Journal of Physical Activity and Health, 2009, 6. 617-624©2009 Human Kinetics, Inc.

Effect of Using a Treadmill Workstation on Performanceof Simulated Office Work Tasks

Dinesh John, David Bassett, Dixie Thompson, Jeffrey Fairbrother, and Debora Baldwin

Although using a treadmill workstation maychange the sedentary nature of desk jobs, it isunknown if walking while woridng affects per-formance on office-work related tasks. .Piir-pose: To assess differences between seated andwalkitig conditions on motor skills and cogni-tive function tests. Methods: Eleven males(24.6 ± 3.5 y) and 9 females (27.0 ± 3.9 y)completed a test battery to assess selectiveattention and processing speed, typing speed,mouse clicking/drag-and-drop speed, and GREmath and reading comprehension. Testing wasperformed under seated and walking condi-tions on 2 separate days using a counterbal-anced, within subjects design. Participants didnot have an acclimation period before thewalking condition. Results: Paired t tests (P <.05) revealed that in the seated condition, com-pletion times were shorter for mouse clicking(26.6 ± 3.0 vs. 28.2 ± 2.5s) and drag-and-drop(40.3 ± 4.2 vs. 43.9 ± 2.5s) tests, typing speedwas greater (40.2 ± 9.1 vs. 36.9 ± 10.2 adjustedwords • min~'). and math scores were better(71.4± 15.2 vs. 64.3 ± 13.4%). There were nosignificant differences between conditions inselective attention and processing speed or inreading comprehension. Conclusion: Com-pared with the seated condition, treadmillwalking caused a 6% to 11% decrease in mea-sures of fine motor skills and math problemsolving, but did not affect selective attentionand processing speed or reading comprehen-sion.

Keywords: light-intensity activity, walk-work,NEAT

The authors are with the Dept of Exercise, Sport, and LeisureStudies, University of Tennessee. Knoxille.

Currently, two-thirds of American adults are eitheroverweight or obese.' Obesity rates are rapidly increas-ing in most modern and industrialized nations.'-^ Tbeobesity epidemic has been attributed to the increasedavaiiability of inexpensive energy-dense foods in con-junction with a modem environment that has reducedopportunities to be physically active.- To combat thisepidemic, tbere has been an empbasis on increasingphysical activity through structured exercise regimens.More recently, it has been suggested that an increase innonexercise activity thermogenesis (NEAT) may helpcontrol weight.^ In 2003. the U.S. Census bureau esti-mated that approximately 52.4% of employed Ameri-cans use a computer at work.^ Because many Americansspend time at work sitting in front of a computer,researchers have identified the work environment as aplace to increase NEAT.' -''

The treadmill workstation was first proposed andbuilt twenty years ago by Edelson to reduce inactivity inoffice workers.* ** It was touted to solve the hazards ofcontinuous sitting (postural fixity). These hazardsincluded aches and pains, stress, and other illnesses.^^Treadmill workstations allow users to alternate betweensitting and walking while working during a regularworkday. Treadmill workstations consist of a conven-tional motorized treadmill that slides under a heightadjustable sit-to-stand table. A regular office cbairallows tbe user to sit, if they choose.

After briefiy disappearing, the treadmill worksta-tion concept was reintroduced as a potential weight lossstrategy by Levine and coworkers.-'' They suggested thatif an obese individual walked 2 to 4 hrs/workday at atreadmill workstation, he/she would increase dailyenergy expenditure by about 500 kcal day^'.^ Tbey esti-mated that this could translate to a yearly weight loss intbe range of 20 to 30 kg.

Although the treadmill workstation seems to havetbe potential for weight loss, the effects of this approachon work performance still need to be assessed. An indi-vidual using a treadmill workstation may need to divideattentional resources between treadmill walking andoffice work, which might compromise work perfor-mance. It remains to be seen if slow walking at a tread-

617

618 John et al

mill workstation affects performance in work relatedvariables. The results of such a study would establishthe feasibility of the treadmill workstation as an effec-tive strategy for weight control. Tiius. the purpose ofthis study was to determine if walking while working ata treadmill workstation affects selective attention andmental processing speed and the performance of simu-lated office work tasks involving 1) fine motor move-ments (typing and mouse movements) and 2) mathe-matical and verbal reasoning.

Methods

Workstation

The treadmill workstation consisted of a sjt-stand table(501-11; Conset A/S. Denmark), height adjustableoffice chair, treadmill (TI450: Vision Fitness. LakeMills. Wl), and a computer. These components werepurchased separately and assembled in the AppliedPhysiology Laboratory {Dept of Exercise, Sport, andLeisure Studies) at the University of Tennessee. Thetable used in the current study could be lowered to aminimum height of 76 cm for seated work and raised toa maximum height of 113 cm above the treadmill deckduring walking. The recommended heights while per-forming normal sitting and standing work are 70 to 78cm and 95 to 115 cm, respectively.'" The treadmillworkstation was constructed in a way that the usercould alternate between treadmill walking and sittingby simply stepping off the treadmill, sitting on thechair, and adjusting the table to a desired height bymeans of an electronic switch. The computer screenwas placed on the table surface and could be moved tosuit treadmill walking or sitting.

Participants

Eleven male (24.6 + 3.5 y) and 9 female (27.0 ± 3.9 y)graduate students from the University of Tennessee vol-unteered to participate in this study. Ability to participatein the study was assessed using a medicai history ques-tionnaire. Ail participants provided written, informedconsent. Although all participants were famihar withtreadmili waiking, none were used to continuous, slowtreadmiii waiking at imph or had previously used atreadmiii workstation. The experimental protocol was

Table 1 Physical Characteristics ofParticipants (N = 20)"

CharacteristicAge (y)Height (m)Weight (kg)BMI (kg/m^)

26.4 (4.04)1.69(0.10)67.05(15.76)22.98 (3.44)

' Data expressed as mean (SD).

approved by the University of Tennessee InstitutionalReview Board. As a safety precaution, a I-meter lanyardwas attached to the participant's clothing that wouldautomatically turn off the treadmill if they moved morethan 1 meter away from the treadmill control panelduring walking, Height and weight of each participantwas measured using a stadiometer and physician's scale,respectively. Participant characteristics are shown inTable 1.

Test Battery

Participants had to complete a test battery comprising of5 tests to assess selective attention and processing speed,cognitive function and fine motor movement. The paper-and-pencil form of the Stroop Color and Word Test(Stoelting Co., Wood Dale, IL.) was used to measureselective attention and processing speed." The Strooptest activates an automatic verbal interference thatimpedes the task ofcolor naming.'-This test is valid andreliable in identifying individual differences in the allo-cation of attentionai resources caused by interference.'^The test had 3 sections of 100 items each and partici-pants had 45 seconds per section to complete as manyitems as possible. During testing, participants had toverbally identify each item so that the test administratorcould verify if items were being correctly identified.The first section required participants to read the namesof colors printed in black ink. The second section haditems represented by 4 consecutive X symbols printedin red, blue, or green and participants had to identify thecolor of the print. Items in the third section were namesof colors (Red, Blue, and Green) printed in a color notrepresented by the word. Participants had to identify thecolor of the printed word rather than simply read theword. For instance, if an item on the test were the wordRed printed in green ink, the correct answer to this itemwould be green. The number of correct items for eachsection were recorded and used to determine /-scores.The method of obtaining i-scores is described in theproduct manual that accompanies the Stroop Color andWord Test."

Fine motor movement performance was assessedwith a typing and 2 computer mouse proficiency tests.These tests were conducted on a Dell OptiPiex GX260desk top computer with an Intei Pentium 4 processor(2.40 GHz) and 256MB RAM. The Mavis BeaconTeaches Typing 17 (Riverdeep Inc., San Francisco, CA.)computer program was the standard typing exercise forall participants. In this test, participants had to replicatesections of text displayed on the screen. On completionofthe assigned paragraphs, typing speed, excluding anyerrors made, was recorded in adjusted words per minute(AWPM). This software has previously been used inresearch studies to assess typing speed.'"*'* Computermouse proficiency was assessed using a visual basicprogram. In this test, participants were instructed to per-form a mouse clicking and a drag-and-drop task."' In themouse clicking task, the participant had to click on 1 of

Treadmill Workstation Use and Work Performance 619

25 icons that randomly turned red in color until all iconswere clicked. The drag-and-drop task involved draggingand dropping 25 icons 1 at a time (when red) into alarger black box. An icon stayed red until successfullydropped into the box. after which another random iconon the screen turned red. The time taken to completeeach task was recorded.

Paper-based graduate record examination (GRE)math and verbal (reading comprehension only) sectionswere used to assess cognitive function of the partici-pants. These sections were obtained from the officialsource for GRE review guide (Educational Testing Ser-vices) that contain examples of actual GRE tests. TheGRE is used to predict academic performance by mea-suring the basic developed abilities related to success ingraduate school.'' Mathematical reasoning and readingcomprehension tests from the GRE have been used inprevious research studies examining cognitivefunction.'^'^ For the verbal section, each participantwas instructed to read a long (600 words) and shortparagraph (200 words), and to answer 11 related ques-tions in 18 min. The reading comprehension part of theGRE is designed to measure the ability to analyze, eval-uate, and synthesize information. For the math sectionparticipants had to answer 30 questions in 30 min. Thissection of the GRE measures the participant's ability toreason and solve quantitative problems under a timeconstraint. Scores for both GRE tests were calculated asthe total percentage of correct responses from eachtest.

Study DesignData collection was conducted at the treadmill worksta-tion during 2 visits separated by 2 days. During theirfirst visit, the study protocol was explained to the par-ticipants and they completed the test battery for eitherthe sitting or treadmill walking (with no prior acclima-tion) condition. The test battery for the remaining con-dition was completed during the second visit. Partici-pants were allowed to adjust the table to a desired heightfor both conditions. During the treadmill walking condi-tion, participants underwent testing while walking at 1mph (27 m-min"'). Participants were allowed to warmup at this speed for a few minutes before the test began.The intensity of level walking at 1 mph is less than 2METS.-'' To avoid a testing effect, 2 different versionsof the GRE and typing test paragraphs were used. Acounterbalanced, within-subjects design was used toavoid confounding due to order effects,

Statistical ProceduresStatistical analyses were conducted using SPSS version14 for windows (SPSS, Chicago, IL). The independentvariable in this study was the treadmill workstation con-dition (seated VA. treadmill walking). The dependentvariables were scores for the Stroop task (t-scores),typing test (AWPM), mouse clicking and drag-and-drop

tests (completion time), and GRE math and readingcomprehension (percentage of correct answers) tests.Test results are presented as means ± standard devia-tions. Paired samples t tests were used to examine dif-ferences between test results from treadmill walkingand the sitting conditions. Repeated-measures ANOVAson test versions were performed to determine if GREtest versions had a confounding effect on the resultsfrom the 2 conditions. Statistical significance for allanalyses was set at /* < .05.

ResultsPaired samples f tests showed significant differencesbetween the 2 conditions on typing, mouse proficiency,and GRE math test scores. In the sitting condition, par-ticipants displayed better scores on the typing (i^ =-3.161, P = .005). recorded lower completion time onthe mouse clicking (/¡q = 2.747, P= .013) and drag-and-drop tests (r|9 = 3.839, P = .001 ). and scored higher onthe mathematical reasoning test (r|9 = -2.I69. P= .017).Eigures 1, 2, and 3 illustrate performances on the Stroopand typing test, mouse proficiency tests, and cognitivetests, respectively. Mouse clicking and drag-and-dropscores (in seconds) for the walking condition werelower hy 8% and 6%, respectively. Typing (AWPM) andmath scores for the sitting condition were 9% and 11%greater than scores for the walking condition, respec-tively. There were no significant differences (P < .05)between the walking and sitting conditions for theStroop and reading comprehension tests. However,repeated-measures ANOVA showed a significant condi-tion X version interaction for reading comprehension {P< .05). In other words, the group that took version 1while walking and version 2 while sitting scored betterin the sitting condition, and the group that took version2 while walking and version 1 while sitting scored betterin the walking condition. However, there was no overallsignificant difference in reading comprehension betweenthe sitting and walking conditions.

Discussion

Test ResultsThe results of the current study indicate that walkingwhile working decreases scores on tests of typing andmouse proficiency, and math solving ability by approxi-mately 6 to 11%. This may be because the added task ofwalking puts an increased load on both mental process-ing and motor control.-^-^ An increased load causesinterference in 1 or both tasks, thereby lowering taskperformance.^"

Significantly lower (P < .05) peribrmances on finemotor movement tests during the walking conditionmay have resulted from the increased complexity ofperforming multiple motor tasks (walking and typing/mouse tasks) that require a more complex interaction

620 John et al

70

60

50

Typing 40Speed(AWPM)

30

20

10 -

Typing Stroop

t-score

Figure 1 — Mean scores and standard deviations on typing and Stroop tests for walking and sitting conditions (N = 20). * Signifi-cantly different from each other {P < .05).

Clicking Drag-Drop

Figure 2 — Mean scores and standard deviations on mouse clicking and drag-and-drop tests for walking and sitting conditions (N= 20). * Significantly different from each other (P < .05).

with cognitive abilities, and iticreased recruitment ofattentional resources.^'•^- Similarly, performing a cogni-tive function like math problem solving and treadmillwalking simultaneously may have increased complexity

and thereby placed a higher than normal demand onattentional resources. This may have resulted in lowerscores on the mathematical reasoning tests during thewalking condition. Research by Giomo et al ** showed

Treadmill Workstation Use and Work Performance 621

Math Reading comprehensionFigure 3 — Mean scores and standard deviations on GRE math and reading comprehension tests for walking and sitting conditions(N = 20), * Significantly different from each other {P < .05).

significantly lower performances (P < .05) on a fingertapping task and a cognitive test during submaximalexercise (75% of ventiJatory threshold) as comparedwith performances in a sitting condition.-''* Although theexercise intensity in the study hy Giomo et al was muchhigher than the current study, a similar trend of decreasedperformance during dual task performance was observedin both studies. ^^

Findings on the typing tests from the current studywere different from those by Edelson and Danoff. Theyfound no significant difference in typing performancesbetween walking and sitting while working at a tread-mill workstation."^ In contrast to this finding, the currentstudy determined that typing speed decreased slightly(3.3 ± 4.7 AWPM) while walking. This discrepancymay have resulted from the fact that participants inEdelson and Danoff's study had an average seatedtyping speed of 61 AWPM in comparison with only 40AWPM in participants from the current study.** Theaverage typing speed of the current participants wassimilar to the average speed (40 AWPM) of 3475 par-ticipants from Ostrach's study.-^ Results from Ostrachwould classify participants tested by Edelson andDanoff in the 2nd decile and the current participants inthe 5th decile.'^ Therefore, the difference between find-ings from Edelson and Danoff and the current studymay have resulted from the fact that participants in thefomier study were more experienced typists than those

in the current study. In addition, Edelson and Danoffexamined only 5 participants and their study may havelacked the statistical power required to show a signifi-cant difference.

Unlike fine motor movement and GRE math tests,results from the Stroop tests were not significantly dif-ferent between walking and sitting conditions. Duringthe Stroop test, participants use their working memoryto resolve a mental conflict between word reading andcolor naming.'--^ In the walking condition, participantsalso had to simultaneously process the task of walking.Findings from the current study were similar to those byGrabiner and Troy who did not report a significant dif-ference iP = .052) in Stroop test results between a sta-tionary condition (standing) and treadmill walking.^"However, they reported a significant decrease in stepwidth variability during treadmill walking while per-forming the Stroop test.-" They concluded that volun-tary gait changes occur to compensate for the reducedvisual resources allocated to walking while performingthe Stroop test.-* The current study did not assess stepwidth variability. In general., it can be said that walkingwhile performing tasks that invoke a mental load similarto the Stroop test, does not affect an individual's perfor-mance on the mental task.

Like the Stroop test, reading comprehension scoresin the sitting condition were not significantly different(P < .05) than scores in the walking condition. These

622 John et al

results are in contrast to those of Barnard, Yi, Jacko, andSears where reading comprehension scores for sittingwere significantly higher (P < .01) than those for walk-ing. ** Unlike Barnard et al who reported a 10% differ-ence between scores, reading comprehension scoreswhile sitting were only 4% bigher than treadmill walk-ing in tbe current study. This difference may be attribut-able to the fact that participants from Barnard et alwalked on a narrow course that meandered around aroom as compared with the simpler task of treadmillwalking in the current study. In addition, the device usedby Barnard et al was a personal digital assistant (PDA)and had a much smaller screen as compared with thecomputer monitor in the current study. ** In summary,reading comprehension may not place as high a demandon attentional resources as other tasks, and thus is notaffected by treadmill walking.

Practical implications of LoweredPerformance

Although performance during the walking conditionwas statistically lower for the typing, mouse proficiency,and math tests, this section of the paper attempts toexamine how these results could potentially impact real-life work productivity. In this study, average typingspeed decreased from 40.2 while sitting to 36.9 AWPMwhile walking. According to Ostrach. the average typingspeed lies between 38 and 43 AWPM.-* It can be specu-lated that, because the reduced typing speed during tbewalking condition falls out of the average typing speedby just 1 AWPM. tbere may be a marginal impact ontyping while walking and working in an office setting.To substantiate the previous argument, we could con-sider tbe example of emailing, which is a popular formof electronic communication.^^ A study that examinede-mail communication between physicians and theirpatients showed that patient emails and physician repliesaveraged 139 and 39 words, respectively.-^ If we com-pute the time taken to compose an e-mail by these indi-viduals based on average typing speeds from Ostrach'sstudy (sitting: 40 AWPM) and tbe current study (walk-ing: 36.9 AWPM), patients and physicians using a tread-mill workstation would take only 17.5 and 4.9 s longerto compose an e-mail, respectively. In addition, accord-ing to the Intemational Organization of Standardization(ISO) ergonomie standard for computer visual displayterminals, if a standard keyboard is replaced with a dif-ferent kind that may cause a decrease in typing speed,the speed obtained using the new one must lie within0.75 standard deviations of the speed for the standardkeyboard to be acceptable.^" Although we did not com-pare different keyboards in our study, we can apply ISOstandards to our results because we induced an experi-mental condition that decreased typing speed. Typingspeed during tbe experimental walking condition (36.9± 10.2AWPM) was within the acceptable range (33.4 to47 AWPM) as per ISO specifications.

The use of a computer mouse depends mainly ontbe type of job being performed. In this study we exam-ined the effect of walking while working on clickingand drag-and- drop tasks only. A previous study exam-ining computer mouse use in office workers showed thatan employee moves and clicks tbe mouse approximately78 limes per hour.'' In the current study, time taken tomove the mouse and click an item took 1.06 and 1.12swhile sitting and walking, respectively. Therefore,moving the mouse and clicking 78 items in an hourwhile walking and working would require only 4.68 smore than sitting, which may not have a substantialeffect on overall work productivity. In general, the taskof dragging-and-dropping items using a mouse at workis less frequent than clicking.

The third test tbat showed significantly differentresults between sitting and walking condifions was theGRE math test. To accommodate errors in the measure-ment process. Educational Testing Services defines a"meaningful difference between scores" as a valuegreater than 2 times the standard error of measurement(SEM) of score differences. In the current study, the dif-ference between sitting and standing conditions for themath test wa.s 2.48 times tbe calculated SEM of scoredifferences and thus meaningful.^^ However, tbe margintbat renders tbe walking GRE math score to be mean-ingfully different from that obtained while sitting is verysmall. In otber words, the score obtained while walking(64.3%) would not have been meaningfully differentfrom the score obtained while sitting (71.4%) if it wasbetween 65.6 and 77%.

The differences between scores for the sitting andwalking conditions on tbe typing, mouse proficiency,and the GRE math tests could be reduced tbroughacclimation. -^ •''' In light of the potential benefits of usinga treadmill workstation, the benefits may outweigh tbedifferences observed.

Restoring Energy Balance

Current data indicate tbat in the past forty years, therehas been an increase in the average American's caloricconsumption and a concurrent decrease in energyexpenditure.'^-'^ As a result, tbe average American adultgains 1.8 to 2 pounds/year." Levine and Miller sug-gested that walking while working at a treadmill work-station increases energy expenditure by approximately165% in obese individuals as compared with sitting at adesk.'' They speculate that the treadmill workstationmay be effective in preventing weight gain or loweringobesity rates in office workers.

Hazards of SittingOccupations involving continuous seated work result inincreased musculoskeletal discomfort.'^ Individualswho sit for more tban 95% of time at work experiencedgreater musculoskeletal discomfort than those who wereable to vary tbeir posture.'^ In response to a subjective

Treadmill Workstation Use and Work Performance 623

assessment of musculoskeletal discomfort, users of thetreadmill workstation report that walking while workinghelped reduce back problems associated with continu-ous sitting.''^ In addition to musculoskeletal discomfort,sedentary behaviors like continuous sitting also affectthe body at a molecular level. Hamilton. Hamilton, andZderic proposed that prolonged hours of sitting mayupregulate specific molecular, physiologic, and bio-chemical responses (also known as inactivity physiol-ogy) that could lead to metabolic disorders.'" A specificexample of molecular adaptations to reduced low-inten-sity activity is the transcription of an inhibitory genethat induces LPL suppression through a posttransla-tional mechanism.'" Hamilton et al state that inactivity(sitting) and low nonexercise activity may produce seri-ous health problems that cannot be explained by exer-cise deficiency alone."' In other words, maintaininghigher levels of low-intensity activity independent ofthe recommended moderate-vigorous physical activitycan lower several metabolic risk factors.'^'

Strengths and LimitationsA limitation of this study is that it was not conducted ina proper office setting. In addition, the participants inthis study were graduate students and not actualemployed office workers. The results of this study maynot be comparabie to resuits from an office settingbecause the duration of a singie testing session (60 min)was less than the recommended duration of 2 to 4 hrs/day. More importantiy. the participants may not have hadsufficient time to acciimatize to waiking and working.Acciimation may reduce or eiiminate the marginal iow-ering of work performance observed initiaiiy. Researchexamining interierence due to duai task performancesuggests that acclimating to performing duai tasks canreduce or even eiiminate interierence, thereby resultingin improved task performance.^^-^ In addition, other fac-tors that may potentiaüy affect work performance are notdiscussed in this study. One such factor could bie vari-abiiity in distance from the area of work due to seiectedconditions (sitting and walking). Assessing the effect ofali potentiai factors that may affect work performancewas beyond the scope of this study.

On the other hand, the current study is the oniy ofits icind to have examined the impact of using a tread-miii workstation on the performance of tasks that simu-iate office work. The selected tasks involved cognitiveand motor skiiis that are a requirement in today's workenvironment. The current study also estabiished thatwalking while working did not greatly affect workperformance.

ConciusionThe treadmiii workstation aims to replace 2 to 4 hoursof sitting at work with low-intensity physical activitythat may help obese office workers achieve a negative

energy balance and reduce obesity relatedUsing a treadmill workstation could also attenuate mus-culoskeletal discomfort, reduce metabolic risk factorsassociated with continuous sitting, and lower stresslevels.^-"' The current study compared the effects ofusing a treadmill workstation on simulated office worktasks. Walking while working was associated with aminor 6 to 11% decrease in math problem solving,mousing, and typing performance. It is possible that thisdecrea.se in performance could he eliminated throughacclimation to walking while working. It is imperativeto investigate work performance in a proper office set-ting after participants are used to walking and workingfor at least 2 to 3 hrs/day. Future studies should examineif using the treadmill workstation helps lower bodyweight, reduce fat mass, and lower employee health carecosts. If these benefits can be shown, it may be possibleto convince employers that the benefits of treadmillworkstations justily the costs.

References1. Ogden CL, Carrüll MD, Curtin LR, McDowell MA, Tabak

CJ, Flegal KM. Prevaiance of overweight and obesity inthe United States. 1999-2004. JAMA. 2006:295:1549-1555.

2. intemational Obesity Task Force and European Asso-ciation for the Study of Obesity. Obesity in Europe.The case for action. London: International Obesity TaskForce and European Association for the Study of Obesity.2002. Available at http://www.iotf.org/media/euubesity.pdf. Accessed January 29, 2008.

3. Levine JA. Non-exercise activity thermogenesis (NEAT).Best Prac! Res Clin Endocrinol Metab. 2OO4;16:679-702.

4. U.S. Census Bureau. Computer Use and Ownership.Available at: http://www.census.gov/prod/2005pubs/p23-208.pdf. Accessed December 13. 2007.

5. Levine JA. Miller J. The energy expenditure of using a'walk-and-work' desk for office-workers with obesity. BrJ Sports Med. 2007;41:558-56L

6. Markoff A. BUSINESS TECHNOLOGY; Feel likeyou're on an office treadmill? Then get on an office tread-mill. New York Times. June 22, 1988.

7. Edelson N. Adjustable portable exercise desk. U.S. Patent5, 257,701. November 2 1993.

8. Edelson N, Danoff J. Walking on an electric treadmillwhile performing VDT office work. SIGCHI Bulletin.1989;21:72-77.

9. Edelson N. Hazards of sitting down on the job. WallStreet Journal. August 31. 1987.

10. Clark DR. Workstation evaluation and design: Bhat-tacharya A, McGlothlin JD, eds. Occupational Ergonom-ics Theory and Applications. New York: Marcel Dekker,1996:279-301.

11. Golden CJ, Freshwater SM. Stwop Color and Ward Test.A Manual for Clinical and Experimental Uses. WoodDale. IL: Stoelting Co.; 2002.

624 John et al

12. Wheeler DD. Locus of interference on the Stroop test.Percept Mot Skills. 1977 ;45:263-266.

13. MacLeod CM. Half a century of research on the Stroopeffect: An integrative review. PsyeholBull. 199 ! ; 109:163-203.

14. Francis M, Good G, Johnson J, et al. The effects of atyping task on the hand volume of the older worker.IVi ni;. 2004;22:liI-115.

15. Ting A, Hedge A. An ergonomie evaluation of a hybridkeyboard and game controller. Proceedings of the humanfactors and ergonomics society 45"' annual meeting.2001:1:677-681

16. Morgan T, Shaffer S. Mouse pad dimensions and theireffects on user performance. Available at: http://www.otal.umd.edu/SHORE98/bs03/index.html. Accessed June5 2007.

17. Kuncel NR, Hezlett SA. Ones DS. A comprehensivemeta-analysis of the predictive validity of the GraduateRecord Examinations: Implications for graduate studentselection and performance. Psyehol Bull. 2001 ;127:162-181.

18. Spencer SJ, Steel CM. Quinn DM. Stereotype threatand women's math performance. J Exp Soe Psyehol.1999;35:4-28.

19. Baumeister RF, Twenge JM, Nuss CK. Effects of socialexclusion on cognitive processes: Anticipated alone-ness reduces intelligent thought. J Pers Sac Psychol.2002:83:817-827.

20. Grabiner MD, Troy KL. Attention demanding tasksduring treadmill walking reduce step width variability inyoung adults. Journal of NeuroEngineering and Reha-bilitation. 2005: 2: 25. Published online 2005. doi: 10.1186/1743-000302025.

21. Regnaux JP, Robertson J, Smail DB. Daniel O. Bussel B.Human treadmill walking needs attention. J Neuroengi-neering Rehabit. 2006:3:19 Published online 2006.

22. Ebersbach G, Dimitrijevic MR, Poewe W. Influence ofconcurrent tasks on gait: A dual-task approach. PerceptMotSkills. 1995:81:107-113.

23. Pashler H. Dual-task interference on simple tasks: Dataand theory. Psyehol Bull. 1994:116:220-244.

24. Giomo JD, Leary KO, Hall E. Bixby WR. Miller. Influ-ence of exercise intensity on cognitive function duringand following submaximal exercise. Paper presentedat: The 36"' Annual Meeting of the Southeast RegionalChapter of the ACSM: February 14-16,2008. Binning-ham. Al.

25. Ainsworth BE, Haskell WL, Whitt MC. et al. Compen-dium of physical activities: an update of activity codes andMET intensities. Med Sei Sports Exerc. 2000:32:S498-S504.

26. Ostrach R. Typing speed: How fast is average. Availableat: http://www.readi.info/TypingSpeed.pdf. AccessedDecember 18,2007

27. Kane MJ. Engle RW. Working -memory capacity and thecontrol of attention: The contributions of goal neglect.

response competition, and task set to Stroop interference.J Exp Psyehol Gen. 2003:132:47-70.

28. Barnard L, Yi JS, Jacko JA. Sears A. Capturing theeffects of context on human performance in mobile com-puting systems. Personal and Ubiquitous Computing.2007:11:81-96.

29. Sittig DF. Results of a content analysis of electronic mes-sages (email) sent between patients and their physicians.BMC Med Inform Decis Mak. 2003:3:11 Publishedonline 2003.

30. Smith WJ, Cronin DT. Ergonomie test of the kinesis key-board. Proeeedings of the Human Factors Society. 1993:318-322.

31. Johnson PW, Hagberg M, Wigaeus Hjelm E, Rempel D.Measuring and characterizing force exposures duringcomputer mouse use. Scand J Work Environ Health.2000:26(5): 398-405.

32. Educational Testing Services. Graduate Record Exami-nations guide to the use of scores. Available at: http://www.ets.org/Media/Tests/GRE/pdf/994994.pdf

33. Van Seist M. Ruthruff E. Johnston JC. Can practice elim-inate the psychological refractory period effect? J ExpPsychol Hum Percept Perform. 1999:25:1268-1283.

34. Ruthruff E, Johnston JC. VanSelst M. Why practicereduces dual-task performance. J Exp Psyehol Hum Per-eept Perform. 2001:27:3-21.

35. National Center for Health Statistics. CDC. Availableat: http://www.cdc.gov/mmwr/PDF/wk/mm53O4.pdf.Accessed January 16, 2007.

36. Brownson RC. Boehmer TK, Luke DA. Declining ratesof physicai activity in the United States: What are thecontributors? Annu Rev Public Health. 2005:26:421-443.

37. Hill JO. Wyatt HR. Reed GW, Peters JC. Obesity andthe environment: Where do we go from here? Science.2003:299:853-855.

38. Ratzon NZ. Yaros T. Mizlik A. Kanner T Musculoskele-tal symptoms among dentists in relation to work posture.WorÄ. 2OOO;15:153-158.

39. Ariens GAM. Bongers PM, Douwes M, et al. Are neckflexion, neck rotation, and sitting at work risk factors forneck pain? Results of a prospective cohort study. OccupEnviron Med. 2001:58:2OO-2O7.

40. Thompson WG, Foster R. Eide D, Levine JA. Feasibilityof a walking workstation to increase daily walking. Br JSports Med. 2008:42:225-228. Published online 2007.

41. Hamilton MY, Hamilton DG. ZdeHc TW. Role of lowenergy expenditure and sitting in obesity, metabolic syn-drome, type 2 diabetes, and cardiovascular disease. Dia-betes. 2007:56:2655-2677.

42. Colditz GA. Economic costs of obesity and inactivity.Med Sei Sports Exerc. 1999;31 :S663-S667.