Influence of Diversity on Conflicts, Turnover, and PerformanceDo Lim Choi & Yun Jeong Kim(Chungnam ational !niv"#IntroductionDiversity issue has been treated as an important element in organizationalmanagement in many advanced countries(Barsade, Ward et al. 2000; Pitts 200!. "sevidence, manypapersondiversity#erepresentedat20$2"merican%ociety&orPublic"dministration'on&erenceheldin(as)egas, *). +nitially, diversitymeantsomething to be related to demographic variables, a&&irmative action (""!, and e,ualemployment opportunity (--.!. +n other #ord, the term #as considered as somethinge,ual totheparticipationo& morecolorsand#omen. /here&ore, theconcept#asmore normative. +t means that more colors and #omen in organization aremeaning&ul. %ince $00s, according as organizations have the diverse #or1&orce, manypractitionersandscholarshavehadinterestsin#or1&orcediversity(WentlingandPalma23ivas $0; 4or2Bara1 $!. /he &ocus o& diversity has been e5panded &romsimple demographic characteristics to belie&, bac1ground, and &unction (/homas 6r$0; %lac1 $7; D8*etto and %ohal $; 9ardens#artz and 3o#e 20$0!. :nli1elyin the past, they #ould li1e to reveal the relationship bet#een diversity andorganizational per&ormance(Pelled, -isenhardt et al. $!. +t, ho#ever, is di&&icult to&ind empirical studies that e5amine the relationship bet#een diversity andper&ormanceinthepublicsector, althoughtherearesomestudiesintheprivatesector(Pitts 200;; Pitts 200!. +n the past decades, empirical studies have &ocused on direct relationship bet#eendemographic characteristic variables and per&ormance, but recently a group o&scholars tries to suggest more sophisticated model to e5plain diversity andper&ormance. or% years $ to $0 years $C 0.$$ to 20 years B0 EB.B2$ to C0 years ;; C7.74ore than C0 years $0 B.03unctional6ac%ground(ess than B BE E7.0B to $0 EB CE.C$$ to $; $$ 0.2$B to 20 7 ;.24ore than 20 B E.;?riginality DaeAeon BB EE.C'hungnam EB C0.'hungbu1 20 $C.Eothers $7 $$.BK> -ach demographic characteristic may have a &re,uency missing #hich indicatesthe number o& respondents that did not mar1 that ,uestion. 'redibility and 'orrelation/able 2sho#s means, standard deviation, 'ronbach alph, and correlationcoe&&icients. /hrough this table, one can obtain basic in&ormation about the thoughtso& respondents &or main variables and to a degree guess the results." 'ronbachalpha #as used to measure the LunidimensionalityD o& multiple ,uestions. "ll variablessho# reliability coe&&icients o& 0.7 or higher #hich con&irms the unidimensionality o&variables. "mong&our,uestionstomeasureemotional con&lict, one,uestion#asremoved, because respondents didnDt recognize it as the same dimension. 3egardingcorrelation, all variablesarestatistically signi&icant #itheachother.-motional con&lict variable is negatively correlated #ith tas1 con&lict and per&ormanceand positively #ith turnover. /as1 con&lict variable is negatively correlated #ithturnover and positively #ith per&ormance. Diversityingender, age, andoriginality#ill havestrongerpositiveassociations #ith emotional con&lict. "s sho#n in table C, gender and originality diversity variables #ere associated #ithemotionalcon&lict, but in the opposite direction. 9ender diversity #as a signi&icantvariable to increase emotional con&lict, but originality diversity #as negativelyassociated #ith emotional con&lict. 4aybe, the samples o& this study areinappropriate to measure the e&&ect o& originality diversity, because most o&respondents are &romDaeAeon(EE.CH! and'hungchung Province(EE.CH!. 0.070KK> signi&icant at 0.0; KKK> signi&icant at 0.0$4ore #or12related variables and tas1 con&lict&y'othesis*> Diversity in education, tenure in organization, and &unctionalbac1ground #ill have stronger positive associations #ith tas1 con&lict. "s sho#n in table E, &unctional bac1ground diversity and tenure diversity in publicservice #ere statistically associated #ith tas1 con&lict. Both variables are positivelyassociated. +t meansthat tas1con&lict increases#henemployeeshavediverse&unctional bac1ground e5perience. "lso, tas1 con&lict increases #hen employees havediverse tenure. "nd educationaldiversity #as not a variable to a&&ect tas1 con&lict./hus, hypothesis 2 #ould be partially accepted. I/able EJ 3elationships bet#een more #or12related variables and tas1 con&lictInde'endent varia6les De'endent varia6le, Tas% conflict$eta 9tandard error t(ducational diversity .0$CC .CBE7 .2E3unctional 6ac%ground diversity .200;; .0707 2.;$KKTenure diversity in 'u6lic service .$0$0B .00$27 2.$;KK< )alue> ;.20 32%,uare> 0.$00E "dA 32%,uare> 0.00$EKK> signi&icant at 0.0; KKK> signi&icant at 0.0$-&&ect on group longevity bet#een diversity variables and con&licts&y'othesis+,9roup longevity #ill diminish the positive associations bet#eendiversity variables and con&licts in #or1 groups. 9roup longevity #as not a signi&icant variable to mediate the relationships bet#eendiversity variablesandcon&lictsin#or1 groups. +nanother #ord, there#asnodi&&erence in the relationships bet#een diversity variables and con&licts according togroup longevity. 3elationship bet#een con&licts and turnover$0&y'othesis.,-motional con&lict #ill be positively associated #ith turnoverintention. &y'othesis/,/as1 con&lict #ill not be a signi&icant variable to in&luence turnoverintention. "ssho#nintable;, emotional con&lict#aspositivelyassociated#ithturnoverintention. +t means that employees have intention to leave current organization #henthey &eel moreemotional con&lict. /hus, hypothesisE#ouldbeaccepted. /as1con&lict #as not a signi&icant variable to in&luence turnover intention. /hus, hypothesis; #ould also be accepted. I/able ;J 3elationships bet#een emotional con&lict and per&ormanceInde'endent varia6les De'endent varia6le, Turnover$eta 9tandard error t(motional conflict .2C72$ .$2B$ 2.EEKKTas% conflict 2.0;00C .$22CC 20.B0< )alue> ;.0$32%,uare> 0.07;$ "dA 32%,uare> 0.0B22KK> signi&icant at 0.0; 3elationship bet#een con&licts and per&ormance &y'othesis0, /as1 con&lict #ill be positively associated #ith per&ormance. &y'othesis1, -motional con&lict #ill decrease organizational per&ormance."s sho#n in table B, both emotional and tas1 con&lict #ere statistically associated#ith per&ormance variable, butinthe oppositedirection.(i1ein hypothesis B,thehigher tas1 con&lict is, the greater organizational per&ormance is. .n the contrary, thehigher emotional con&lict is, the lo#er organizational per&ormance is. /hus,hypothesis B and 7 #ould be accepted. I/able BJ 3elationships bet#een emotional con&lict and per&ormanceInde'endent varia6les De'endent varia6le, Performance$eta 9tandard error t(motional conflict 2.$B$7E .072C 22.2KKTas% conflict .B$E20 .070E0 0.7$KKK< )alue> ;.0$32%,uare> 0.07;$ "dA 32%,uare> 0.0B22KK> signi&icant at 0.0; KKK> signi&icant at 0.0$$$Im'lications and Conclusion+nt#oaspects, theuse&ulnesso& thisstudycanbee5plained> theoretical andpractical. ;C02;E7.9ardens#artz, (. and ". 3o#e (20$0!. 4anaging Diversity> " 'omplete Des1 3e&erence R Planning9uide. %F34, %ociety 2$2C7.Pelled, (. F. ($B!. ODemographic Diversity, 'on&lict, and Wor1 9roup .utcomes> "n +nterveningProcess /heory.O .rganization %cience 1(B!> B$;2BC$.Pelled, (. F., N. 4. -isenhardt, et al. ($!. O-5ploring the Blac1 Bo5> "n "nalysis o& Wor1 9roupDiversity, 'on&lict, and Per&ormance.O "dministrative %cience Quarterly ..($!> $220.Pitts, D. W. (200;!. ODiversity, 3epresentation, and Per&ormance> -vidence about 3ace and -thnicityin Public .rganizations.O 6ournal o& Public "dministration 3esearch and /heory )/(E!> B$;2BC$.Pitts, D. W. (200!. ODiversity 4anagement, 6ob %atis&action, and Per&ormance> -vidence &rom :.%. C202CC0.%lac1, 6. D. ($7!. O %F+ 7;207./aA&el, F., 4. 9. Billig, et al. ($7$!. O%ocial categorizationandintergroupbehavior.O-uropean6ournal o& %ocial Psychology )> $E2$77./homas 6r, 3. 3. ($0!. O