Control/Display Relationship of The
Total Page:16
File Type:pdf, Size:1020Kb
Human Factors: The Journal of the Human Factors and Ergonomics Society http://hfs.sagepub.com/ Control/Display Relationship of the Four-Burner Stove: A Reexamination Sheng-Hsiung Hsu and Yu Peng Human Factors: The Journal of the Human Factors and Ergonomics Society 1993 35: 745 DOI: 10.1177/001872089303500413 The online version of this article can be found at: http://hfs.sagepub.com/content/35/4/745 Published by: http://www.sagepublications.com On behalf of: Human Factors and Ergonomics Society Additional services and information for Human Factors: The Journal of the Human Factors and Ergonomics Society can be found at: Email Alerts: http://hfs.sagepub.com/cgi/alerts Subscriptions: http://hfs.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://hfs.sagepub.com/content/35/4/745.refs.html >> Version of Record - Dec 1, 1993 What is This? Downloaded from hfs.sagepub.com at HFES-HUMAN FACTORS on October 24, 2011 HUMAN FACTORS, 1993,35(4),745-749 SHORT NOTE Control/Display Relationship of the Four-Burner Stove: A Reexamination SHENG-HSIUNG HSU! and YU PENG, National Taiwan Institute of Technology, Taipei, Taiwan Several previous studies concerning the arrangement of the controllburner rela- tionship of a four-burner stove showed discrepancy in their results. For this reason, a further analysis of this arrangement is necessary and worthwhile. Two research methods were adopted to duplicate earlier studies. One was the paper-pencil test in which subjects took three different questionnaire forms that used alphabetical, sign, and numerical code systems. The other method was computer simulation in which subjects took part in a performance test of four arrangements of con troll burner designs, and reaction time and error rate were measured. The results in- dicated the existence of a suggestive effect, which is a tendency by some stimuli- cues to induce a specific response unawares in subjects. This confounding variable must be controlled in the test tool design. Results also revealed a population stereotype in the controllburner linkage relationship that was different for Chinese subjects and for American subjects. The equivalence of research methods between the paper-pencil test and computer simulation was not completely assured. INTRODUCTION sponse time, fewer errors, greater safety, and reduced mental workload. Compatibility refers to the agreement be- After reviewing the literature of four- tween the stimuli/response relationship and burner stove studies, we found that the clas- human expectations and as such is a central sic example of control/response compatibility human factors issue. In this regard, a system was worth reexamining. Chapanis and Lin- that is ergonomically designed should be denbaum (1959) and Ray and Ray (1979) pre- compatible with population stereotypes. We sented several arrangements of controls and postulate that a greater degree of compatibil- burners (as shown in Figure 1)to subjects and ity will result in faster learning, shorter re- asked them to turn off or on specific burners. The stove model was simulated using colored perspex disks to represent the burners, each 1 Requests for reprints should be sen! to Sheng-Hsiung of which could be illuminated by bulbs. The Hsu, Department ofIndustrial Management, National Tai- wan Institute of Technology, 43. Section 4, Keelung Road, results from both studies showed that Type II Taipei 10772, Taiwan. appeared to be superior. @ 1993, The Human Factors and Ergonomics Society, Inc. All rights reserved. Downloaded from hfs.sagepub.com at HFES-HUMAN FACTORS on October 24, 2011 746-December 1993 HUMAN FACTORS than were Types II, III, and IV, which did not A c:=::> c:=::> C differ significantly from each other. This means that no particular stereotype emerged B D c:> c:> as predominant. After considering the cause of those incon- 0000 sistent and indefinite findings, we supposed they might have resulted from the code sys- tem used in Shinar and Acton's study (see Type Assigned Control Order Burner Figure 2). Because the alphabetic codes A, B, Sequence C,and D are essentially sequential, some sub- A jects might have been influenced unknow- II B D C U ingly by this. Consequently, they filled in the III A B C D four controls with those letters consecutively III from left to right, thus facilitating a quasi- IV B A D C predominance of the Type III choice. There- N fore, we were encouraged to conduct a study V B A C D n to test whether a suggestive effect on the subject's judgment actually existed with a Figure 1. Control·burner arrangement relationship used in previous studies. code system of a sequential nature. Sugges- tive effect is defined here as the tendency to induce unawares a specific response in a Shinar and Acton (1978) presented to sub- receptive person by some kind of cue or jects a questionnaire containing a drawing stimuli. and asked them to point out which of the un- We were also interested to know whether or marked controls they thought controlled each not a difference in the population stereotype of the burners. As in Figure 2, the burners existed between Chinese and American sub- were labeled "A," "B," "C," and "D," and the jects regarding the controllburner relation- controls were unlabeled. Shinar and Acton ship. This aspect of the controllburner rela- found that the most frequently chosen ar- tionship had not been studied before. rangement (see Figure 1) was Type III (31%). Finally, we wanted to study the equiva- Arrangement Type II, which induced fewer lence of two methods: the paper-pencil test to errors than did Type III in previous studies, study choice and computer simulation to was chosen by only 25% of the subjects. Nev- study reaction times and errors. ertheless, a chi-square analysis showed that Type V was chosen significantly less often METHODS Paper-Pencil Test ~ CC) There were 423 college students-291 males and 132 females-who volunteered to serve as subjects in this study. In order to CD CD evaluate the possible suggestive effect of se- quential codes, three different forms of ques- 0000 tionnaires were developed. The code systems Figure 2. Drawing from the questionnaire used in employed in each of these three question- Shinar and Acton (1978). naires were as follows (see Figure 3): Downloaded from hfs.sagepub.com at HFES-HUMAN FACTORS on October 24, 2011 CONTROUDISPLAY RELATIONSHIP December 1993-747 CDCA:> en@) 0000 Qa Qs Figure 3. The three forms of questionnaires used in this study. 1. Alphabetical Questionnaire (Qa): The alpha- as the ready position, on the table with the betic codes A, B, C, and D were labeled over keyboard. the four burners, as in Shinar and Acton's study (1978). When subjects were ready, the computer 2. Sign Questionnaire (Qs): A set of sign codes, displayed a graphic stove on the screen; then 1:r, #, 6, and D, devoid of sequentiality, were a light spot would appear on one of the assigned arbitrarily to each burner. 3. Numerical Questionnaire (Qn): A string of stove's four burners simultaneously with an numerical codes, 1, 2, 3, and 4, were allo- acoustic signal. The subjects' task was to turn cated not to each burner but to each control off the light spot by pressing the "correct" from left to right. key control as quickly as possible. If subjects The subject's task was to fill in each control made an error, they had to immediately try blank with its corresponding burner's code another key until the right one was found. (for Oa and Os), according to his or her ste- Before the test began, subjects were allowed reotyped judgment, and conversely to fill in practice to make sure that they were familiar each burner blank with its corresponding with the experimental task and control- control's code (for On). burner configuration. The three types of questionnaires were ran- The testing order for the graphic burner domly distributed to the 423 subjects: Oa light and the linkage type was randomized. went to 153 subjects, Os to 143,and On to 127. Because linkage type was varied within sub- jects, a two-day break was instituted between Computer Simulation testing sessions on the different types. Each light burner was turned on and off 20 times, A computer simulation was developed to so each type of linkage was tested a total of 80 model the four aforementioned types (II, III, times. The intertrial interval between light IV, and V) of control/burner linkage. The ex- off and on was fixed at about 1 s. perimental procedure was similar to that of The performance measures collected in this Chapanis and Lindenbaum (1959). Fifteen experiment consisted of reaction times and male college students aged 18 to 20 years vol- number of errors, which were automatically unteered to participate in this simulation recorded by the computer. study. At the beginning of each test session, the RESULTS AND DISCUSSION experimenter told subjects the purpose of the Paper-Pencil Test experiment and mentioned the type of con- trol/burner linkage that would be used. The Table 1 summarizes the numbers of intu- four keys (E, T, U, and 0) on the keyboard itive choices (and the percentage these repre- were set up to act as stove controls. Subjects sent) of each possible controlfburner linkage were asked to place their index fingers on (or arrangement made by subjects in the three return to) a predetermined point, designated forms of questionnaires. Downloaded from hfs.sagepub.com at HFES-HUMAN FACTORS on October 24, 2011 748-December 1993 HUMAN FACTORS