Диссертация (1136702), страница 22
Текст из файла (страница 22)
Thus the ranges ofscores on the UWES-9 subscales do not vary across gender groups, and the subscaleshave the same relationship in male and female samples. Finally, the model of factormean invariance (Model 6a) also demonstrated an acceptable fit, and statistically didnot differ from Model 3a. These results suggest that the 3-factor model of the UWES9 is invariant across the male and female samples, and thus allows for the comparisonof male and female groups on work engagement.117Table 3. Mean values, standard deviations, and correlations between the scales of the Russian version of the UWES-9 and othervariablesM1.2.3.4.5.6.7.8.9.10.11.12.13.UWES-9VigorDedicationAbsorptionEmotional exhaustionDepersonalizationPersonal accomplishmentsJob satisfactionLife satisfactionTurnover intentionAgeLength of employment at organizationGeneral length of employment4.804.574.984.843.242.695.114.663.312.7336.366.9015.58SD1234567891011121.00 (.92)1.08 .90** (.79)1.14 .93** .79** (.87)1.08 .88** .68** .73** (.75)0.96 -.49** -.51** -.50** -.33** (.88)1.06 -.37** -.37** -.40** -.26** .67** (.80)0.80 .52** .45** .51** .47** -.39** -.31** (.84)1.15 .72** .63** .74** .60** -.52** -.43** .48** (.91)0.85 .35** .33** .36** .29** -.35** -.26** .33** .43** (.89)1.54 -.44** -.38** -.48** -.35** .48** .48** -.28** -.53** -.28**–9.55 .15** .16** .14** .11** -.10** -.08** .04.11** -.04-.14** –6.86 .06**.05* .05.06** .04.07** .01.03-.08** -.03 .54** –9.99 .13** .14** .12** .10** -.09** -.07**.05* .09** -.03-.11** .92** .54**Note.
*p < .05; **p < .01. On the diagonal in parentheses are values of the α-Cronbach.118Table 4. The fit of multi-group models of the modified 3-factor model of the UWES-9χ2pdfRMSEA[90% CI]CFITLISRMRMales/FemalesModel 1a (configural)Model 2a (metric)Model 3a (scalar)Model 4a (factor variance)Model 5a (factor variance and covariance)Model 6a (factor mean)345.23354.26371.12376.27381.08403.54< .001< .001< .001< .001< .001< .001445056596259.09 [.08–.10].08 [.08–.09].08 [.07–.09].08 [.07–.09].08 [.07–.08].08 [.08–.09].95.95.95.95.95.94.92.93.93.93.94.93.04.04.04.04.04.042a vs 1a3a vs 2a4a vs 3a5a vs 3a6a vs 3a4.87, p = .5606.09, p = .4120.58, p = .9009.77, p = .13538.99, p < .001Age groupsModel 1b (configural)Model 2b (metric)Model 3b (scalar)Model 4b (factor variance)Model 5b (factor variance and covariance)Model 6b (factor mean)401.63412.97470.02485.33485.04514.86< .001< .001< .001< .001< .001< .0016678909610296.10 [.09–.10].09 [.08–.09].09 [.08–.09].08 [.08–.09].08 [.08–.09].09 [.08–.09].94.94.94.94.94.93.91.92.92.93.93.92.04.04.05.07.07.062b vs 1b3b vs 2b4b vs 3b5b vs 3b6b vs 3bModelsModelcomparisonΔχ2, pΔdfΔCFI66363.004.002.001.001.00511.10, p = .520 1254.26, p < .001 1212.60, p = .050 616.78, p = .158 1247.77, p < .001 6.000.008.001.000.006Note.
df – degree of freedom; RMSEA – root mean square error of approximation; CFI – comparative fit index; TLI – Tucker Lewis index; SRMR –standardized root mean square residual; AIC – Akaike information criterion.119Second, a multi-group CFA was conducted across different age groups.Participants were divided into three age groups: under 30 (career start), 31-50 (inprime of working lives), and over 50 (past the peak of one’s career and approachingretirement). Models of configural invariance (Model 1b), metric invariance (Model2b), and scalar invariance (Model 3b) had an acceptable fit, but the Δχ 2 testdetermined a considerable deterioration in Model 3b fit in comparison to Model 2b.However, a ∆CFI was below the benchmark of .01; consequently, it was decided notto proceed to examine partial invariance and consider the following models nestedwithin the scalar invariance model (Model 3b).
The models of factor varianceinvariance (Model 4b) and factor covariance invariance (Model 5b) showed anacceptable fit and did not statistically differ from Model 3b. Thus the range of scoreson the UWES-9 subscales does not vary, and has the same relationship acrossdifferent age groups. The model of factor mean invariance (Model 6b) alsodemonstrated an acceptable fit and did not significantly differ from Model 3b.
Thismeans that the 3-factor model of the UWES-9 is invariant across the age groups andallow for comparison on work engagement.Convergent, Discriminant, and Incremental Validity of the Russian version ofthe UWES-9Table 3 demonstrates the correlations between the subscales of the UWES-9 andburnout. The UWES-9 subscales correlated slightly more strongly with personalaccomplishments (r ranging from .45 to .51), than with the other burnout subscales (rranges from -.26 to -.51). All subscales and the total UWES-9 score revealed asignificant positive correlation with job satisfaction (r ranges from .60 to .74) and lifesatisfaction (r ranges from .29 to .36), although the correlations with the former werestronger than with the latter. These results support the convergent and discriminantvalidity of the Russian version of the UWES-9.Finally, a series of hierarchical regression analyses were conducted, where jobsatisfaction, life satisfaction, and turnover intention were dependent variables.
Thefirst step included the burnout subscales in the regression equations; the second step120added the UWES-9 general score (we did not use the UWES-9 subscales because ofmulticollinearity problem — r between subscales ranging from .68 to .79). Theresults of the regression analyses are summarized in Table 5.
In all the cases, modelsincluding the UWES-9 predicted dependent variables more accurately. Wald testsshowed that the models in the second steps are significantly better than in the firststeps. Changes in the AICs and R2 from step 1 to step 2 confirm this conclusion. Thismeans that the UWES-9 added a significant amount of variance for predicting jobsatisfaction (20%), life satisfaction (2%), and turnover intention (5%). This supportsthe incremental validity of the Russian version of the UWES-9.Table 5. Hierarchical multiple regression modelsEmotional exhaustionDepersonalizationPersonalaccomplishmentsUWES-9InterceptFR2Job satisfactionStep 1Step 2-.385*** -.174***(.032)(.027)-.109*** -.081***(.028)(.023).471***.133***(.030)(.027).668***(0.023)3.799*** 1.563***(.197)(.177)326.32** 587.34****.362.577Life satisfactionStep 1Step 2-.244*** -.198***(.027)(.028).000.006(.024)(.023).265*** .192***(.025)(.027).145***(.023)2.743*** 2.257***(.164)(.179)133.88** 112.78****.189.207Turnover intentionStep 1Step 20.394***0.265***(0.046)(0.046)0.426***0.409***(0.040)(0.039)-0.146***0.060(0.042)(0.045)-0.407***(0.038)1.056***2.421***(0.281)(0.301)220.80*** 204.51***.277.322AIC461339063985394758495741ΔR2.215.018.045Wald test (Step 2 vs874.61***40.314**112.72***Step 1)*Note.
Standard errors in parentheses. ***p < .001. AIC – Akaike information criterion.Characteristics of the Russian version of the UWES-9 in different demographicand professional groupsAn additional descriptive analysis of the UWES-9 was held in different gender(male/female) and age (under 30, 31–50, and over 50) groups. A series of ANOVAswere conducted, in which the UWES-9 and its subscale scores were the dependentvariables. The analysis showed that the mean values of the UWES-9 scores and the121scores on the vigor and absorption subscales differed significantly between groups ofmen and women (see Table 6), with women, on average, having higher scores.However, these differences are extremely small (the effect size [ηp2] did not exceed.008). Employees from different age groups also differed on the mean level of workengagement (see Table 6).
Employees who are in their early career stages (under 30)displayed the lowest level of the overall UWES-9 and its subscales except on theabsorption subscale, whereas employees who are past the peak of their careers andapproaching retirement (over 50) had highest level. There is a linear relationshipbetween work engagement and age. However, age had only a minor effect on workengagement (the effect size [ηp2] ranges from .007 to .022). This was also confirmedby the correlations between the overall UWES-9 and its subscales and age, which arepositive and small (r ranges from .11 to .16).Table 6. Gender and age differences in UWES-9UWES-9VigorDedicationAbsorptionUWES-9VigorDedicationAbsorptionMenn = 516M (SD)4.70 (1.02)4.47 (1.09)4.95 (1.18)4.69 (1.11)Womenn = 1213M (SD)4.84 (1.00)4.62 (1.07)5.00 (1.12)4.91 (1.07)F (1, 1727)6.90 (p = .009)7.32 (p = .007)0.58 (p = .447)14.62 (p < .001)ηp20.0040.004<0.0010.008ηp2Gr. 1Under 30n = 554M (SD)Gr.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
