Enhancing Employee Engagement - link.springer.com978-3-319-54732-9/1.pdf · each questionnaire into...

31
APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS METHOD Four separate eld studies were conducted. Two of these studies were conducted in the southwestern United States and two other involved participants in Brazil. Participants in each sample were asked to answer a set of self- and social-report questionnaires via an online survey. The samples from each country are described below. For the studies based in Brazil, we followed the guidelines for multi-cultural research (Bracken & Barona, 1991; Brislin, 1970; Sechrest, Fay & Zaidi, 1972) and translated each questionnaire into Portuguese and then back-translated into English to assure an accurate presentation of the questions. The original survey was translated from English into Portuguese by one of the researchers who is a Portuguese native speaker. The survey was translated back from Portuguese to English by a business consultant who is also a Portuguese native speaker and uent in English. The versions were then compared in order to identify if any adjustment was necessary. PARTICIPANTS UNITED STATES SAMPLES Two eld studies were conducted in the southwestern United States. The rst study involved students enrolled in a professional MBA program. The second study involved employees of independent insurance agencies located throughout Texas. In each of these studies, participants were asked to voluntarily participate in a lengthy online survey of various aspects © The Author(s) 2017 J.L. Whittington et al., Enhancing Employee Engagement, DOI 10.1007/978-3-319-54732-9 109

Transcript of Enhancing Employee Engagement - link.springer.com978-3-319-54732-9/1.pdf · each questionnaire into...

APPENDIX A: DESCRIPTION OF THE STUDY

AND PROFILE OF PARTICIPANTS

METHOD

Four separate field studies were conducted. Two of these studies wereconducted in the southwestern United States and two other involvedparticipants in Brazil. Participants in each sample were asked to answer aset of self- and social-report questionnaires via an online survey. Thesamples from each country are described below. For the studies based inBrazil, we followed the guidelines for multi-cultural research (Bracken &Barona, 1991; Brislin, 1970; Sechrest, Fay & Zaidi, 1972) and translatedeach questionnaire into Portuguese and then back-translated into Englishto assure an accurate presentation of the questions. The original surveywas translated from English into Portuguese by one of the researchers whois a Portuguese native speaker. The survey was translated back fromPortuguese to English by a business consultant who is also a Portuguesenative speaker and fluent in English. The versions were then compared inorder to identify if any adjustment was necessary.

PARTICIPANTS UNITED STATES SAMPLES

Two field studies were conducted in the southwestern United States. Thefirst study involved students enrolled in a professional MBA program. Thesecond study involved employees of independent insurance agencieslocated throughout Texas. In each of these studies, participants wereasked to voluntarily participate in a lengthy online survey of various aspects

© The Author(s) 2017J.L. Whittington et al., Enhancing Employee Engagement,DOI 10.1007/978-3-319-54732-9

109

of their workplace. As an incentive to participate in the survey, participantswho completed the survey were entered into a random drawing to receiveone of several iPads. These two subsamples provided a total of 389participants in a combined United States subsample. The majority ofthese respondents were employed full time (87%), female (62%), andWhite/Caucasian (68%).

Graduate Student Sample: One hundred and seventy-two respon-dents (N = 172) fully completed the questionnaire. The majority weremale (60%), and White/Caucasian (46%). Most participants (71%)reported to be working in a full-time job. The majority of respondents(63%) reported to be Christian, while 10% reported to be Muslim, 9%Hindu, and 10% reported to belong to other religious backgrounds.

Insurance Agency Sample: Two hundred and sixty (N = 260) employ-ees fully completed the questionnaire. The majority of the respondentswere female (78%), White/Caucasian (82%) and reported to have a bache-lor’s degree or some college (76%). Most participants (54%) worked injobs related to customer service and support. A vast majority of partici-pants (87%) reported to follow a Christian faith.

BRAZILIAN PARTICIPANTS’ SAMPLES

Two separate field studies were conducted in Brazil. The first studyinvolved employees from 14 different organizations. The second studywas conducted with employees from a large rental car company thatoperates in all regions of Brazil. These studies provided a combined sampleof three hundred and sixty-seven participants (N = 367) in Brazil who fullycompleted the questionnaire. The majority of these respondents weremale (57%), White/Caucasian (78%), and reported to have a bachelor’sdegree (69%).

Brazil Sample A – Multi-organizational Sample: This sample hadparticipants from 14 different organizations in various industries, such ascivil construction, logistics, legal, safety, consulting, accounting, andretail. The majority (90%) were employed in full time in jobs. Job cate-gories represented in this sample included sales (20%), operations (17%),and administrative roles (14%). Most of the respondents were male (56%)and White/Caucasian (80%), with a bachelor’s degree (69%). The major-ity of the participants (76%) indicated Christian as their religiousaffiliation.

110 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

Brazil Sample B – Car Rental Company: The majority of respondentswere male (58%), and White/Caucasian (77%). Additionally, the majorityof participants (69%) reported to have a college degree and to practice aChristian faith. Most participants (93%) reported to be working full time.Within the organization, most participants (85%) reported to work at thebranches, while the remainder worked at the headquarters office. In termsof tenure, the majority of respondents (37%) reported to be with theorganization for 2 to 5 years. These employees were deployed in a varietyof organizational roles. Most participants (29%) reported to work in sales,17% in administrative roles, and 11% in customer service.

RESULTS

Means, standard deviations, and correlations between the constructsexplored in this study are presented in Table A.1. The correlation coeffi-cients below the diagonal are from the combined US sample. The scoresabove the diagonal are from the combined Brazilian sample. Means andstandard deviations for the US sample are presented in the columns and forthe Brazil sample they are presented in the two first lines below the tableheadings. Reliability coefficients are presented in the diagonal of the corre-lation matrix. Reliabilities for the US sample are depicted before the doubledashes and for the Brazil sample they are indicated after the double dashes.

We used a variety of self- and social-reports; however, all of our mea-sures were obtained from a single source. Because we were concernedabout common source variance, a Harman single factor analysis wasperformed. A common method bias is a potential problem when thevariance is attributable to the measurement method instead of the con-structs the measure is supposed to represent. A common method biaswould be an issue if the single factor explained majority of the variance.The Harman single factor based on the US model accounted for 41% ofthe variance and for Brazil it was 35%. Therefore, we concluded that ourresults were not subject to a common source bias.

MEASURES

Meaningful work was measured using items from the Work and MeaningInventory (WAMI) (Steger, Dik, & Duffy, 2012). This self-report instru-ment assesses three dimensions of meaningfulness: positive meaning

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 111

Tab

leA.1

Correlatio

n,Means,S

tand

ardDeviatio

nsandReliabilities-Braziland

US

BRAZI

L

Mean

Std.

Deviation

NWAM

IPM

NG

OCB

LM

XTFL

CR

AFF

CON

NOR

SATIS

FHRVC

ENGAGE

MRS

MPS

Mean

4.14

3.55

4.13

4.92

5.12

4.92

5.25

4.30

4.47

4.86

3.50

4.09

4.09

54.01

Std.

Deviation

0.69

0.79

0.36

1.09

1.08

1.47

0.96

0.86

0.91

0.75

0.60

0.82

0.82

22.18

N39

336

735

333

333

633

432

232

232

232

031

931

631

636

3UNIT

ED

STATES

WAM

I3.85

0.74

559

0.89

//0.91

0.40

8**

0.36

7**

0.26

0**

0.35

6**

0.20

5**

0.51

9**

–0.02

50.37

8**

0.47

9**

0.40

7**

0.24

9**

0.24

9**

0.38

9**

PM

NG

3.53

0.90

504

0.51

0**

0.90

//0.92

0.21

6**

0.42

9**

0.48

9**

0.45

0**

0.52

3**

0.13

2*0.35

1**

0.64

6**

0.61

6**

0.19

5**

0.19

5**

0.38

5**

OCB

4.12

0.44

487

0.42

9**

0.26

9**

0.90

//0.83

0.19

7**

0.23

1**

0.09

10.45

0**

0.03

40.34

3**

0.31

0**

0.23

5**

0.23

9**

0.23

9**

0.25

3**

LM

X5.29

1.27

464

0.46

6**

0.54

1**

0.39

1**

0.91

//0.80

0.72

5**

0.58

8**

0.34

0**

0.17

6**

0.16

9**

0.53

4**

0.38

4**

0.09

90.09

90.35

4**

TFL

5.03

1.17

467

0.49

4**

0.62

6**

0.38

9**

0.84

2**

0.96

//0.95

0.69

7**

0.39

7**

0.18

6**

0.27

4**

0.65

6**

0.46

5**

0.15

6**

0.15

6**

0.28

8**

CR

4.90

1.45

467

0.37

6**

0.55

8**

0.26

3**

0.70

4**

0.74

4**

0.93

//0.92

0.26

6**

0.11

3*0.13

0*0.56

8**

0.37

2**

0.11

8*0.11

8*0.24

8**

AFF

4.87

1.28

433

0.58

6**

0.57

2**

0.41

2**

0.55

6**

0.58

4**

0.46

6**

0.88

//0.80

0.19

3**

0.53

7**

0.60

4**

0.53

4**

0.23

1**

0.23

1**

0.34

2**

CON

4.31

1.02

433

0.01

50.06

8−0.03

0.01

70.03

8−0.04

0.14

5**

0.69

//0.57

0.18

1**

0.13

2*0.21

9**

0.17

1**

0.17

1**

−0.09

5

NOR

4.57

1.06

433

0.41

8**

0.31

7**

0.29

6**

0.34

5**

0.36

2**

0.25

7**

0.66

1**

0.24

3**

0.79

//0.70

0.41

8**

0.42

9**

0.33

1**

0.33

1**

0.08

6

112 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

SATIS

F4.81

0.93

433

0.61

2**

0.68

8**

0.39

2**

0.69

0**

0.73

6**

0.63

8**

0.76

3**

0.02

50.43

9**

0.94

//0.91

0.70

2**

0.19

7**

0.19

7**

0.42

1**

HRVC

3.37

0.68

434

0.44

8**

0.61

6**

0.27

0**

0.49

4**

0.59

2**

0.50

6**

0.54

2**

0.04

30.36

8**

0.69

0**

0.95

//0.93

0.18

3**

0.18

3**

0.30

2**

ENGAGE

4.10

0.63

434

0.64

9**

0.42

6**

0.62

8**

0.40

0**

0.44

1**

0.29

2**

0.57

2**

0.06

70.43

0**

0.53

6**

0.41

2**

0.96

//0.95

1.00

0**

0.05

4M

RS

4.03

0.92

429

0.30

4**

0.16

6**

0.25

4**

0.10

4*0.11

2*0.07

90.20

2**

-0.032

0.28

7**

0.17

0**

0.18

1**

0.26

5**

0.98

//0.96

0.05

4

MPS

62.98

27.59

488

0.24

8**

0.20

5**

0.23

1**

0.20

4**

0.17

4**

0.14

8**

0.29

4**

0.01

80.19

4**

0.29

5**

0.18

6**

0.26

5**

0.03

8N/A

**Correlatio

nissign

ificant

atthe0.01

level(2-tailed)

*Correlatio

nissign

ificant

atthe0.05

level(2-tailed)

0.00

//–UNIT

ED

STATES

//0.00

–BRAZIL

UNIT

ED

STATES

BRAZIL

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 113

(PM), meaning making through work (MMTW), and greater good moti-vations (GGM). PM captures the idea that individuals evaluate their workto matter and to be meaningful. A sample question from the PM scales is“I have found a meaningful career.” MMTW translates the concept ofindividuals having a broader context of their work. This was measuredthrough statements such as, “My work helps me make sense of the worldaround me.” Finally, GGM reflects the perspective that work has moremeaning if it impacts other individuals (Steger et al., 2012). The GGMsubscale included questions such as “The work I do serves a greaterpurpose.”

Each item in the WAMI scale was measured on a 5-point Likert scaleranging from 1 (Absolutely Untrue) to 5 (Absolutely True). PM,MMTW, and GGM were measured with 4, 3, and 3 items, respectively.We did not differentiate among these subscales in our theoretical model.Therefore, we submitted the data to confirmatory factor analysis tovalidate a single construct for meaningfulness. The results from theconfirmatory factor analysis supported this. We created a total score formeaningfulness by averaging the responses to the three dimensions. Thisscale was used in our analyses. The meaningfulness engagement scale hada reliability of 0.89

Engagement was measured using a self-report scale consisting of 18items from Rich et al. (2010). This measure captures three distinct dimen-sions: physical engagement, emotional engagement, and cognitiveengagement. Physical engagement refers to the energy exerted and thephysical involvement within the job. This was assessed through questionssuch as “I devote a lot of energy to my job.” Emotional engagementapprehends the level of pleasantness and enthusiasm experienced at work.This was captured through statements such as “I am enthusiastic in myjob.” Finally, cognitive engagement measures the level and intensity offocus and concentration applied while performing work-related tasks.Cognitive engagement was measured through statements like “At work,I pay a lot of attention to my job.” Each item was measured with a 5-pointLikert scale that ranged from 1 (Strongly Disagree) to 5 (Strongly Agree).We did not differentiate among the engagement subscales; therefore, wesubmitted the data to confirmatory factor analysis to validate a single,global scale for engagement. Based on the supportive results of thisanalysis, we created a total score for engagement by averaging theresponses to the three dimensions. This scale was used in our analyses.The engagement scale used in our analyses had a reliability of 0.88.

114 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

Integrated faith was measured using 14 items from the MatureReligiosity Scale developed by Pieper, Van Uden, and De Vries-Schot(2012). This scale captures the degree to which an individual’s spiritualityis integrated throughout all dimensions of life. A 5-point Likert scaleranging from 1 (Strongly Disagree) to 5 (Strongly Agree) was used tomeasure responses to statements such as “I have the idea that I entrustmyself more and more to God or a Higher Power,” “The meaning andsignificance of my life is in my relationship with God”, “My faith is orientedto values that transcend physical and social needs”, and “I believe sincerely,not mainly out of obligation or fear.” A total score for faith integration wasobtained by averaging responses across all of these items. The reliabilitycoefficient for faith integration is 0.97 in our combined data set.

The Human Resource Value Chain (HRVC) captures the full set ofmanagerial practices that encompass the entire relationship between orga-nizations and employees from recruitment, selection, orientation, andsocialization, passing through compensation and career development,until employee separation. A scale of 18 items was created for this parti-cular study based on Galpin, Whittington, and Bell (2012). These itemsare provided in Table A.2. Participants were asked to evaluate their orga-nization’s HR practices on a 5-point Likert scale ranging from 1 (StronglyDisagree) to 5 (Strongly Agree).

An exploratory factor analysis revealed three distinct factors within theHRVC responses. The first factor includes eight questions that were designedto evaluate the general HR practices of the firm. This dimension was capturedthrough statements such as, “The HR department is responsive to the needsof the business.” The second dimension included six items designed toevaluate the organization’s support of the employee’s development. One ofthe statements used in this case was “The HRDepartment assists me with mycareer management objectives.” A score for this factor was obtained byaveraging responses from six items. The third dimension focused on respon-dents’ perceptions of how well the organization handled negative events suchas disciplinary actions and severance. This dimension was measured throughstatements like “Employee disciplinary actions are handled fairly.”

The hypotheses developed in our model did not differentiate amongthe HRVC subscales individually. Therefore, we submitted the data toconfirmatory factor analysis to validate a unidimensional scale. Results ofthe CFA are presented in Figures A.1 and A.2. The results from theconfirmatory factor analysis strongly supported our expectations for US,Brazil, and the total sample. While only the US sample statistics are

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 115

discussed here, the results for Brazil and the total sample are presented inTable A.3. The fit statistics suggested a well-fit model (Comparative fitindex = 0.957; Tucker-Lewis index = 0.95; residual mean square error =0.037). The chi-square for the model was 261.26 (p < .001). However,the significance of the chi-square is not surprising given that our sample islarge (Kenny, 2015). We compared the unidimensional model with themulti-dimensional model to determine the best-fit model and as can beseen in Tables A.4 and A.5 the unidimensional model was a better fit.Table A.6 summarizes the results of these CFA analyses. Therefore, we

Table A.2 Human Resources Value Chain Items

General HR Practices

The HR Practices in my organization are integrated with the company’s strategy.

The HR department is responsive to the needs of the business.

Communications about changes in HR policy are communicated in a timely manner.

The HR Department does a good job of administering the company’s benefit package.

There is adequate information about new job opportunities within the organization.

The company has an effective employee assistance program.

The HR Department supports the business with succession planning.

The HR department is responsive to my requests for assistance.

Organization’s support of the employee’s development

Our HR department recruits an adequate applicant pool for job openings.

The HR department does a good job of screening potential applicants.

The performance evaluation process is effectively managed by the HR Department.

The Organization supports my professional development needs.

The HR Department assists me with my career management objectives.

I receive adequate training for my job.

Handling Negative Events

Employee’s personal issues are handled with discretion.

Employee disciplinary actions are handled fairly.

The organization provides adequate severance pay for employees who are terminated.

The organization assists employees in finding new employment when there is a downsizingevent.

116 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

.86

HRVC23.7

HRVC13.7

HRVC73.7

HRVC113.9

HRVC123.1

HRVC133.6

HRVC163.4

HRVC203.8

HRVC33.5

HRVC43.5

HRVC63

HRVC83.6

HRVC92.8

HRVC103.4

HRVC143.6

HRVC153.6

HRVC173.6

HRVC18 3.7

1

.3

2

.36

3

.37

4

.45

5

.5

6

.5

.83.8

.79

.74HR_Practice

1.71

.71

.7

.72

7

.51

.898

.48

9

.43

10

.41

11

.49

12

.48

13

.4

.75

.77

.71

Emp_Develop1

.72

.77

.69

14

.53.78

15

.45

16

.48

17

.69

.74

.72

Hand_nEvents1

.56

.56

18

.68

Fig. A.1 CFA of HRVC for US Sample as Multi-Construct

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 117

HRVC14.7

HRVC24

HRVC74.1

HRVC113.9

HRVC123.5

HRVC133.7

3.7HRVC16

HRVC204.4

HRVC33.8

HRVC43.7

HRVC63.3

HRVC83.3

HRVC93.1

HRVC103.3

HRVC144.6

HRVC154.5

HRVC175

HRVC183.3

1 .51

2 .4

3 .55

4 .54

5 .55

6 .61

7 .54

8 .62

9 .51

.56

10

.51

11 .59

12 .5

13

.42

14

.65

15 .8

.34

16

.65

.7

.78.67

.68.67

.63.68

.61

.7

.7

.64

.7

.76.59.45.59.43.22

HRVC1

17 .81

–.16

18

.95

Fig. A.2 CFA for HRVC US Sample as Unidimensional Construct

118 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

Table A.3 Performance Management System Items

Goal Characteristics

Specific and clear goals.

Goals for all key areas relating to their job performance.

Challenging but reasonable goals (neither too hard or too easy).

Deadlines for accomplishing their goals.

Outcomes and Rewards

Performance appraisals are based on objectives set in the performance planning session.

Rewards (i.e., pay, promotions) allocated to them according to howwell they reach their goals.

Goals reflect organization-wide initiatives as well as individual performance.

Rewards are tied to organizational achievement as well as individual performance.

Support

Sufficient skills and training to achieve their goals.

Sufficient resources (i.e. time, money, equipment) to achieve their goals.

Frequent feedback on how well they are progressing toward their goals.

VoiceThe opportunity to participate in setting their goals.

A say in deciding how to implement their goals.

Table A.4 CFA of HRVC as Multi-Construct

USA Brazil Total

RMSEA 0.064 0.081 0.068

SRMR 0.043 0.061 0.044

CFI 0.927 0.864 0.909

TLI 0.915 0.843 0.895

AIC 16765.84 12012.82 29085.27

χ2 358.29*** 391.01*** 568.73***

df 132 132 132

n 415 297 712

*p < 0.05; **p < 0.01; ***p < 0.001

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 119

created a total score for HRVC by averaging the three dimensions. Oursubsequent analyses utilize a single scale to represent the HRVC. The HRvalue chain scale used in our analyses had a reliability of 0.87.

Performance Management (PMNG) captures to what extent the orga-nization plans, evaluates, and contributes to enhance individuals’ perfor-mance. A scale of thirteen items was created for this particular study basedon Galpin et al. (2012). Each of these items is listed in Table A.3. Responses

Table A.5 CFA of HRVC as Single Construct

USA Brazil Total

RMSEA 0.049 0.058 0.053

SRMR 0.037 0.050 0.038

CFI 0.957 0.933 0.946

TLI 0.950 0.921 0.937

AIC 16631.744 11835.684 26833.514

χ2 261.26*** 258.66*** 389.26***

df 130 130 130

n 415 297 712

Results reported in table 1 and 2 are based on Satorra Bentler testingRMSEA = Root mean square error of estimationSRMR = Standard root mean squared residualCFI = Comparative fit indexTLI = Tucker-Lewis indexAIC = Akaike’s information criteriondf = Degrees of freedom*p < 0.05; **p < 0.01; ***p < 0.001

Table A.6 CFA Model Comparison

US1 US2 Brazil1 Brazil2 Total1 Total2

χ2 261.26 358.29 258.66 391.01 389.26 568.73delta χ2 97.03*** 132.35*** 179.47***

*p < 0.05; **p < 0.01; ***p < 0.0011 = single-construct2 = multi-construct

120 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

were recorded on a 5-point Likert scale ranging from 1 (Hardly Ever) to 5(Almost Always).

An exploratory factor analysis identified four distinct factors within thePMNG scale responses. The first dimension evaluated the characteristics ofthe individual’s performance goals, identifying if they were clear, specific,and attainable. This dimension was measured based on four questions,such as “In my organization we have challenging but reasonable goals(neither too hard nor too easy).” Responses to the four items wereaveraged in order to obtain an overall score for the dimension. The secondfacet evaluated outcomes and rewards, capturing the congruence betweenindividual goals and organizational objectives, as well as between goals andindividual rewards and recognition. We used four questions to evaluatethis dimension, such as, “Rewards are tied to organizational achievementas well as individual performance.” Responses to the four items wereaveraged in order to obtain a score for this dimension. The third factorof the PMNG system encompasses perceived organizational support. Itcaptures individuals’ perception of support from the organization,through skill development, available resources, and feedback in order toperform accordingly. Perceived organizational support was measuredthrough three questions, such as “In my organization individuals receivefrequent feedback on how well they are progressing toward their goals.”The fourth dimension involved two questions concerning the employee’sopportunity to participate in the goal-setting and implementation pro-cesses. The statements used to capture this dimension were “In my orga-nization individuals have the opportunity to participate in setting theirgoals” and “In my organization individuals have a say in deciding how toimplement their goals.”

The hypotheses developed in our model did not differentiate amongthe PMNG subscales individually. Therefore, we submitted the data toconfirmatory factor analysis to validate a single, global scale. The resultsfrom the confirmatory factor analysis strongly supported our expectations.

Results of the CFA are presented in Figures A.3 and A.4. While only theUS sample statistics are discussed here, the results for Brazil and the totalsample are presented in Table A.7. The fit statistics suggested a well-fitmodel (Comparative fit index = 0.968; Tucker-Lewis index = 0.955;residual mean square error = 0.065). The chi-square for the model was168.90 (p < .001). As discussed earlier, the statistical significance of thechi-square was expected as the sample size is large and for large sample sizeit is a known fact that chi-square tends to be significant (Kenny, 2015). We

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 121

PMNG13.3

PMNG23.3

PMNG33.5

PMNG63.5

PMNG102.7

PMNG112.6

PMNG12 3

PMNG132.8

PMNG73.3

PMNG83.3

PMNG92.7

PMNG42.8

PMNG52.9

2 .22

3 .19

4

.32

5 .61

7

.45

8

.42

9

.23

10

.33

12

.34

.88

.9

.82

.63

.74

.76

.88

.82

.82

1 .19

Goal

6 .22

Outcomes

11 .15

.9

.89

.92

PMNG1

13 .37

14

.35

16

.19

17

.17

.79

.81

.9

.91

Support

15 .34

Voice

.82

Fig. A.3 CFA of PMNG for US Sample as Multi-Construct

122 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

PMNG13.3

PMNG23.3

PMNG33.5

PMNG63.5

PMNG102.7

PMNG112.6

PMNG12 3

PMNG132.8

PMNG73.3

PMNG83.3

PMNG92.7

PMNG42.8

PMNG52.9

1 .36

.45

2

.35

.2–.24

3

.37.011

–.21 4

.8.62

.8

.85

.41

.62

.776

.52

.69

.17

.41

PMNG7

.381

8

.49

9

.48

.44

10

.51

.79

.72

.72

.7.81

.72

.72

11

.35

–.02912

.6213

.48

.48

Fig. A.4 CFA for PMNG US Sample as Unidimensional Construct

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 123

compared the unidimensional model with the multi-dimensional model todetermine the best-fit model and as can be seen in Table A.8 and A.9, theunidimensional model was a better fit. Therefore, we created a total scorefor PMNG by averaging the four dimensions and proceeded in our

Table A.7 CFA of PMNG as Multi-Construct

USA Brazil Total

RMSEA 0.072 0.084 0.067

SRMR 0.038 0.054 0.038

CFI 0.957 0.921 0.957

TLI 0.945 .899 0.945

AIC 15627.156 11523.441 27352.001

χ2 217.68*** 209.22*** 285.85***

df 61 61 61

n 490 343 833

*p < 0.05; **p < 0.01; ***p < 0.001

Table A.8 CFA of PMNG as Single Construct

USA Brazil Total

RMSEA 0.065 0.064 0.056SRMR 0.033 0.041 0.031CFI 0.968 0.958 0.973TLI 0.955 0.941 0.962AIC 15574.476 11433.73 27240.947χ2 168.90*** 133.34*** 196.16***df 55 55 55n 490 343 833

Results reported in table 1 and 2 are based on Satorra Bentler testingRMSEA = Root mean square error of estimationSRMR = Standard root mean squared residualCFI = Comparative fit indexTLI = Tucker-Lewis indexAIC = Akaike’s information criteriondf = Degrees of freedom*p < 0.05; **p < 0.01; ***p < 0.001

124 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

analysis using a single scale to represent PMNG. This PMNG scale used inour analyses had a reliability of 0.88.

Enriched Jobs. Participants were asked to evaluate the characteristicsby responding to items from the job characteristics inventory developed bySims, Szilagyi, and Keller (1976). Responses were placed on a 5-pointLikert scale ranging from 1 (Very Little) to 5 (A Great Deal). Each of thefive core job dimensions (task variety, task significance, task identity,autonomy, and feedback) was measured with two items. FollowingHackman and Oldham (1975), we used scores on the five core jobdimensions to calculate a motivating potential score (MPS). MPS wascalculated as follows: MPS = (Skill Variety + Task Identity + TaskSignificance)/3*Autonomy*Feedback.

Leader Member Exchange (LMX) was measured with a six-item scaledeveloped by Liden et al. (1993). Employees were asked to evaluate thequality of their relationship with their manager using a 7-point Likert scaleranging from 1 (Strongly Disagree) to 7 (Strongly Agree). This scaleincludes statements such as, “My supervisor recognizes my potential”and “I usually know where I stand with my supervisor.” The reliabilityfor this scale is 0.87.

Transformational leadership (TFL) was measured using twenty twoitems from the scale developed by Podsakoff, MacKenzie, Moorman, andFetter (1990). Participants were asked to describe the behavior of theirimmediate supervisor on a 7-point Likert scale ranging from 1 (StronglyDisagree) to 7 (Strongly Agree). Transformational leadership was cap-tured through statements like “My manager inspires others with his/herplans for the future.” Our exploratory factor analysis revealed a perfectsimple structure reflecting 6 distinct factors in our TFL scale, in line with

Table A.9 CFA Model Comparison

US1 US2 Brazil1 Brazil2 Total1 Total2

χ2 168.90 217.68 133.34 209.22 196.16 285.85

delta χ2 48.78*** 75.88*** 89.69***

*p < 0.05; **p < 0.01; ***p < 0.0011 = single-construct2 = multi-construct

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 125

results presented by Podsakoff et al. (1990). However, our model did notdifferentiate among the transformational leadership subscales. Therefore,we submitted the data to confirmatory factor analysis to validate a single,global scale. The results from the confirmatory factor analysis stronglysupported our expectations. Therefore, a total score for transformationalleadership was created by averaging the responses from the six dimensions.We proceeded in our analysis using this single scale to represent transfor-mational leadership. The transformational leadership scale used in theseanalyses had a reliability of 0.86.

Contingent reward (CR) occurs when a leader identifies perfor-mance requirements and clarifies the conditions under which rewardsare available for meeting these requirements. Based on these conditionsand requirements, the leader acknowledges when good performancetakes place. CR was measured by averaging responses of five items fromPodsakoff et al. (1990). It was captured through statements such as, “Mymanager personally complements me when I do outstanding work.”Responses were placed on a 7-point Likert scale ranging from 1(Strongly Disagree) to 7 (Strongly Agree). Cronbach’s alpha for theCR scale in this study was 0.93.

Job satisfaction captures how much employees like or dislike their jobs(Spector 1997). Job satisfaction is a state-like emotional evaluation basedon feelings and thoughts that one has toward various aspects of his or herjob (Weiss, 2002). Job satisfaction was measured using 33 items fromSpector (1997). This measure captures the attitudinal perspective of anindividual’s perception of the job from nine different facets: pay, promo-tion, supervision, benefits, rewards, operating conditions, coworkers,work itself, and communication. These facets were captured throughstatements such as, “I am satisfied with my chances for promotion,” “Ifeel that the work I do is appreciated,” and “I feel a sense of pride in doingmy job.” Responses were placed on a 7-point Likert scale ranging from 1(Strongly Disagree) to 7 (Strongly Agree). Following Spector (1997), wecreated a total score for job satisfaction that was created by averaging thescores for the nine subscales. The total job satisfaction scale had a relia-bility of 0.88.

Affective commitment was measured by averaging responses to theeight-item affective commitment scale developed by Allen and Meyer(1996). Responses were placed on a 7-point Likert scale ranging from 1(Strongly Disagree) to 7 (Strongly Agree). Affective commitment refers tothe employee’s “emotional attachment to, identification with, and

126 APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS

involvement in the organization” (Allen & Meyer, 1996:11). Affectivecommitment reflects the idea that individuals stay within the organizationbecause they “want to.” It was captured through statements such as “Thisorganization has a great deal of personal meaning for me.” The affectivecommitment scale had a reliability of 0.86.

Continuance commitment was measured through the average resultof responses to the 8-item continuance commitment scale (Allen &Meyer, 1996). Responses were placed on a 7-point Likert scale rangingfrom 1 (Strongly Disagree) to 7 (Strongly Agree). Continuance commit-ment captures one’s perception of the costs involved in leaving the orga-nization (Allen & Meyer, 1997). This facet depicts the idea thatindividuals remain in an organization because they cannot afford toleave. A sample continuance commitment item is “Right now, stayingwith my organization is a matter of necessity as much as desire.” Thecontinuance commitment scale had a relatively low level or reliability(Cronbach’s alpha = 0.64).

Organizational citizenship behavior (OCB) was measured using 24items from a self-report questionnaire developed by Podsakoff et al. (1990).The original OCB scale identified five factors: altruism, conscientiousness,civic virtue, courtesy, and sportsmanship. Responses were recorded on a5-point Likert scale ranging from 1 (Strongly Disagree) to 5 (StronglyAgree). The various dimensions of OCB were captured through statementssuch as, “I believe in giving an honest day’s work for an honest day’s pay,”“I attend meetings that are not mandatory, but are considered important,”“I am mindful of how my behavior affects other people’s jobs,” and“I willingly help others who have work related problems.” We did notdifferentiate among the OCB subscales. Therefore, we developed a totalscore for OCB by averaging the scores from the five dimensions. Thecombined OCB scale had an acceptable level of reliability 0.80.

APPENDIX A: DESCRIPTION OF THE STUDY AND PROFILE OF PARTICIPANTS 127

APPENDIX B: TESTING THE FULL MODEL USING

STRUCTURAL EQUATIONS MODELING (SEM)

Figure B.1 shows our initial conceptual model. This conceptualmodel was tested using structural equation modeling (SEM).Following Bollen (1990), we tested the overall fit of the model inFig. B.2 using multiple indices. While the chi square and root meansquare residuals (RMSE) are usually used to test the fit of the model,they are significantly influenced by the sample size. The comparativefit index (CFI) and Tucker–Lewis index (TLI) are less susceptible andrelatively stable for small to moderate sample sizes (Hu & Bentler,1995). Also, since the data are not normally distributed, the SEM wassubjected to the Satorra–Bentler scaled chi square test for modelgoodness of fit vs. the saturated model. This is done becauseSatorra–Bentler chi square test is robust to non-normality (Yuan,Chan, & Bentler, 2000). The SEM for the initial theoretical modelpresents the following results: χ2 = 2599 (p < 0.0001); RMSEA =0.09; SRMR = 0.17; CFI = 0.79; and TLI = 0.77.

Using the modification index option of STATA in SEM and elim-inating those variables that had no impact on the theoretical model,we removed both quality of leader–follower relationship (LMX) andcontingent reward (CR) variables. We subjected this model to morerestrictions, including covariates, as indicated by the modificationindex and theoretical justifications.

The alternative SEM model in Fig. B.3 presents the following results:χ2 = 1822 (p < 0.0001); RMSEA = 0.08; SRMR = 0.10; CFI = 0.86 and

© The Author(s) 2017J.L. Whittington et al., Enhancing Employee Engagement,DOI 10.1007/978-3-319-54732-9

129

An

tece

den

ts o

fE

ng

agem

ent:

HR

Val

ue

Ch

ain

P

erfo

rman

ceM

anag

emen

tS

yste

m

En

gag

emen

t

Em

plo

yee

Ou

tco

mes

:

Job

Sat

isfa

ctio

nO

rgan

izat

ion

alC

om

mit

men

t O

rgan

izat

ion

alC

itiz

ensh

ipB

ehav

ior

(OC

B)

An

tece

den

ts o

fM

ean

ing

fuln

ess:

Mea

nin

g a

t W

ork

:

Tra

nsf

orm

atio

nal

Lea

der

ship

Mea

nin

g i

n W

ork

:

Job

Des

ign

Bri

ng

ing

Mea

nin

g t

o W

ork

:

Inte

gra

ted

Fai

th

Mea

nin

gfu

lnes

s

Co

nti

ng

ent

Rew

ard

Qu

alit

y o

fL

ead

er-F

ollo

wer

Rel

atio

nsh

ip

• •

• • •

• • •

Fig.

B.1

Con

ceptualM

odel

ofMeaning

fulnessandEmploy

eeEng

agem

ent

130 APPENDIX B:TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . .

1 .047

CO

RE

_TF

L

3.6

.98

6

.15

7

.36

8

.42

17

.77

18

.42

19

.77

PA

Y2.

721

.6

.47

2 .7

2 T

FL_

D4 3.7

3 .3

5 T

FL_

D5

4 .3

9 T

FL_

D6 3.2

2.2

MP

S1

.4

CR

.53

.81 .7

8

–.05

7

TF

L_L

.46 –.1

PM

MM

TW

WA

MI_

L

5

.56

GG

M

.76

PH

YS

EM

O

EN

GA

GE

_L

16 .2

CO

G

.48

.99 S

AT

ISF

_L

20

.024

.5

.77 .3

9

.62

PR

OM 2.

5

SU

P

BE

N3.

3

RE

W3.

1

OP

S

22

.67

23

.47

24

.75 .2

9

25

.41

26

.85

.29 .4

6

.39

.54

.1.7

1

.068 .0

88

LMX

1

4.1 1

4.3

.72

.67 .1

3

.24

.27

.36

.79

.044

.84

.77

3.4

CO

W4.

6

.26

27

.61

MR

S1

WO

RK 3.

928

.2

9

9 .2

9P

MN

G_D

1 4

10 .2

8P

MN

G_D

2

.84

.85

.48

CO

MM

3

AF

F

29

.4

30

.37

3.2

11 .2

5P

MN

G_D

33.

5.8

6 .78

PM

NG

_L

.68

HR

VC

_L

.9.8

8.7

5

3.2

CO

N4.

231

1

12 .4

PM

NG

_D4 3

HR

VC

_D1 4.

5H

RV

C_D

2 4.1

HR

VC

_D3 4.

7C

ON

SC 7.4

33

.45

13

.18

Chi

Squ

are

= 2

598.

60 (

p<

0.00

1), d

f = 6

02

RM

SE

A =

0.0

9S

RM

R =

0.1

7

14

.22

15

.43

OC

B_L

.74 .57

.63

SP

OR

T 6

CIV

IC35

.6

CF

I = 0

.79

TLI

= 0

.77

32

.77

.82

.67

6.2

CO

UR

T 8

ALT

R7.

5

36

.32 .3

8

37

.55

Fig.

B.2

Con

ceptualM

odel

with

Results–US

APPENDIX B: TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . . 131

1 .0

57

CO

RE

_TF

L3.

8.9

76

.1

27

.3

48

.4

117

.7

418

.3

719

.7

3

PA

Y2.

721

.6

.46

2 .7

4 T

FL_

D4 3.

8

3 .3

7 T

FL_

D5

.51

.8

TFL_

LP

MM

MTW

GG

MP

HY

SE

MO

CO

G

PR

OM 2.

5

SU

P

22

.65

23

.39

.3

.46

4 .4

2 T

FL_

D6 3.

4

.76

.3

.32

WA

MI_

L

.77

EN

GA

GE

_L

.52

1.2

BE

N

.53.

324

.7

5 .3.36

.094

.68

.57 .35

.084

5

.59

.099

.26

–.78

16 .2

6

SA

TIS

F_L

20 .2 .7

9

.047

.79 .3

8

.61

.79

.78

RE

W3.

1

OP

S 3.3

CO

W4.

6

WO

RK

4

25

.37

26

.85 .2

6

27

.63

28

.38

9 .3

1P

MN

G_D

1 4.1

10

.3P

MN

G_D

2

.83

.84

.51

CO

MM

3

AF

F

29

.39

30

.38

3.3

11 .2

8P

MN

G_D

3 3.7

.85 .7

6PM

NG

_LH

RV

C_L

.6

.91

.88

.75

3.3

CO

N4.

231

1

12 .4

3P

MN

G_D

4 3.1

HR

VC

_D1

4.5

HR

VC

_D2

4.1

HR

VC

_D3

4.7

CO

NS

C 7.4

33 .4

5

13

.17

Chi

Squ

are

= 1

821.

80 (

p<

0.00

1), d

f = 5

36

RM

SE

A =

0.0

8S

RM

R =

0.1

0

14

.23

15

.44

OC

B_L

.74 .58

.63

SP

OR

T 6

CIV

IC35

.6

1

CF

I = 0

.86

TLI

= 0

.85

32

.74

.82

.67

6.2

CO

UR

T 8

ALT

R 7.5

36 .3

3 .39

37.5

5

2.3

MP

S1

4.3

MR

S1

Fig.

B.3

Alte

rnativeMod

elwith

Results–US

132 APPENDIX B:TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . .

TLI = 0.85. This is still not a perfect model, but this model has a better fitthan our initial model.

There is significant debate among researchers and statisticiansabout the acceptable fit indexes. While many earlier researchers havecalled for χ2 to be insignificant with an RMSEA value of less than0.10 and CFI values of greater than 0.90 (Hu & Bentler, 1995;Browne & Cudex, 1993), researchers have also cautioned againstusage of such strict cutoff indexes to determine the fit of the model(Marsh, Hau, & Wen, 2004; Weston & Gore, 2006). Weston andGore (2006) suggest a look at the sample size and complexity of themodel to determine acceptable fit indexes, especially when theRMSEA values are between 0.05 and 0.10 and SRMR values arebetween 0.08 and 0.10 with more stringent criteria applied if themodel is simple and the sample size is large and less stringent criteriaif the model is complex and the sample size is less than 500, both ofwhich apply to our model presented in Fig. B.5. Thus, although thechi square statistic is still significant, the alternative model presentedis considered to be a good fit given other supportive indexes.

Table B.1 shows the results of the model comparisons and theresulting chi square difference test. The objective of this differencetest is to determine if the reduction in chi square is significant, and ascan be seen the delta between the two models is very significant withthe model in Fig. B.2 having lesser chi square. This indicates thatthe theoretical model has non-significant paths and paths that do notadd to the understanding of factors leading to employee engagement

Table B.1 SEM Model Comparison

Model χ2 RMSEA CFI TLI AIC delta χ2

Theoretical(Figure 4)

2599*** 0.09 0.79 0.77 35397

Alternative(Figure 5)

1822*** 0.08 0.86 0.85 32933 777***

Results reported are based on Satorra–Bentler testingRMSEA = Root mean square error of estimationCFI = Comparative fit indexTLI = Tucker–Lewis indexAIC = Akaike’s information criterion*p < 0.05; **p< 0.01: ***p< 0.001

APPENDIX B: TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . . 133

as compared to the alternative model. This can also be inferred fromthe lower Akaike’s information criterion (AIC) for the alternativemodel compared to the theoretical model.

Having established that the alternative model presented in Fig. B.3is an acceptable fit, the research question was to see if the same modelof employee engagement would hold true in a different culturalsetting. To test this, we used the data collected from Brazil and ranan SEM of the alternative model and the results are presented inFig. B.4. As can be seen from Table B.2, both models have similargoodness of fit index. This indicates that the employee engagementmodel presented in this chapter may be applicable across culturalsettings.

As expected, we found a significant path between transformationalleadership (TFL) and meaningfulness (WAMI) in both samples(US = .46, p < 0.0001; Brazil = .34, p < 0.001). The presence of anenriched job (MPS) was also significantly related to work meaningfulnessin both samples (US = .32, p < 0.0001; Brazil = .11, p < 0.001). We alsofound integrated faith (MRS) to be positively related to meaningfulness;however, the beta weights were weak at best for both countries(US = .084, p < 0.001; Brazil = 0.048, p < 0.001).

The analysis also indicates that the organization’s performance man-agement system (PMNG) is a significant predictor of engagement(ENGAGE) in both countries (US = .099, p < 0.0001; Brazil = .17,p < 0.001). The human resource value chain is also significantly relatedto the level of employee engagement (HRVC; US = .26, p < 0.0001;Brazil = .27, p < 0.001).

As expected, there is a strong relationship between meaningfulness andengagement in both samples (US = 0.87, p < 0.001; Brazil = 1.0, p <0.001). In turn, engagement had a significant positive relationship withoverall job satisfaction (US = 1.2, p < 0.001; Brazil = 1.4, p < 0.001),affective commitment (US = .79, p < 0.001; Brazil = .76, p < 0.001), andorganizational citizenship behavior (US = .51, p < 0.0001; Brazil = .52,p < 0.001).

As shown in Fig. B.5, the results are generally consistent acrossthe cultures. However, there are some differences between the USmodel and the Brazil model in the significance of the individualrelationship paths. For example, while in the US sample the pathbetween performance management (PMNG) and engagement(ENGAGE) was weak at best (β = 0.08, p < 0.05), it was moderately

134 APPENDIX B:TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . .

1 .1

7C

OR

E_T

FL 4

.91

6

.24

7

.23

8

.33

17

.81

18

.37

19

.8

PA

Y2.

721

.6

9 .44

2 .5

9 T

FL_

D4 4.

4

3 .4

2 T

FL_

D5

.64

.76 .8

4

TFL_

LP

MM

MT

WG

GM

.82

PH

YS

EM

OC

OG

.45

PR

OM

2.6

SU

P

22

.47

23

.45

.32 .4

8

.35

4 .3

TFL_

D6 3.

9.2

1

.11

WA

MI_

LE

NG

AG

E_L

1.4

.49

BE

N3.

424

.7

6

.23

.018

.53

.45

.24

.048

5

.85

.17

.27

–.76

16 .45

SA

TIS

F_L

20 .63 .7

6

.14

.77 .1

9

.42

.7

.65

RE

W3.

3

OP

S2.

8

CO

W5.

6

WO

RK

6.4

25

.41

26

.96 .2

9

27

.82

28

.51

9 .4

6P

MN

G_D

1 5.1

10 .3

2P

MN

G_D

2 3.6

11 .2

7P

MN

G_D

3 3.9

.74

.83

.85 .6

6

PM

NG

_L

.65

HR

VC

_L

.89

.88

.72

.52

CO

MM

5.2

AF

F5.

4

CO

N4.

9

29

.58

30

.42

31

.98

12 .5

7P

MN

G_D

4 3.1

HR

VC

_D1 5.4

HR

VC

_D2 4.4

HR

VC

_D3 6.6

CO

NS

C 7.7

33

.73

13

.21

Chi

Squ

are

= 1

385.

88 (

p<

0.00

1), d

f = 5

36

RM

SE

A =

0.0

7S

RM

R =

0.1

1C

FI =

0.8

5T

LI =

0.8

4

14

.23

15

.49

OC

B_L 32

.73

.52 .44

.75 .7

6

.69

SP

OR

T 7.6

CIV

IC7.

6

CO

UR

T 9.4

ALT

R9.

1

35

.43

36

.43 .1

5

37

.53

2.4

MP

S1 5

MR

S1

Fig.

B.4

Alte

rnativeMod

elwith

Results–Brazil

APPENDIX B: TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . . 135

1C

OR

E_T

FL

2T

FL_

D4

3T

FL_

D5

TF

L_L

6

PM

US

0.4

6***

7

MM

TW

8

GG

M

17

PH

YS

US

0.8

7***

18

EM

O

19

CO

G

PA

Y21

PR

OM

22

SU

P23

4T

FL_

D6

WA

MI_

LE

NG

AG

E_L

Bra

zil 1

.4**

*B

EN

24

US

0.3

2***

Bra

zil 0

.11*

**5

US

0.0

8***

Bra

zil 0

.04*

*

16S

AT

ISF

_L

20

RE

W25

OP

S26

CO

W27

9P

MN

G_D

1

Bra

zil 0

.17*

**

US

0.2

6***

Bra

zil 0

.27*

**

US

0.7

9***

WO

RK

28

CO

MM

29

10 11

PM

NG

_D2

PM

NG

_D3

PM

NG

_LH

RV

C_L

Bra

zil 0

.76*

**

US

0.4

7(ns

)B

razi

l 0.1

4***

AF

F30

CO

N31

12P

MN

G_D

4H

RV

C_D

1H

RV

C_D

2H

RV

C_D

3B

razi

l 0.5

2***

CO

NS

C33

1314

15S

PO

RT

34

US

Sam

ple

Bra

zil S

ampl

eC

hi S

quar

e =

182

1.80

(p

<0.

001)

, df =

536

R

MS

EA

= 0

.08

CF

I = 0

.86

TLI

= 0

.85

Chi

Squ

are

= 1

385.

88 (

p<

0.00

1), d

f = 5

36

RM

SE

A =

0.0

7

CF

I = 0

.85

TLI

= 0

.84

OC

B_L

32

CIV

IC35

CO

UR

T36

ALT

R37

MP

S

MR

S

Fig.

B.5

Com

parativ

eResultsof

theUnitedStates

andBrazil

136 APPENDIX B:TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . .

stronger in the Brazil sample (β = 0.14, p < 0.001). On the otherhand, in the Brazil sample, the path between enriched jobs (MPS) andmeaningful work (WAMI) was relatively weak (β = 0.11, p < 0.001),while it was moderately stronger in the US sample (β = 0.31, p <0.001). Similarly, we found the link between integrated faith (MRS) tobe significant, but relatively weak in both countries. However, thisrelationship was much weaker in Brazil (US = .084, p < 0.001;Brazil = 0.048, p < 0.001). Of interest in our study was the fact thatthe path between engagement (ENGAGE) and continuance commit-ment (CON) was not significant in the US sample but turned out tohave a weak but significant relationship in the Brazil sample (β = 0.1,p < 0.01).

Table B.2 Overall SEM Fit Results US and Brazil Sample

Model χ2 RMSEA CFI TLI

US 1822*** 0.08 0.86 0.85Brazil 1386*** 0.07 0.85 0.84

Results reported are based on Satorra–Bentler testingRMSEA = Root mean square error of estimationCFI = Comparative fit indexTLI = Tucker–Lewis index*p < 0.05; **p < 0.01: ***p < 0.001

APPENDIX B: TESTING THE FULL MODEL USING STRUCTURAL EQUATIONS . . . 137

INDEX

AAppreciation language, 33

CCalling, 13, 14, 54–58, 62, 106, 107

See also Work; Work orientationCulture

country culture, 92, 93cross-cultural, 6, 15cultural differences, 91, 99, 107culture-brazil, 91, 92, 93, 107culture-united states, 91, 92, 93,

99, 107dimensions of culture, 93, 94See also Organizational culture

EEngagement (employee)

antecedents, 3, 4, 5, 14, 26, 54, 62,99, 105

benefits, 11consequences, 4, 24costs, 1, 33defined; trait approaches, 2; state

approaches, 2, 3, 4, 23disengagement, 1

global issue, 5, 15job satisfaction, 4, 13, 24, 25, 26,

37, 98measures, 1, 2, 3, 9, 23, 103organizational citizenship behavior

(OCB), 107, 126organizational commitment, 2,

13, 14performance, 2–5, 13, 14, 32, 47,

68, 77, 81, 82, 85, 88, 97, 98See also Gallup Q12

FFaith

and calling, 14, 58, 62, 63extrinsic religiosity, 55integrated faith, 5, 14, 54, 55–58,

59, 60, 62, 67, 106, 126, 129intrinsic religiosity, 58, 62and meaningful work, 5, 14,

54, 129See also Religiosity

GGallup Q12, 3

© The Author(s) 2017J.L. Whittington et al., Enhancing Employee Engagement,DOI 10.1007/978-3-319-54732-9

139

HHuman resource value chain

components of; careermanagement, 72–75;compensation systems, 73;employee development, 70, 73;employee separation, 68, 74,75; HR planning, 68, 70;job analysis, 68, 70; jobrequirements, 68, 70;job specifications, 68;organizational entry, 68, 70,71; orientation, 70, 72, 75;realistic job preview, 71;recruitment, 70, 72, 75, 104;selection, 70, 71, 72, 75, 104;socialization, 70, 72, 75;training, 72, 74, 84, 86

and engagement, 67, 68, 71, 74–79,86, 97

measure of HR value chain, 75, 76model of HR value chain, 10

IIndividual differences, 14, 97

personality, 97See also Integrated faith; Work

orientation

JJob

job characteristics, 11, 12, 14, 43,44, 46–48, 55, 105, 106

job characteristics inventory, 47job characteristics measure, 47job design, 99, 107job enrichment, 47–50, 55job redesign, 55, 106See also Task

Job satisfactionfacets of job satisfaction, 121,

126, 127measure of, 4

LLeadership

4i’s of transformationalleadership, 33; idealizedinfluence, 35; individualizedconsideration, 35, 36;inspirational motivation, 35;intellectual stimulation, 35, 36

contingent reward, 15full-range leadership, 10, 77measure of, 77transactional leadership, 34, 77–78,

94, 97transformational leadership, 5, 11, 13,

31, 32, 34–40, 43, 49, 55, 62,67, 77, 94, 97, 105, 106, 107

transformational leadership–measure, 77

MMeaning; meaningfulness

bringing meaning to work, 53, 62defined, 13, 14, 15meaning at work, 5, 14, 26, 31, 32meaningful work, 5, 14, 19, 20, 22,

34, 47, 54, 71, 105, 108meaning in work, 43, 106measure, 22

OOrganizational citizenship behavior

(ocb)Components, 127

140 INDEX

defined, 33measure, 47

Organizational commitmentaffective commitment, 24, 25, 26,

37, 47continuance commitment, 98defined, 13, 14measure, 46

Organizational culture, 71leadership and, 77–78, 79

PPerformance management system

components of, 15and engagement, 5, 14, 15, 67, 78,

81, 87, 88, 97, 98feedback, 3, 14, 82, 83, 84, 85, 98implementation, 14measure, 47performance evaluation

(appraisal), 82–85, 98performance planning, 83–86, 98

Positive organizational scholarship(POS), 11, 21

Positive organizationspositive communication, 32, 33positive culture, 31positive leadership, 34positive relationships, 32, 33

RReligiosity

extrinsic religiosity, 55

intrinsic religiosity, 58, 62See also Faith

SSpirituality, 5, 14, 28, 53, 54, 58, 60

See also Faith; Religiosity

TTask, 3, 14, 23, 31, 44, 46, 47, 55, 62,

68, 82, 85, 94, 106, 108

VVirtue, 20, 21, 32Virtuous organization, 21, 22

WWork

as worship, 14, 57, 58, 62,63, 107–108

See also Calling; FaithWork orientation

calling, 14, 54–58, 62, 63, 106, 108career, 55, 56, 58, 72–75, 106craftsmanship, 56, 58job, 2, 4, 5, 9, 11, 12, 13, 14,

23–27, 31, 37, 43–51, 53, 55,56, 61, 67–73, 82–84, 86, 91,94, 97–99, 105–108

kinship, 56–58, 63serving, 31, 56–58, 108

INDEX 141