Notes
Early Comparative Office Study
Kevin Manning, Seattle University, 1975
Author Note:
With assistance by a grant from the National Science Foundation
Abstract
Post-Industrial design of the Workplace has experienced several epochs in the last one hundred and fifty years. Driving these changes were a variety of factors, e.g., intuition of the designer, monetization of furniture manufactures, etc. Little hard empirical, behavioral data existed to guide office design. In 1975, the author attempted to provide this guidance, with help from his University (Seattle University), Robert Sommer of UC Davis and a grant from the National Science Foundation. The author conducted a large (and perhaps the first) comparative statistical study across four distinct office/production environments. This study was conducted at Unigard Insurance’s headquarters in Bellevue, Washington with a study population of 145 employees. Data from this study showed the factor “intrusive noise” to be the only statistically significant condition aggravating all office employees. Moving forward forty-four years, most new studies have found the same to be true in newer “Open”, "Agile" or "Activity-based" office design environments.1
Keywords: Open Office, Agile, Activity-based office design, Burolandshaft,
Early Comparative Office Study
This submission to the 50th Conference of the Environmental Design Research Association (EDRA) is intended to satisfy EDRA’s desire to showcase early scientific studies commemorating its half-century of existence. Therefore, this study is not intended to reveal earth shattering conclusions, but to offer evidence of early attempts to use Behavioral Science techniques in support of office design.
Background
As an eighteen-year-old high school graduate, I was hired as a mail clerk at Safeco Insurance Company in Seattle, Washington during June of 1970. Jobs were scarce because of the deep recession in progress at that time and I was fortunate to find one that would allow me to pay for my college education beginning in the Fall of that year. Little did I know that this rather menial job would define my interests for the next fifty years.
Safeco was in the process of designing a new twenty-two story tower to be used as its headquarters and the CEO of the company and I had forged a relationship when I found him in the elevators while delivering mail. He knew that the mailroom of the new building would be below grade and would lack daylight. He asked me to work with the architects (NBBJ) to make the space bright and pleasant. I took my assignment seriously and found it fascinating. I used the public library (long before personal computers or the internet) for material that might help me but found very little. I worked with NBBJ by conducting space/time observations (another early study) and listened to topics discussed like sound attenuation, color optimization and workflow. I was hooked. I eventually found EDRA in 1972 and joined the organization in 1975.
I could not enter college immediately due to lack of funds, but approached the University of Washington in 1972, asking the Dean of the School of Architecture if he would allow me to
devise a Special Projects, interdisciplinary major combining architecture and psychology. The Dean told me this was a nascent field and he would not support such a major. I was heartbroken but then approached Seattle University (a private Jesuit school) with the same question. Although Seattle University (SU) did not have a school of architecture, it did have robust schools of engineering and psychology. SU thought it was a great idea and approved my major. It was during my Senior year at SU that I performed the subject study. I had gained a great deal of knowledge through EDRA in the interceding years and knew exactly the type of study I wanted to conduct.
Hypothesis
The subject study was to take a large, homogenous population residing in three or more office configurations to determine if twenty-three self-reported factors (delivered by means of a questionnaire) changed as office configuration changed [Hypothesis: Self-reported perceptions of factors will change depending upon the influence of the office design surroundings]. If the self-reported factors did in fact change at a statistically significant level (0.05) then we could infer that certain office designs were possibly instrumental in changing perceptions. This would support the Determinist school of thought that held architectural or office designs could mold perceptions and perhaps behavior.
Methodology
In an effort to preserve the historical value of this study, the decision has been made to not recreate the study in a new written format. Consequently, attached are the original, typewritten pages of the study and correspondence from Robert Sommer and to the management of the company where the study was performed.
Footnotes
1 University of Sydney, 2013
https://www.sciencedirect.com/science/article/abs/pii/S0272494413000340
THE 7JNIGARD BUILDING PERFOR:4ANCE QUESTIONNAIRE
Unigard's physical layout provided the researcher with
a natural setup for a comparison of factors among different
working conditions (primarily furniture arrangements).
These conditions may be titled:
1) Burolandschaft (or office landscaping)
2) File partition
3) Bull Pen arrangement
4) Service arrangement (mail room, press room:
Each of these furniture and office arrangements offered distinct attributes, all of which might possibly have had an effect on the employees working in them. The purpose of the research then, was to administer the same questionnaire to the employees in each working arrangement, and determine if perceptions of different physical conditions changed as office arrangement changed.
METHODOLOGY
The Unigard questionnaire consisted of 23 questions of which 17 tested physical factors, and 5 elicited demographic information. Because of limited time for analysis, only factor questions 9 thru 17 were analyzed. These factors where as follows:
Sound transmission
To keep out elements
Orientation
2
Ventilation
Liking the building
5) Equipment
Maintenance
Earthquakes
Fire protection
Two hundred and fifty questionnaires were printed and distributed to managers for all departments, who in turn distributed them to the various supervisors under them, who in turn dispersed them to their employees who were asked to respond on a voluntary basis. It should also be noted that managers and supervisors also were included among the respondents.
Upon return of the questionnaires, each respondent's questionnaire was sorted into one of the four previously mentioned furniture categories on the basis of department information obtained from the questionnaire.
A Kolmogorov-Smirnov one-sample test was performed, on each of the four groups. Although this test provided information about the likes and dislikes of those individuals occupying like surroundings in terms of furniture style, it did not provide quantitative information regarding the relationships between groups. For this information an Analysis of Variance (AOV) was performed. Traditionally, a 't' ratio for each pair of means would have been performed, however the AOV is more efficient on several counts. First, the 'F' ratio tells whether any mean differs signi2icantly from any other. If 'F' is significant then a 't' test may be performed
between any two means to locate the point of significant difference. Secondly, the within groups variance based on (in this case) 141df, gives a better measure of experimental error than could be obtained from the SEd of any two means based on a lower df if only two groups were involved.
DATA
The total number of respondents for the research (N), was 145. The group breakdown was as follows:
Burolandschaft 25
File Partition 45
Bull pen 45
Services 30
total 145
The total number of employees at the Unigard facility was approximately 200, which gives a total return percentage of 72.5%. The distribution of the questionnaire took place over a three-day period so as to be as little disruptive as possible given the time restriction of one quarter.
FINDINGS
To provide the managers of Unigard with the relevant data derived from the research, the K-S test results will be shown here only where the null hypothesis has been rejected, indicating a significant departure from the equal distribution among scores. The direction or 'average feeling' will be indicated by a mean score for each furniture group type. Those factors where the mean score was low will be listed
first, with the rest of the factors listed in an increasing
hierarchy of preference. For the mean evaluation, 1 would
be the lowest, 3 average, and 5 best. (results on following page)
ANALYSIS OF K-S FINDINGS
It appears from the mean data that excessive noise may be a problem for those in office situations. However, in that those in the service areas (mailing & printing) did not find the noise in work areas to be excessive (where one would intuitively suspect a noise problem due to machinery etc. might lead to the theory that expectations of what a work area 'should' be like may play a role in the perception of excessive or less than excessive noise levels. Though this may be true, the individuals working in these areas may have a legitimate complaint, and with the significant number of people depicting this situation in their questionnaires, it is obviously a disturbing factor in their working lives.
Other factors were not as predominant as Sound Transmission. However, 'Liking the building' seemed to score moderately well, again with services being atypical; this time however, with a lower rating than the other situations. It might also be noted that a large number of people seemed to think there was far too little filing space (that some of the files were in the restrooms). There were also many comments concerning what one woman described as 'the unpredictable heating and cooling system'.
ANALYSIS OF VARIANCE TEST RESULTS
An Analysis of Variance between groups showed only one statistically significant difference between factors tested.
The two groups involved were the service group and the Bull pen group.
The results of that test are shown below:
After the AOV test was completed a 't' test was conducted to determine what groups in this factor contributed to the rejection of Ho.
On the factor VENTILATION, the two groups mentioned above differed significantly at the 0.05 level of confidence. Consequently, the services departments employees find they have better ventilation than any other group tested. On the other hand, the employees in the Bull pen arrangement found the ventilation in their area to be the worst of any group tested.
This difference must in part be attributed to the relative ceiling height between the two areas. The Bull pen areas have very low ceilings compared to rather high ceilings in the service areas. There were many comments concerning a claustrophobic feeling in the questionnaires from the Bull pen areas, indicating that perhaps the desks are too crowded and the ceilings too low.