60 L. LU, H-J. TSENG AND C. L. COOPERFig. 1. Ð Relationships among work <strong>stress</strong>ors, moderators <strong>and</strong> <strong>job</strong> stra<strong>in</strong>Table 4 Ð Predict<strong>in</strong>g <strong>job</strong> stra<strong>in</strong>Variables b R 2 R 2 D F ConstantJob <strong>satisfaction</strong>‡ Marital status 0.14* 0.03 0.03**‡ Rank 0.04 0.05 0.02*‡ Stress 0.16** 0.09 0.04***‡ WLCS 0.20*** 0.19 0.10***OSI Ð Control 0.22***Total 0.19 12.23*** 23.46Mental <strong>health</strong>‡ Age 0.21*** 0.05 0.05***‡ Stress 0.13*** 0.16 0.11***‡ WLCS 0.33*** 0.20 0.04***‡ Stress WLCS 1.14*** 0.23 0.03**Total 13.65*** 35.88Physical <strong>health</strong>‡ Age 0.19** 0.07 0.07***Marital status 0.12*‡ Smok<strong>in</strong>g 0.17* 0.09 0.03**‡ Type A Ð impatience 0.22*** 0.14 0.05***Total 12.50*** 19.28*p 4 0.05; **p 4 0.01; ***p 4 0.001.smok<strong>in</strong>g <strong>and</strong> the `impatient' characteristic of theType A behaviour pattern were negatively relatedto it. The <strong>in</strong>teraction between impatience <strong>and</strong> work<strong>stress</strong> did not reach statistical signi®cance, <strong>in</strong>dicat<strong>in</strong>gno evidence of Type A as a moderator for the<strong>stress</strong>±stra<strong>in</strong> relationship.Copyright # 1999 John Wiley & Sons, Ltd. Stress med. 15, 53±64 (1999)
MANAGERIAL STRESS IN TAIWAN 61DISCUSSIONThe impact of managerial <strong>stress</strong>Be<strong>in</strong>g one of the very few concerted e€orts<strong>in</strong>vestigat<strong>in</strong>g managerial <strong>stress</strong> across the barriersof a s<strong>in</strong>gle organization, a s<strong>in</strong>gle occupation or asection of the national economy, <strong>and</strong> <strong>in</strong> a nonwesterncountry, this study generated valuableevidence of the impact of managerial <strong>stress</strong> <strong>in</strong> arapidly <strong>in</strong>dustrializ<strong>in</strong>g society. Among the topthree work <strong>stress</strong>ors reported by the <strong>Taiwan</strong>esemanagers, `personal responsibility' <strong>and</strong> `relationships'may be particularly <strong>in</strong>terest<strong>in</strong>g <strong>and</strong> relevant<strong>in</strong> a Ch<strong>in</strong>ese culture. In Hofstede's 14 sem<strong>in</strong>alwork on the cross-cultural comparison of workrelatedvalues, <strong>Taiwan</strong> was identi®ed as moderate<strong>in</strong> power distance (ranked 29/30 among 53countries), low <strong>in</strong> <strong>in</strong>dividualism (ranked 44among 53 countries), <strong>and</strong> moderate <strong>in</strong> uncerta<strong>in</strong>tyavoidance (ranked 26 among 53 countries).Corroborat<strong>in</strong>g this general description of the<strong>Taiwan</strong>ese culture, managers would be di<strong>stress</strong>ed<strong>in</strong> tak<strong>in</strong>g personal responsibilities, which often<strong>in</strong>volve exercis<strong>in</strong>g power <strong>and</strong> tolerat<strong>in</strong>g uncerta<strong>in</strong>ties;<strong>Taiwan</strong>ese managers would also be ratherconcerned about the <strong>in</strong>terpersonal relationships atwork. The empirical evidence seemed quite consistentwith this theoretical reason<strong>in</strong>g. As noted byother researchers, 15,16 Ch<strong>in</strong>ese workers, especiallymanagers, value guangxi, the aliative relationshipbetween superiors <strong>and</strong> coworkers, which is acharacteristic of the large power distance culture.Although empirical evidence was <strong>in</strong>conclusive 16<strong>and</strong> not particularly strong, it did po<strong>in</strong>t to thevalue of conduct<strong>in</strong>g more cross-cultural comparativestudies to map out universality as wellas speci®city <strong>in</strong> the work <strong>stress</strong> experiences <strong>in</strong>di€erent societies.On the quantitative aspects of the experience ofmanagerial <strong>stress</strong>, <strong>Taiwan</strong>ese managers perceivedmore <strong>stress</strong> than the general workforce <strong>in</strong> thecountry <strong>and</strong> were also more <strong>stress</strong>ed than managers<strong>in</strong> Hong Kong 17 (t ˆ 3.78, p 5 0.001), the UK <strong>and</strong>Germany 18 (t ˆ 13.10 <strong>and</strong> t ˆ 8.69, p 5 0.001).This level of work <strong>stress</strong> was related to <strong>job</strong><strong>satisfaction</strong>, mental <strong>health</strong>, physical <strong>health</strong> <strong>and</strong>turnover <strong>in</strong>tentions (Table 3). Furthermore, work<strong>stress</strong> was a powerful predictor of <strong>job</strong> <strong>satisfaction</strong>(account<strong>in</strong>g for 4 percent of the variance) <strong>and</strong>mental <strong>health</strong> (account<strong>in</strong>g for 11 percent of thevariance). These results are consistent withprevious ®nd<strong>in</strong>gs. 18,19 The problems of smok<strong>in</strong>g,dr<strong>in</strong>k<strong>in</strong>g, lack of exercise <strong>and</strong> absenteeism were notsevere among this sample of managers (Table 1).While the behavioural symptoms of <strong>stress</strong> were notparticularly evident, due to the Ch<strong>in</strong>ese culturalcensorship of excessive smok<strong>in</strong>g or dr<strong>in</strong>k<strong>in</strong>g, theself-reported <strong>health</strong> <strong>in</strong>dicators were rather clear:<strong>Taiwan</strong>ese managers disclosed worse mental <strong>and</strong>physical <strong>health</strong> than the general workforce <strong>in</strong> thecountry. Although the comparison methodadopted may warrant some methodological concernsÐ simply compar<strong>in</strong>g the work <strong>stress</strong> <strong>and</strong><strong>health</strong> of managers <strong>in</strong> this study with other samples<strong>in</strong> the literature Ð the cross-sectional data suggestthat <strong>Taiwan</strong> managers may be under <strong>stress</strong> <strong>and</strong> riskof ill-<strong>health</strong>.In recent years, some research has demonstratedthat female managers are under higher<strong>stress</strong> than their male counterparts, 20±22 whereasother research yielded <strong>in</strong>conclusive results. 23In this study, a t-test revealed that <strong>Taiwan</strong>esefemale managers did report more <strong>stress</strong> relatedto the `managerial role' than their male counterparts(mean ˆ 15.20 vs 14.37, t ˆ 2.49, p 5 0.05),thus partially corroborat<strong>in</strong>g the previous studies.This seem<strong>in</strong>gly consistent pattern of femaledisadvantage <strong>in</strong> managerial positions is very<strong>in</strong>terest<strong>in</strong>g as it provides <strong>in</strong>direct support forthe role con¯ict thesis, which states that thesocially approved characteristics of fem<strong>in</strong><strong>in</strong>ity(such as nurturance <strong>and</strong> submission) may oftencome <strong>in</strong>to con¯ict with the requirements of a`mascul<strong>in</strong>e' managerial role (such as toughness<strong>and</strong> assertion). The prevail<strong>in</strong>g hostility aga<strong>in</strong>stfemales, the lack of female role models, unsupportivesuperiors (usually males), long-establishedbus<strong>in</strong>ess practices with sexual overtures, <strong>and</strong>the dilemma between career commitment <strong>and</strong>family duties may all compound to produce thehigh level of work <strong>stress</strong> encountered by femalemanagers <strong>in</strong> a paternalistic Ch<strong>in</strong>ese societysuch as <strong>Taiwan</strong>. 24,25 Hopefully, these ®nd<strong>in</strong>gswill serve to raise awareness of the unduly highwork <strong>stress</strong> endured by these Ch<strong>in</strong>ese femalemanagers <strong>and</strong>, <strong>in</strong> turn, facilitate steps to reduceits detrimental e€ects on <strong>health</strong> <strong>and</strong> workmorale.Structural relationships among work <strong>stress</strong>,moderators <strong>and</strong> stra<strong>in</strong>The structural relationships depicted <strong>in</strong> Fig. 1were largely borne out by correlational <strong>and</strong>regression analyses. More speci®cally, most of theCopyright # 1999 John Wiley & Sons, Ltd. Stress med. 15, 53±64 (1999)