Professional Documents
Culture Documents
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/211380792
CITATIONS
READS
58
334
3 authors, including:
Aliosha Alexandrov
University of Wisconsin - Oshkosh
7 PUBLICATIONS 195 CITATIONS
SEE PROFILE
Ugur Yavas
East Tennessee State University
Keywords:
357
358
FIGURE 1
Research Model
Perceived Management
Concern for
Employees
(CE)
Research Model
H1a
Job
Satisfaction
(JS)
H1b (+)
H3 (-)
Turnover
Intentions
(TI)
H5 (+)
H2a (+)
Concern for
Customers
(CC)
H2b (+)
H4 (-)
Affective
Organizational
Commitment
(AOC)
Moderator:
Employment Status
(full-time vs. part-time)
Control Variable: Tenure
Appraisal
Emotional Response
Behavior
NOTE: Moderator hypotheses (H6H9), which predict varying strengths of the relationships in the model, are not shown to keep the diagram simple.
359
360
361
362
METHOD
Sample and Data Collection
Data for this study were collected as part of a comprehensive employee and customer survey sponsored by
a national retail chain. Store managers distributed 8,500
questionnaires, with a cover letter from the company
CEO, to frontline employees in 1,200 stores across 25
states that the retailer operates in. Store managers were
instructed, by a letter from the company CEO, not to
pressure respondents in any way. Employees were
assured of confidentiality and allowed to respond to the
questionnaire anonymously during work hours by keying
in their responses electronically via the companys internal e-mail system. Participation was totally voluntary.
The responses were transferred into a data file at the corporate office under the supervision of a member of the
research team. After a 3-week period, 5,591 usable
responses, for a response rate of 66%, were obtained. The
number of responses across the stores ranged from 1 to
20 with an average of 5 responses per store (there was no
response from 50 stores). About 63% of the respondents
were full-time employees. Seventy-six percent of fulltime and 79% of part-time employees were male. The
average tenure was 2.1 years for full-time and 1.1 years
for part-time employees. The relatively low average
tenure is a reflection of rapid growth (opening of new
stores) experienced by the retailer. A comparison of the
363
TABLE 1
Confirmatory Factor Analysis and Measurement Properties of the Scales
Scale
t-Value
.74
.76
.77
.75
.79
.77
.79
62.2
65.0
66.3
64.6
69.1
66.5
69.9
.77
.79
.72
.63
.72
.69
65.5
67.6
59.5
49.8
59.2
56.8
Job satisfactionc
1. Coworkers
2. Store managers
3. Teamwork
4. Pay
.77
.75
.79
.53
63.1
60.9
65.5
39.6
.69
.80
.71
.75
.67
56.3
69.1
58.0
62.8
54.2
Turnover intentionse
1. I frequently think of quitting my job
2. I am thinking about leaving my company
.83
.85
55.9
57.0
NOTE: Model fit statistics for metric invariance test: X2509 = 3,811.19, root mean square error of approximation = .053, goodness-of-fit index = .94, normed
fit index = .98, non-normed fit index = .98, comparative fit index = .99. Item scores range from 1 (strongly disagree) to 5 (strongly agree).
a. Full-time (FT): = .90, average variance extracted (AVE) = .57, 2 = .06.49; part-time (PT): = .92, AVE = .61, 2 = .04.49.
b. FT: = .86, AVE = .51, 2 = .08-.45; PT: = .87, AVE = .54, 2 = .03-.52.
c. FT: = .78, AVE = .50, 2 = .12-.41; PT: = .81, AVE = .53, 2 = .08-.50.
d. FT: = .84, AVE = .53, 2 = .26-.49; PT: = .84, AVE = .53, 2 = .15-.52.
e. FT: = .83, AVE = .71, 2 = .06-.26; PT: = .82, AVE = .70, 2 = .03-.15.
RESULTS
Measurement Results
The dimensionality, convergent, and discriminant
validity of the measures were initially assessed via a series
364
three items each in job satisfaction and affective organizational commitment scales did not perform well (i.e.,
factor loadings were below the cutoff value of .50). One
item in the TURNOVER INTENTIONS SCALE heavily
loaded on a different factor. When the same analysis was
repeated using full-time and part-time sample data separately, similar results emerged. In addition, item-to-total
correlations identified the same set of items as weak contributors to their respective scales (corrected item-to-total
correlations < .50). Therefore, these items were discarded
and the exploratory factor and reliability analyses were
repeated using the remaining 24 items. The results indicated that all scale items loaded heavily on their respective underlying factors and none of the items showed a
large cross-loading. These results were consistent across
the subsamples and the total sample with slight differences in the magnitudes of the loading estimates.2
For each group, the remaining 24 items were subjected to a confirmatory factor analysis with a five-factor
measurement model using the sample covariance matrix
as input to LISREL 8.51 (Jreskog and Srbom 1993).
The fit statistics indicated that the measurement models
fit the data well for both groups (full-time employee sample: 2242 = 2,436.72, root mean square error of approximation [RMSEA] = .053, goodness-of-fit index [GFI] =
.94, normed fit index [NFI] = .98, non-normed fit index
[NNFI] = .98, comparative fit index [CFI] = .98, and parttime employee sample: 2242 = 1,340.07, RMSEA = .050,
GFI = .94, NFI = .99, NNFI = .99, CFI = .99).
Next, we examined measurement invariance following
the steps suggested by Hair et al. (2006). We first tested
the measurement model by simultaneously analyzing the
sample covariance matrices of full-time and part-time
groups with no restrictions. This result (2485 = 3,777.08)
served as the basis for testing measurement invariance.
Second, we repeated the analysis with equality constraints imposed only on factor loadings (2509 =
3,811.19). A 2 difference test at the .05 level between the
base model (with no equality constraints) and the
restricted model indicated no significant deterioration in
model fit. This result suggests that the measures exhibit
metric invariance across the two groups of frontline
employees. Third, we repeated the analysis with the additional constraint of equal covariance matrices across the
two groups (2519 = 3,871.54). The 2 difference test indicates a significant deterioration in model fit with the
additional constraint. This means that at least one of the
relationships among model variables will be significantly
different between full-time and part-time employees, a
requisite initial evidence for justifying the tests of
specific moderator hypotheses. Table 1 presents metric
invariance results, the list of items making up each scale,
scale reliabilities, average variance extracted (AVE), and
the range of shared variances (2).
365
1.000
.640
.489
.454
.614
.492
.332
.334
.337
.334
.366
.357
.344
.398
.380
.349
.378
.325
.288
.292
.264
.321
.154
.136
.143
.679
1.000
.550
.432
.579
.498
.378
.398
.377
.363
.412
.395
.401
.408
.425
.390
.424
.366
.291
.318
.287
.351
.171
.165
.103
CC2
3.96
.807
4.04
.785
4.00 4.09
.794 .835
3.96
.818
3.90
.797
.467
.474
.540
1.000
.398
.470
.305
.322
.314
.322
.358
.361
.349
.366
.334
.341
.329
.307
.267
.278
.195
.296
.113
.094
.091
CC4
3.75
.910
3.66
.933
.635
.604
.503
.453
1.000
.444
.373
.331
.381
.347
.350
.332
.338
.390
.350
.334
.354
.310
.273
.269
.268
.286
.139
.137
.137
CC5
4.10
.802
4.00
.846
.530
.535
.559
.504
.489
1.000
.346
.408
.357
.352
.426
.431
.398
.400
.384
.396
.411
.365
.310
.348
.240
.343
.192
.173
.098
CC6
3.35
1.035
3.29
1.065
.377
.407
.356
.335
.419
.330
1.000
.548
.680
.592
.525
.513
.544
.395
.394
.337
.404
.360
.297
.348
.387
.367
.201
.208
.118
CE1
3.81
.885
3.82
.905
.406
.452
.439
.421
.420
.442
.553
1.000
.534
.515
.635
.578
.597
.350
.365
.369
.360
.344
.282
.298
.301
.336
.158
.169
.066
CE2
3.43
.982
3.34
1.063
.381
.413
.384
.348
.424
.358
.710
.566
1.000
.655
.535
.546
.568
.409
.394
.324
.397
.372
.308
.354
.384
.374
.228
.223
.125
CE3
3.57
.966
3.49
1.032
.365
.402
.374
.338
.377
.355
.588
.567
.649
1.000
.570
.544
.580
.389
.386
.332
.391
.361
.310
.350
.349
.376
.198
.178
.091
CE4
3.85
.885
3.84
.900
.418
.452
.464
.382
.424
.430
.527
.650
.576
.613
1.000
.642
.638
.402
.392
.387
.391
.375
.297
.339
.306
.348
.166
.182
.069
CE5
3.72
.924
3.72
.959
.368
.428
.428
.363
.409
.415
.536
.605
.611
.586
.664
1.000
.628
.390
.382
.372
.365
.351
.282
.330
.302
.342
.168
.177
.056
CE6
3.80
.886
3.73
.941
.412
.460
.478
.394
.431
.435
.540
.674
.579
.595
.694
.664
1.000
.397
.392
.370
.388
.360
.341
.369
.323
.428
.194
.178
.070
CE7
3.70
.911
3.65
.950
.421
.431
.409
.362
.373
.396
.372
.371
.373
.364
.392
.406
.397
1.000
.539
.517
.495
.461
.294
.317
.288
.335
.254
.239
.104
AOC1
3.84
.864
3.81
.898
.426
.431
.422
.362
.358
.416
.362
.381
.373
.361
.395
.372
.420
.567
1.000
.600
.632
.537
.346
.382
.383
.385
.402
.397
.094
AOC2
3.88
.860
4.02
.815
.393
.411
.373
.366
.355
.357
.340
.379
.341
.316
.381
.354
.387
.502
.540
1.000
.520
.475
.296
.300
.251
.306
.286
.299
.049
AOC3
3.91
.921
3.83
.959
.404
.423
.397
.370
.365
.394
.378
.387
.376
.359
.408
.377
.421
.483
.598
.523
1.000
.499
.338
.411
.384
.391
.348
.350
.117
AOC4
.390
.411
.383
.322
.366
.380
.358
.395
.378
.374
.383
.381
.433
.360
.407
.362
.398
.364
1.000
.561
.351
.626
.187
.167
.130
JS1
.363
.402
.399
.334
.358
.393
.374
.403
.413
.408
.423
.422
.458
.401
.425
.363
.426
.375
.627
1.000
.386
.561
.250
.241
.135
JS2
.369
.367
.357
.319
.322
.323
.375
.374
.370
.354
.402
.371
.384
.487
.532
.489
.496
1.000
.314
.355
.318
.353
.254
.291
.089
AOC5
3.30
.995
3.21
1.017
.321
.333
.304
.271
.343
.284
.422
.324
.386
.343
.321
.348
.344
.324
.397
.301
.379
.340
.387
.413
1.000
.367
.233
.209
.137
JS3
.118
.098
.093
.082
.062
.087
.124
.112
.106
.111
.131
.117
.109
.154
.277
.195
.236
.218
.163
.194
.154
.185
1.000
.717
.105
TI1
3.74 2.53
.983 1.181
3.52 2.57
1.014 1.195
.394
.429
.405
.366
.390
.411
.410
.434
.417
.388
.420
.420
.483
.390
.440
.391
.407
.391
.657
.579
.411
1.000
.234
.200
.120
JS4
2.57
1.173
2.57
1.190
.116
.125
.122
.101
.086
.115
.144
.123
.115
.106
.122
.121
.129
.190
.283
.240
.260
.242
.153
.205
.150
.190
.694
1.000
.117
TI2
1.00
1.156
2.06
1.156
.102
.081
.061
.054
.053
.083
.050
.024
.060
.070
.026
.006
.036
.071
.036
.006
.074
.057
.094
.102
.055
.103
.090
.023
1.000
TEN
NOTE: Perceptual measures were recorded on a 5-point scale, where a higher number indicates a more favorable rating. CC = Perceived management concern for customers; CE = Perceived management concern for employees; AOC = Affective organizational commitment; JS = Job satisfaction; TI = Turnover intentions; TEN = Tenure. Above the diagonal are the correlations for full-time, and below the diagonal are the correlations for part-time
frontline employees.
M
SD
.547
.568
1.000
.492
.456
.547
.364
.415
.375
.352
.434
.421
.399
.416
.375
.385
.393
.342
.288
.320
.253
.341
.179
.183
.102
CC3
3.97 4.04
.812 .845
Part-time (n = 2,057)
M
SD
Full-time (n = 3,534)
CC1
CC2
CC3
CC4
CC5
CC6
CE1
CE2
CE3
CE4
CE5
CE6
CE7
AOC1
AOC2
AOC3
AOC4
AOC5
JS1
JS2
JS3
JS4
TI1
TI2
TEN
CC1
TABLE 2
Correlations, Means, and Standard Deviations of the Observed Variables
366
TABLE 3
Tests of Structural Model and Research Hypotheses
Two-Group Resultsa
(nFT = 3,534, nPT = 2,057)
Combined Sample
Results (n = 5,591)
Structural Model
Parameter
TEN
TEN
CE
CC
TEN
CE
CC
JS
TEN
JS
AOC
CE
CC
TEN
Total effects of:
CE on TI
CC on TI
Model fit statistics:
Standardized Estimate
CE
CC
JS
JS
JS
AOC
AOC
AOC
AOC
TI
TI
TI
TI
TI
(1)b
(2)b
(1)
(2)b
(3)
(3)b
(4)
(5)
(4)
(6)
(7)b
(8)c
(9)c
(5)b
.07
.12
.41
.32
.10
.23
.38
.29
.03
.08
.62
.06
.23
.05
t-Value
8.4
5.0
20.9
16.3
8.22
12.4
20.0
16.1
3.2
3.3
20.7
2.39
9.0
3.5
.01
.02
Common Metric
Standardized Estimate
.47
.61
.26
.19
8.4
.08
3.7
2 = 3,572.20, df = 261, RMSEA = .053,
GFI = .94, NFI = .98, NNFI = .98, CFI = .99
(FT/PT)
.04 / .17
.09 / .24
.41 / .42
.29 / .36
.07 / .11
.25 / .17
.39 / .35
.28 / .34
.02 / .01
.08 / .11
.66 / .53
.05 / .08
.23 / .24
.04 / .15
t-Value
(FT/PT)
2.5 / 7.3
5.0 / 10.4
16.7 / 13.9
10.6 / 12.0
4.4 / 5.8
10.9 / 5.7
16.9 / 11.5
12.7 / 11.0
1.7 / .6
2.7 / .2.5
18.3 / 10.8
1.69 / 2.0
7.2 / 5.4
2.09 / 6.2
R2
(FT/PT)
.01 / .03
.01 / .06
.42 / .56
.61 / .63
.30 / .21
.22 / .13
7.9 / 4.9
.10 / .05
3.6 / 1.9
2 = 4,758.16, df = 543, RMSEA = .055,
GFI = .93, NFI = .98, NNFI = .98, CFI = .98
NOTE: Measurement parameter estimates were almost identical to the results from confirmatory factor analysis. Therefore, they are not presented in
this table. FT/PT = Full-time/part-time; CC = Perceived management concern for customers; CE = Perceived management concern for employees; JS
= Job satisfaction; AOC = Affective organizational commitment; TI = Turnover intentions; TEN = Tenure; RMSEA = root mean square error of approximation; GFI = goodness-of-fit index; NFI = normed fit index; NNFI = non-normed fit index; CFI = comparative fit index.
a. Two-group simultaneous analysis with all structural model parameters set free across groups.
b. Significant difference between full-time and part-time frontline employees.
c. Not hypothesized.
DISCUSSION
We developed and tested a model to investigate the
effects of two critical dimensions of psychological climate (perceived management concern for employees and
customers) on frontline employees turnover intentions
and examined their differential effects across full-time and
part-time employees. Our model proved to be viable, and
all of the overall hypotheses and three of the seven
specific hypotheses concerning the moderating role of
employment status were supported.
Our study yields some insightful findings and makes
several contributions. First, our findings lend credence to
367
368
369
NOTES
1. It should be noted that when the variability of climate perceptions
across individuals is low, they also reflect organizational climate
(Schneider and White 2004). For the purposes of our study, the individual (frontline retail employee) is the unit of theory and analysis because
our focus is on the affective and behavioral consequences of psychological climate.
2. A set of tables summarizing the results of exploratory factor
analyses is available from the authors. We used relatively stringent criteria for retaining a scale item (i.e., factor loading > .50 and corrected
item-to-total correlation > .50) to prepare the scales for the examination
of measurement invariance across part-time and full-time groups via
confirmatory factor analysis.
370
REFERENCES
Abbott, Geoffrey N., Fiona A. White, and Margaret A. Charles (2005),
Linking Values and Organizational Commitment: A Correlational
and Experimental Investigation in Two Organizations, Journal of
Occupational and Organizational Psychology, 78 (4), 531-551.
ACAS (1994), Absence and Labor Turnover. London, UK: ACAS
Handbook.
Allen, David G. and Rodger W. Griffeth (2001), Test of a Mediated
Performance-Turnover Relationship Highlighting the Moderating
Roles of Visibility and Reward Contingency, Journal of Applied
Psychology, 86 (5), 1014-1021.
Allen, David G., Lynn M. Shore, and Rodger W. Griffeth (2003), The
Role of Perceived Organizational Support and Supportive Human
Resource Practices in the Turnover Process, Journal of
Management, 29 (1), 99-118.
Anderson, James C. and David W. Gerbing (1988), Structural
Equation Modeling in Practice: A Review and Recommended TwoStep Approach, Psychological Bulletin, 103 (3), 411-423.
Babakus, Emin, Ugur Yavas, Osman M. Karatepe, and Turgay Avci
(2003), The Effect of Management Commitment to Service
Quality on Employees Affective and Performance Outcomes,
Journal of the Academy of Marketing Science, 31 (3), 272-286.
Babin, Barry J. and James S. Boles (1998), Employee Behavior in a
Service Environment: A Model and Test of Potential Differences
between Men and Women, Journal of Marketing, 62 (2), 77-91.
Bagozzi, Richard P. (1992), The Self-Regulation of Attitudes,
Intentions, and Behavior, Social Psychology Quarterly, 55 (2),
178-204.
Baldrige National Quality Program (2005), Criteria for Performance
Excellence. Gaithersburg, MD: NIST.
Barker, Kathleen (1993), Changing Assumptions and Contingent
Solutions: The Cost and Benefits of Women Working Full- and PartTime, Sex Roles: A Journal of Research, 28 (1/2), 47-71.
Boles, James D., John A. Wood, and Julie Johnson (2003),
Interrelationships of Role Conflict, Role Ambiguity, and WorkFamily Conflict with Different Facets of Job Satisfaction and the
Moderating Effects of Gender, Journal of Personal Selling & Sales
Management, 23 (2), 99-113.
Borucki, Chester C. and Michael J. Burke (1999), An Examination of
Service-Related Antecedents to Retail Store Performance, Journal
of Organizational Behavior, 20 (6), 943-962.
Boshoff, Christo and Janine Allen (2000), The Influence of Selected
Antecedents on Frontline Staffs Perceptions of Service Recovery
Performance, International Journal of Service Industry Management, 11 (1), 63-90.
Brashear, Thomas G., Elzbieta Lepkowska-White, and Christian
Chelariu (2003), An Empirical Test of Antecedents and
Consequences of Salesperson Job Satisfaction among Polish Retail
Salespeople, Journal of Business Research, 56 (12), 971-978.
Brown, Steven P. and Robert A. Peterson (1993), Antecedents and
Consequences of Salesperson Job Satisfaction: Meta-Analysis and
Assessment of Causal Effects, Journal of Marketing Research, 30
(1), 63-77.
Burke, Michael J., Chester C. Borucki, and Amy E. Hurley (1992),
Reconceptualizing Psychological Climate in a Retail Service,
Journal of Applied Psychology, 77 (5), 717-729.
Chiu, Randy K. and Anne Marie Francesco (2003), Dispositional
Traits and Turnover Intention: Examining the Mediating Role of Job
Satisfaction and Affective Commitment, International Journal of
Manpower, 24 (3), 284-299.
Cohen, Aaron (1993), Organizational Commitment and Turnover:
A Meta-Analysis, Academy of Management Journal, 36 (5),
1140-1157.
Connelly, Catherine E. and Daniel G. Gallagher (2004), Emerging
Trends in Contingent Work Research, Journal of Management, 30
(6), 959-983.
371
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.