergonomic design guidelines over those constructed in accordance with standard architectural references. Freshmen, sophomore, junior, and senior students (N = 145) evaluated the four lecture halls in which they regularly attended classes. Using a 5-point Likert-type rating scale questionnaire consisting of 152 items, organized into 46 major interior environmental factors, the students rated each item as it existed in each of the rooms. The results of this activity were then analyzed in an effort to determine the validity of current architectural standards and recommendations as applied to the design of lecture halls.

A comparison was then made in the 1991 study between what the students found acceptable and the current architectural standards in the design and construction of lecture halls. A similar comparison was then made between the student's ratings and the recommendations found in published ergonomic sources. A t test, developed by Kruskal-Wallis and Dunn (Dunn, 1964), and simultaneous confidence intervals were used to analyze statistically the student's responses and to validate their significance.

36.2.2.7. Results and Discussion (1991 Study). The 1991 study built on the work of the 1973 investigation and extended the user assessment techniques used in that study in order to evaluate the efficacy of applying ergonomic recommendations to the design of lecture halls. The four lecture halls used in the 1991 study were all built or remodeled within the last 10 years and therefore reflected current architectural practices. .

A comprehensive set of measurements was taken of each architectural feature in each of the four rooms, and each item was categorized as being either in agreement with architectural standards or ergonomic guidelines. Modifications were made to the original questionnaire by eliminating a number of audiovisual display system items not present in the current rooms and by setting up four cross-comparisons between rooms. This resulted in 38 individual test items and a total of 152 evaluation factors. The questionnaire was distributed to each of the 145 students taking part in the study. The statistical analysis first tried to determine what a student's response should be to a hall that was acceptable both in terms of architectural standards and ergonomic recommendations.

This resulted in a mean score of 3.554, with a standard deviation of 0.411 (p < .01). A Kruskal-Wallis test and Dunn's method were used to show that the results from the four different lecture halls were statistically related, and "simultaneous confidence" intervals were used to compare the responses within a specific question. The evaluation concluded that the results both between the individual halls and the individual questions were reliable at the 95% confidence level.

Table 36-2 shows the mean responses to the individual items in the questionnaire. It was concluded that any response of less than 3.000 indicated student dissatisfaction, as 3.000 is greater than one standard deviation from the 3.554 expected satisfactory mean response. The same mean satisfaction level of 3.000 can be applied to the data in Table 36-1 from the 1973 study, showing a statistical similarity in student's assessments of their classroom environments despite the 20 years between the studies.

The results of the evaluation showed that of the 152 measured factors, students agreed with ergonomic recommendations 82% of the time. For 103 of the factors, the agreement was positive; that is, the factor was in agreement with ergonomic recommendations, and the student's mean assessment was within the satisfaction level. For 22 of the factors, students found specific room features to be unacceptable, a position consistent with ergonomic sources but not with architectural standards. There were 16 cases where students found specific factors acceptable which were not supported by the ergonomic sources. These factors were primarily related to seat spacing. Interestingly, students accepted frontal seating closer (34") than recommended by ergonomic sources (40") for seating comfort during notetaking, but rejected that-same spacing when evaluating it for ease of access and egress. Similarly, 11 items related to lateral seat spacing that were within ergonomic recommendations (24") were found to be too close and therefore unsatisfactory.

The results of the 1991 study confirmed the conclusions of the 1973 study and serve to point out that when it comes to the design of lecture halls, there is still a discrepancy between architectural standards and ergonomic recommendations, differences that are noticeable to college students and who for the most part find greater agreement with the ergonomic guidelines than with the architectural standards. Students were invited to make written comments throughout the questionnaire. These qualitative responses in general supported the quantitative results presented above, as well as most of the findings of the 1973 study.

Some of the interesting student responses included the following:

• Only the small students or those of average height registered complaints with the seating. But the lack of response from the very tall or large students was apparently due not to their satisfaction but simply to the resignation they had developed over the years with the unsatisfactory seating to which they were continuously subjected.
• The combination of tight lateral and frontal seat spacing resulted in very low scores for the tablet armchairs relative to book and coat storage. Frequently, this situation combined with poor tablet arm design led to books and notes spilling onto the floor when students tried to enter or leave their seats.
• While Nick H was rated acceptable as a movie theater, its use as a lecture hall was very poorly received by the students. This result speaks to the frequent failure of multiuse designs for lecture halls, as well as the unwise assignment of such spaces as large-group classrooms by college administrators.
• Provisions for left-hand tablet arms were absent in all of the rooms, used in this study. This omission was duly noted in the students' evaluations.

36.2.2.8. Conclusion. Both of these studies, although performed approximately 20 years apart, indicate the firm and continuing existence of a strong student preference for lecture halls designed in accordance with existing ergonomic guidelines than for those designed in accordance with standard architectural references and building codes. Consequently, it is recommended that now and in the future, facility planners and architects make every effort to utilize existing ergonomic guidelines and standards in their educational facility design, construction, and remodeling efforts. Many of these important guidelines and standards will be found in section 36.3. Others can be found in existing ergonomic handbooks and in the growing number of journals that are directing their focus toward ergonomic applications.

While the two studies just presented raise a number of interesting questions about the current state of educational facilities design, their results also generate some concerns relative to the current preparation of architects, inasmuch as valuable ergonomic information continues to be ignored in their work product. One of the questions that surfaces is: In the face of such evidence, why haven't architectural programs adopted the ergonomic guidelines that have been available for more than 20 years? A second question is: Why have architects ignored past research findings that clearly have demonstrated the benefits of designing educational facilities in accordance with ergonomic guidelines?

These questions go beyond the scope of this study, but they are clearly implied by the study's results. A third question raised by the study is: In light of the fact that both studies clearly showed that college students are reliable and objective evaluators of their learning environments, why are they not consulted more often in educational facilities planning and evaluation? While these questions merit considerable discussion, one can conclude from the findings of both the 1973 and 1991 studies the following recommendations:

• Educational facility planning and architectural design references and standards should adopt time-tested and proven ergonomic guidelines.
• Architects should seek out the observations and comments of college students in determining the successful and unsuccessful elements of their educational facilities before making final determinations as to the specific design features to be contained in new and proposed facilities.
• Architectural educational programs should include the study of ergonomic principles and guidelines.

36.2.3.1. Background and Rationale for Study. In the mid-60s, the work of E. T. Hall and specifically his book The Hidden Dimension (Hall, 1966) advanced the science of "proxemics" and the concepts of "socieopital" and "sociefugal" spaces, i.e., those spaces that promoted human interaction and those that promoted noninteraction. This work followed the acceptance of earlier works on social interaction by Bass and Klubeck (1952), Steinzor(1950), and Leavitt (1951). It was also during this period that the sociologist Robert Sommer published research findings(1959, 1965, 1967, 1969, 1970, 1974) that strongly supported and justified specific seating arrangements to promote desired patterns of social interaction. It was in this context that the following study was rationalized and conducted-

Fulrath (1976, p. 23) states:

Increasingly, educational and business facilities are being designed with spaces for small-group work. Schools, libraries, offices, and hotels have conference areas designated for, or classrooms that accommodate, both small-group work and media use. Typically, the furniture arrangement of the classroom is still that of the straight row, though rectangle arrangements prevail in offices. However, these arrangements may not be best for the individual, especially in view of increased media use, if they contribute to a negative response condition by the individual, either behaviorally or attitudinally. Other arrangements, such as the circle/hexagon, may better help to facilitate a positive response set, particularly where media tasks are concerned. The design implications of this issue concern future use of furniture arrangements that facilitate the optimal response of the user. In studying user responses to these varied arrangements, the question of whether to design predominately for comfort or for interaction may be partially answered.

Similarly, use of the small group as a problem-solving mechanism is common in both education and business, as is the attendant use of media for information transfer. In the examination of the relationship of time to these factors, the results of this study may provide some specific answers about appropriate limits to task duration and involvement for the individual. That information would, in turn, augment existing knowledge about the appropriate utilization of media and small-group work.

36.2.3.2. The Research Question and Method. This study focused on two environmental variables and their effects on selected behavioral and attitudinal. responses of the learner. The specific purpose was to determine the effects of seating arrangement and task duration on fatigue, attention, participation, and preference of individuals engaged in small-group media tasks. Circle, rectangle, and straight-row patterns constituted the three levels of the independent variable seating arrangement. Similarly, three levels of the independent variable task duration were I hour, 2 hours, and 3 hours. There were three hypotheses in the study:

1. There would be significant differences in fatigue and participation as task duration increased from I to 3 hours.
2. There would be significant increases in fatigue, attention, and participation between people in the circle seating arrangement and those in the rectangle or straight-row patterns.
3. There would be significant increases in preference for a seating arrangement between people in the circle pattern and those in the other arrangements.

The sample consisted of 54 adult students from various schools in the university who were noontide majors. Subjects were randomly assigned to six small groups of nine members each, and two groups each were then assigned to one of three seating arrangements.

The task used in the study was a 3-hour, visual-verbal media task in which groups rated 35-min slide images according to criteria of visual design. The measurement instruments for fatigue included the Pearson Fatigue Checklist (Pearson, 1956), and a Visual Discomfort Evaluator developed by the researcher based on the work of Hultgren, Knave, and Werner (1974). Videotape recordings were used to assess attention and participation; and preference was measured by Mehrabian's Approach-Avoidance Test (Mehrabian & Russell, 1974). Testing occurred at four time intervals of the study, including the start of task, first hour, second hour, and third hour. The resulting data were scored analyzed by multifactor analysis of variance procedures.

Figure 36-2. below shows the layout of the experimental setting used in this study.

Figure 36-2. Layout of experimental room used in Fulrath study.

36.2.3.3. Results and Analysis of Findings. Results confirmed the first hypothesis and revealed significant increases in fatigue between the start of task and the first, second, and third hours (p < .001). Similar results were recorded for three symptoms of visual discomfort as the incidence of tired eyes, sore eyes, and headache increased significantly between start of task and the third hour (p < .05). However, no significant differences in discomfort were found for five other symptoms (itchy eyes, watery eyes, sandy eyes, blurred vision, and double vision). Attention increased significantly between the third hour of the task and each of the other task times (p < .001),.but here were no significant differences in participation.

These results held in all groups regardless of seating arrangement, and, with one exception, no significant differences were found between groups for either the second or third hypotheses. The single exception was that attention in the straight-row arrangement was significantly greater than that recorded in the other two patterns (p < .01). Additionally, the use of blink rate as an index of fatigue and the relationship between fatigue and performance were examined and found to confirm the reservations that some behavioral researchers have found with this device as a reliable index of fatigue.

Several specific conclusions based on the general acceptance of the first hypothesis were formulated from the results of the study. The conclusions also reflect the clinical observations video recorded during testing. Those conclusions led to several recommendations, including two primary ones. It was recommended that, when using media tasks with. groups of adults, optimal task duration should not exceed I hour without a break, or 2 hours if rest breaks are used. Study results also suggested that a single 15-minute break after 1.25 hours might foster changes in arousal levels and would be more conducive to continued task performance. Given media use, the straight-row arrangement may be preferred if attention to the screen were most important, but the rectangle may be preferred if attention to the group, or group interaction, were most important. Though equal to the other two patterns, the circle/hexagon arrangement could neither be recommended or rejected as being superior based on the results of the study.

Secondly, the Pearson Fatigue Checklist (Pearson, 1956) was found to be a valid, sensitive instrument, and its use was recommended for fatigue studies that would investigate other task types, task durations, learning situations, and age groups. It was further recommended that such studies also investigate the nature and pattern of the fatigue response as well as the psychophysical law to which it conforms. Finally, the study confirmed the value of television recording as a valued tool in the analysis of student behavior.

36.2.4.1. Background and Rationale for the Study. In the 1950s and early 1960s, Domina Spencer (1954 ) and her associate at M.I.T., Parry Moon (1961), created a stir in academic and engineering circles with their studies regarding luminance contrasts that attempted to quantify the patterns of photometric brightness that promoted comfortable and accurate viewing through enhancing figure/ ground separation and three dimensionality of objects in interior environments. At the same time, Darrell Boyd Harmon's monograph The Coordinated Classroom (1951) related this work and that of others in the illuminating engineering profession to the classroom settings of elementary school children.

Two decades later, the lighting designer William Lam (1977) took major steps toward quantifying the perceptual aspects of lighting in school and office environments. Around the same time, LaGuisa and Perney of the Illuminating Engineering Society's Research Institute conducted their research of supplementary illumination on visual displays (1973, 1974) and discovered that attention was. sustained longer and distractions reduced when classroom charts were illuminated in excess of the surround. Thus it was only natural that educational researchers in the field of media and technology would look toward establishing, through their own research, operational guidelines for establishing appropriate contrast ratios in media-related, rooms. If such luminance contrast ratios could be verified, they could then be adopted with confidence in future educational facility design practices.

Such was the background of the DesRosier's study. Surprisingly, this study remains one of the only experimental studies to deal specifically with photometric brightness contrast ratios and remains uncited in the literature, even though influential organizations such as the Illuminating Engineering Society of North America and the Human Factors and Ergonomic Society regularly include similar luminance ratios in their published Standards and Recommendations. One would expect that those organizations could find their recommendations strengthened by referencing the DesRosiers' study (Fig. 36-3).

36.4.2.2. Method. One aspect of the visual environment, the photometric brightness contrast ratio (BCR) between a projected image and its surround, was investigated in two experiments. The nature of these two experiments was established by DesRosiers to include the four important features called for by Chapanis (1965): (a) controlled observations in (b) an artificial situation with (c) the deliberate manipulation of some variables in order to answer (d) specific hypotheses. In describing the experimental setting DesRosiers states:

The experimental setting was essentially the same for both experiments I and 2. The room dimensions were 17'X 17' X 11'. The main feature of the room was a vinyl rear projection screen 10' X 15' which served as a translucent wall dividing the room into two smaller rooms, one for projection, one for viewing. Portions of the rear projection screen were designated for specific tasks:

Visual task, the centrally located portion. 26" X 17" on which the test slide presentation image was projected. Distractor, that portion 8" X 12" on which the distractor slides were projected. The position of this screen was 45' left of the center of the test screen. This distance was chosen based on earlier research findings. Surround, the remaining portion of the rear projection screen was transilluminated by supplementary lighting. The rear projection screen used had a gain of 120% at 0'. Projection area, the inner portion of the laboratory, 11.5' X IT, served as the projection area. Test presentation system two carousel slide projectors, an audiocassette tape recorder with sync pulse capacity, a dissolve unit, a buffer relay system, a dimmer to control image brightness, and a rectangular image frame, a device to ensure that all illumination from the projectors (and only illumination from them) reached the portion of the screen where the visual task (image) was projected. The slide presentation system was controlled by a custom power supply and relay system to ensure that all luminances were as specified.

The videotape observation recording system consisted of a low-light sensitive video camera, a camera adapter, and a TV monitor. Forty one-half-inch 30-minute videotapes were used. The audio-grid is an instrument devised by DesRosiers to ensure consistent reading of the videotapes after the test period. This system required an audiocassette player and an audiocassette with oral cues to announce the change of slides. When played in sync with the videotape, the investigator could know what slide was being projected at any moment of the videotape. The student station consisted of a chair-desk combination positioned two screen widths from the screen. A dimmer for controlling the brightness of the image was positioned on the student's desk. The luminous environment of the viewing area was defined by identifying the range of the illumination on the subject and on the desk. A photo research illuminance meter and 2' luminance meter were used in the measurements.

36.2.4-3. Problem Statement and Methodology of Experiment 1. The first experiment attempted to answer these questions: What are the image/surround brightness contrast ratios preferred by students? How do these correspond to the task/surround brightness contrast ratios recommended by the literature? The currently recommended task/surround BCRs for conventional tasks suggest the hypothesis that, given a specific surround brightness, students will select an image brightness no less than that of the surround (1: 1), and no greater than 10 times that of the surround (10: 1). In a controlled laboratory environment, 14 high school students, given the surround brightness levels of 2, 6, and 20 footlamberts (FL), were asked to select their preferred image brightnesses. The mean brightness levels preferred were established, and the resultant BCRs were calculated.

36.2.4.3.1. Results of Experiment 1. AT 20 FL, the students selected a BCR of approximately 1: 1 ratio; at 6 FL, a 3:1 ratio, and at 2 FL, an 8:1 ratio. Since the BCRs chosen were within the 1:1 to 10:1 range, there was strong indication given that the general recommendations for task/surround BCRs are equally applicable to the projected image/surround BCRs. Furthermore, students chose a similar image brightness (17 FL) regardless of the surround brightness. As a result of this stable choice, a recommendation of 15 ± 2 FL was established as a "brightness task requirement for projected images." It should be noted that this recommendation is approximately the same as that recommended by the Society of Motion Picture Engineers (Kloepfel, 1969) for motion pictures for more than 3 decades.

36.2.4.4. Problem Statement and Methodology of Experiment 2. Current literature has suggested that the root of problems like inattention, visual discomfort, and visual fatigue encountered in media environments seems to lie in the characteristics of the visual display system and how it interfaces with its viewers, that is, the viewing angles, the luminance contrast between the display and its surround, etc. Experiment 2 dealt with this problem and investigated the effect of image/surround brightness contrast ratios on attention, visual comfort, and visual fatigue. DesRosiers states:

Thirty-seven students between ages 14 to 19, assigned randomly to one of four experimental groups, viewed a slide presentation individually at an image/surround BCR of 40:1, 10:1, 3: 1, or 1: 1. The footlambert levels for these were 20:0.5, 20:2, 20:6.7, and 20:20, respectively. The Kruskal-Wallis One-Way Analysis of Variance by Ranks was applied to see if any relationship existed within the four BCRs. Each student was tested for attention by videotaped observation, for visual comfort by subjective evaluation using a checklist, and for visual fatigue by the two objective measures of critical flicker fusion (CFF) and threshold and eye blink rate recorded by videotape.

36.2.4.4. 1. Results of Experiment 2. The hypothesis tested was: When the brightness contrast between image and surround is high, attention is greater, visual comfort is lower, and visual fatigue is greater. The Mann-Whitney U test for k independent samples was applied to the combined attention scores at 40:1 and 10:1 and to the combined scores at 3:1 and 1: 1, yielding results that permitted acceptance of the first part of the hypothesis (that when BCR is high, attention is greater).

Results of the subjective evaluation of visual comfort gave some evidence that the second part of the hypothesis is true, that comfort is lower when BCR is high. Because of problems with the experimental process and the instrumentation used to measure visual fatigue (CFF and eye blink), there was no way of concluding whether or not visual fatigue was greater at high BCRs. It may be concluded that when the image/surround BCR is high, attention is greater and visual comfort is lower. It was also found that a "contamination effect" may occur when CFF tests and eye blink counts are employed in close succession.

In discussing her rationale for using the Mann-Whitney U test, DesRosiers stated:

This test is used in order to determine whether the two independent groups have been drawn from the same population. One of the most powerful of the nonparametric tests, it is most useful as an alternative to the parametric t test when the researcher wishes to avoid the t test's assumptions (Siegel, 1956). The U test does not require that data be normally distributed or that sample variances be equal. It calls for a nominal independent variable and an ordinal dependent variable (Tuckman, 1972).

36.2.4.5. Implications of Study. In her study, DesRosiers concluded that:

This study pointed out that often visual discomfort is the cost paid for a media environment conducive to high attention. Research is needed to determine a brightness contrast ratio which is the best compromise between these competing elements. The findings in this study have direct application to designers and users of media facilities, especially in preparing for the projection of slides, filmstrips, and motion pictures with rear-screen projection for high school students. They apply less directly for front-screen projection and for a broader range of audience than the target population in the current study. They have possible implications to designers and users of facilities for any visual presentation, including nonaudio visual setting, since the structural patterns of brightness used by architects are basically the same for all designs and environmental settings.

At the time of DesRosiers' study (1976), VDT use was in its infancy relative to educational environments, and so it is understandable that DesRosiers did not relate her findings to lighting and viewing conditions in VDT workstations. If provided that opportunity today, there is no doubt that given the validity and reliability of her findings, she could and would apply them to such educational environments with a high degree of confidence.

36.2.5 The Accuracy Recognition of Positive and Negative Symbols on Front- and Rear-Projection Screens Under Self-Selected Illumination (Hamilton, 1983)

36.2.5.1. Rationale for the Study. This study represents a response to two basic questions most educational media specialists were asking themselves during the early 1980s: How can I be sure that I am producing legible instructional materials? Should I use front- or rear-screen projection to display those materials? Adding confusion to the issue was the fact that most educational media guidelines available to the media specialist specified print size by dimensions such as 1/8", 1/4", etc., without any consideration of the distance at which those materials would be viewed. And second, there was considerable disagreement among information display specialists as to which was the preferred medium for information display, rear- or frontscreen projection. In an effort to provide other media specialists with some "hard science" for guidance relative to making such choices, Mark Hamilton (1983) decided to investigate both sides of this issue and include as a modifying factor the se If-selection of illumination level. By doing this, Hamilton anticipated discovering luminance-illuminance interrelationships that would affect performance and possibly serve as a guide for establishing ambient illuminance levels for future media presentation spaces. In noting his motivation, Hamilton stated:

This investigation originated from recent visibility-orientated research and concern for possible misconceptions regarding the use of front and rear screens for projecting positive and negative symbols.... The study of projected visual images in teaching is one area in which there has been a move toward more systematic research. The increased use of projected images has amplified the potential benefits to be gained from good message design and optimum projection practice and has magnified the negative effects of inadequate materials and poor projection. An examination of published standards and research findings, as well as current classroom practices, show widespread confusion and disagreement on what should be the proper level of various factors of the projection environment.

36.2.5.2. Problem Statement This study was designed to determine the effects of projection screens, symbol polarity, and symbol size on accuracy recognition of projected symbols. The dependent variable was accuracy recognition. Three independent variables were: film-based projection systems, consisting of front and rear screens; symbol contrast, consisting of positive (white symbols on black) and negative (black symbols on white); and subtended visual angle, consisting of 7, 5, and 3 minutes of arc.

36.2.5.3. Procedure. Hamilton describes his procedure as follows:

One hundred sixty college students with tested visual acuities ranging from 20/40 to 20/10 and including those with normal or corrected-to-normal vision were individually presented 81 randomly ordered letters for a total of nine slides displayed in a laboratory specifically set up for this purpose during evening sessions at a small private New England college. Subjects were randomly assigned to one of four treatments, i.e., front screen with positive symbols; front screen with negative symbols; rear screen with positive symbols; and rear screen with negative symbols, each having symbol sizes of 7, 5, and 3 minutes of arc. Subjects attempted to. correctly identify the self-paced stimulus materials. Additionally, subject's preference for a general illumination level and responses were recorded. Data from the 2 X 2 X 3 factorial design were subjected to a three way analysis of variance. Where significance at the .05 level was evident, individual differences between means were examined with the Newman-Keuls test.

36.2.5.4. Apparatus, Description, and Specification of Stimulus Materials, Display System, Lighting System. An excerpt of Hamilton's lengthy and precise description of his experimental setting and controls follows (Fig. 36-4):

The photographic equipment used throughout was a 35-mm, single-lens reflex camera (Canon AEI), with a 50-mm macro lens (0.5). The exposure used was f/1 I for I second for the positive materials and f/8 for 4 seconds for the negative materials. The letters were originally derived from a Kroy lettering machine using the Helvetica Regular (24 point) disk. Letters were graphically laid out on grid paper to provide uniform spacing. The single sheet of letters and sample Es were then photographically reduced on a stat to a uniform height of 10 turn. When this appropriate size was attained, the letters were reduced by 30, 50, and 70% to produce symbol sizes of 7, 5, and 3 millimeters.... To eliminate slide "popping" due to heat, Gepe glass slide mounts were used. Each test slide, in the horizontal format, was projected on a front or a rear screen at an appropriate distance (12 feet) to provide the specified subtended visual angle of 7, 5, and 3 minutes of arc for the symbol. Projected image brightness for each of the test slides was 18 footlamberts. Two multiple slide sets were produced to control potential bleaching of the slides due to exposure of the projector lamp.

A single Kodak Ektagraphic AF-2 slide projector with a Kodak 100-150-mm f/3.5 flat field lens with a 300-watt ELH lamp was used in the tests. Screen luminance was determined by projecting stimulus slides on the projection screen and measuring it with a photometer calibrated in footlamberts. Test instructions and procedures were pre-taped and synchronized on a cassette tape recorder to provide uniform presentation of test procedures. The range of illumination intensities available at the work plane was approximately 0 to 50 foot-candles. General illumination Was derived from two table top lamps, each fitted with one 150-watt bulb, positioned behind and to either side of the subject. One commercial, 100-volt (600 watts maximum) rheostat dimmer, wired to the 115-volt, 60-cycle electrical source, was utilized to provide a range of 0 to 50 foot-candles at the work plane. The dimmer was contained in an enclosure and located to the right of the subject to allow easy selection of room light for each of the four treatments. The rheostat dimmer was preset to 0 before each subject entered the testing room.

In the remaining portion of his lengthy description of the experimental setting and controls, Hamilton provides specifications of the measuring instruments and display system characteristics, and describes his measurement procedures.

36.2.5.5. Research Design and Data Analysis. The dependent variable for the study was the number of judgments made on the criterion task, which consisted of one self-paced recognition accuracy task of symbols projected via 35-mm, 2 X 2 slides. Subject responses were judged either correct or incorrect, and therefore ratio data were obtained from the tests. The number of correct recognitions made by the subjects were computed and analyzed. There were three independent variables manipulated in this investigation. The first independent-variable, film-based projection system had two levels: (1) front-projection screen and (2) rear-projection screen. The second independent variable, direction of symbol contrast, also had two levels: (1) positive (W/B) projected symbols and (2) negative (B/W) projected symbols. The third independent variable manipulated in the study was subtended visual angle. This variable had three levels: (1) 7 minutes of arc, (2) 5 minutes of arc, and (3) 3 minutes of arc.

The basic design for the study was a true experimental design, Design 6, the Posttest-Only Control Group Design (Campbell & Stanley, 1963). This design was utilized because it can be delivered to students or groups as a single natural package and eliminates the awkwardness of a pretest. This design was also used because it permits an evaluation of treatment effects upon the criterion task while minimizing the effects of confounding variables (Issac & Michael, 1981).

The Posttest-Only Control Group Design consisted of a 2 (film-based projection system) X 2 (direction of symbol contrast) X 3 (subtended visual angle) factorial design with repeated measures on the third factor. The use of this design enables the investigator to assess interaction between the three independent variables (Ary, Jacobs & Razavieh, 1979). A three-way analysis of variance of the criterion task scores was applied to ascertain the presence of differences between treatments (Ary, Jacobs & Razavieh, 1979; Campbell & Stanley, 1963). Results with a statistical significance at the .05 level were considered demonstrative of reportable treatment effects. Kerlinger (1973) suggests the .05 level as adequate, being "... neither too high nor too low for most social scientific research" (p. 170). The Newman-Keuls multiple comparison test was used to compare means on the criterion task scores after analysis of variance had been performed if the results indicated its appropriateness. The Newman-Keuls test is an a posteriori comparison test and is employed when the investigator intends to make all possible simple pairwise comparisons among means if a significant overall F ratio is obtained in the analysis of variance (Howell, 1982).

The range of brightness patterns created in the testing environment for the present investigation was caused by the reflectance of the walls, ceiling, and floor surfaces, as well as the reflectance of screen areas adjacent to the projected image. Figure 18 illustrates the visual field of the present investigation, as seen by the subject during projection of stimulus materials. Specific target areas were preselected by the investigator, and footlambert readings were taken at each to determine a specific brightness pattern selected by the subject. All testing was conducted during the evening to eliminate external light sources from the testing environment. The 10 preselected target points are numbered.

36.2.5.6. Results and Discussion. Hamilton offers the following in his Results and Discussion section:

1. Given the relatively ideal viewing conditions and display systems used in this study, film-based projection systems, i.e., front and rear screens, of similar performance did not significantly affect the accuracy recognition of either positive (W/B) or negative (B/W) projected symbols of either 7, 5, or 3 subtended arc minutes. In fact, for all the single and multiple interactions with the other independent variables, there were no significant effects revealed in the accuracy recognition scores (F = 0.44).

2. For 7 minutes of arc little difference was found between the accuracy recognition of positive (W/B) symbols (94.6%) and negative (B/W) symbols (98.7%). For 5 and 3 minutes of arc, negative symbols (B/W) were more accurately recognized than the positive (W/B) symbols (p <.01).

It should be noted here that when determining the symbol sizes to be used in this study, it was assumed that any size symbol less than 10 arc minutes would result in accuracy recognition reductions, and that these losses would be directly proportional to the amount of reduction. This study proved this to be true, with 100% accuracy scores recorded in the field test conducted prior to the start of the study employing a symbol size of 10 arc minutes, and the following results recorded during the Hamilton study with the three smaller symbol sizes (see Table 36-3).

3. In the self-selection of ambient illuminance when displaying white symbols on the front screen, 31% preferred a light level between 0-1 FC, while the remainder of the test population were evenly distributed (about 15% each) in their selections of the following levels: 1-5 FC; 5-15 FC; 15-30 FC; over 30 FC. In the self-selection of ambient illuminance when displaying black symbols on the front screen, there were no majority preferences for any of the 5 foot-candle ranges and were found to be fairly equal (20-25%) at the various foot-candle ranges. Only at the 1-5 FC range did subject preferences drop to 9%. One interpretation of these findings is that when subjects view white symbols on a black background, projected on front screen, they prefer low illumination. On the other hand, when subjects view black symbols on a white background, projected on a front screen, they prefer a wide range of illumination.

4. In the self-selection of ambient illuminance when displaying white symbols on the rear screen, subject's selections for ambient light conditions were fairly evenly distributed for each of the 5 foot-candle ranges, but with a slight reduction in preference for the 5-15 FC range. In the self-selection of ambient illuminance when displaying black symbols on the rear screen, a majority of the subject's selections (60.8%) fell on or with the 0-5 FC range, with the remainder equally spread over the range of 5 to over 30 ranges.

In explaining these results, it should be remembered that black symbols projected on a white background produce a bright image and hence a brighter room than do white symbols projected on a black surround. As a result of conditioning or believing that the existing light in the room was sufficient, subject preferences did not group around the higher foot-candle ranges. In fact, in treatment 4, 24% of the subjects selected no additional illumination at their workplace. It should also be noted that this finding might have been different had the subjects been required to take notes during their period of participation.

5. The range of luminance contrasts for the target area and its near and far surrounds as expected were higher when projecting black symbols on a white background than when projecting white symbols on a black background. This was true for both front- and rear-screen display, with front-screen projection having more variance and slightly higher luminance levels recorded than rear screen. However, in all cases there were no luminance ratios that exceeded 3:1.

There is a need for educators, architects, and ergonomists to continue the kind of research sampled in this section. User assessments have yielded important findings that have proved valid and reliable. Current research has applied a modified version of the instrument used in the McVey (1979) and Bethune (1991 studies in the evaluation of music education multimedia workstations. Preliminary results from this study appear to support the findings of the two earlier studies (Badolato, 1995).

But more importantly, there is a critical need to conduct the kind of epidemiological research that was conducted by Harmon and Bennett decades ago and unfortunately not repeated since. We need to find out how our young students are being affected by the learning environments in which they are expected to dedicate increasingly more time in VDT workstations and carrels. Are their maladaptations to the current nonergonomic facilities simply creating surmountable stress and fatigue? Or are they being exposed to conditions that threaten their normal growth and development, due to the fact that their physiological and sensory systems are yet fully developed? It is my own personal belief that we educators are sitting on a time bomb in this regard. And while substantive and well-sponsored research in the field of office design has produced corrective designs to mitigate if not eliminate repetitive motion disorders, no such mandate has yet been directed toward the learning environment. Hopefully, some concerned and well-positioned educational leaders will discover this author's quiet alarm signal and respond accordingly.