Open Journal of Social Sciences, 2014, 2, 18-27
Published Online September 2014 in SciRes. http://www.scirp.org/journal/jss
http://dx.doi.org/10.4236/jss.2014.29004
How to cite this paper: Chen, J.A. (2014) Which Characters of Knowledge-Based Employees Have Higher Turnover Inten-
sion in Chinese Culture Industry? Open Journal of Social Sciences, 2, 18-27. http://dx.doi.org/10.4236/jss.2014.29004
Which Characters of Knowledge-Based
Employees Have Higher Turnover Intension
in Chinese Culture Industry?
Jing’an Che n
Economy Management Department, Southwest Jiaotong University, Chengdu, China
Received April 2014
Abstract
By use of multiple logistic regression analysis of the questionnaire data of 465 know le dge -based
staff in the culture industry of a province in the Western China, the author has found that the three
main factors related to job satisfaction and the turnover intention are not simply negatively cor-
relative, which differs from the former classical studies of relationship between job satisfaction
and turnover intention. This research has discovered that 1) the factor of career development sa-
tisfaction is not predictable on the turnover intention probability; 2) the turnover intention of the
knowledge-based staff who have low satisfaction with the job itself is not higher than those who
have high satisfaction with their jobs; 3) the turnover intention of the staff with low pay satisfac-
tion is lower than those with high pay satisfaction, that is, the knowledge-based employees with
higher pay satisfaction have correspondingly higher probability of turnover intention, compared
with those with lower pay satisfaction; 4) such covariates as gender, age, educational background,
professional title, income level, position, working years, the number of previous work units being
considered, all mentioned demographic characteristics above but the income level have no signif-
icant influence on turnover intention; the employees with low income levels are more unwilling to
leave than those with higher monthly income level.
Keywords
Job Satisfaction, Turnover Intention, Cultural Industry, Western China
1. Introduction
In recent years, the rapid development of Chinese culture industry has been accompanied by the increasing
turnover of knowledge staff in some cultural enterprises. What are the relationships between the job satisfaction
of knowledge staff of those enterprises and their turnover? Is the mode similar to that of the knowledge staff in
IT industry, colleges and research institutes? Who have the highest turnover intention among the knowledge
staff of cultural industry? All these questions need to be further studied so as to establish a precise early warning
system against the resign of knowledge staff and to prevent excessive loss of knowledge staff of culture industry

J. A. Chen
19
in western China.
The research results of most scholars at home and abroad have indicated that job satisfaction and turnover in-
tention are negatively correlated and positively correlated with their tendency of stay; some researchers have
even found that there is no significant correlation between job satisfaction and turnover (Porter et al., 1973) [1].
Existing research have shown that there are differences in demographics in terms of job satisfaction. When it
comes to the relationship between age and job satisfaction, some studies have found that age and job satisfaction
is positively correlated, that is, job satisfaction grows higher with age (Bluedorn, 1982; Martin, 1979), negative
correlated (Muchinsky et al., 1978), not significantly correlated (Ronen, 1978; Nassab, 2008). And some other
research has proved that the correlation between age and overall job satisfaction is shown as a U-shaped model
(Herzberg, Mausner, 1957), inverted U-shaped (Wang Zhigang, Jiang Huiming, 2004), and J-shaped model (Sa-
leh, 1964); as for the correlation between gender and job satisfaction, some experts believe that no simple con-
clusion can be drawn on the distinction of job satisfaction of men and women (Herzberg, 1957), while Chinese
data has shown that female employees have higher job satisfaction than males (Wang Zhigang, Jiang Huiming,
2004), and Ismael & Richard (1991) has discovered that female teachers have higher job satisfaction than male
teachers; the correlation between years of working and job satisfaction remains inconclusive. According to some
research, the longer the length of service is, the higher the employees’ job satisfaction (Martin, 1979), although
some others have proved the job satisfaction of the employees with more years of working is lower than that of
those with shorter length of service (Lin Zheng, 1999; Gibson & Klein, 1970). There is also a study which has
found that age only has significant influence on the working relationship of women (Shott, Albright & Glennon,
1963); the correlation between the degree of education and job satisfaction is uncertain, because experts have
argued that they are positively correlated (Shi Pu, 1991), negatively correlated (Blegen & Muller, 1987) and not
significantly correlated (Ding Hong, 1987).
In terms of the research objects of the knowledge staff, we mainly focus on teachers and professors in colleg-
es and universities and the employees of scientific research institutions and IT companies. On the research me-
thods, Chinese scholars used to apply correlation analysis to study the relationship between job satisfaction and
turnover intention and have currently employed structural equation model in their analysis. Correlation analysis
can only verify the correlation between job satisfaction and turnover intention, but cannot manifest the causality
between the two; in essence, structural equation model is the regression analysis of exploratory factor analysis
basis which can analyze the causal relationship or path relationship of them but cannot tell the probability of job
satisfaction affecting turnover intention. Based on the questionnaires of the knowledge employees in the cultural
industry of a province in Western China and PCA (principal component analysis), this study adopts multiple lo-
gistic regression model to analyze the degree of probability of work satisfaction affecting the turnover intention.
This study takes the knowledge staff in the cultural industry of the western region as the research objects; there-
Shi Yinlei (2007), through relevant analysis, found that all dimensions related to job satisfaction but career development were negative
ly
correlated to the turnover intention. He Guifen (2009) concluded from certain research that the job satisfaction of the employees in IT ente
r-
prises had significant negative correlation with their turnover intention and had certain predictive effect on turnover intention; the most i
m-
portant dimension affecting the turnover intention of IT employees is the work content, followed by
supervisors’ supervision, salary and
welfare, role load, role ambiguity and career development successively. Wang Dongqian (2013)
believed that the five dimensions related to
the job satisfaction of knowledge employees are all significantly negatively corr
elated to the turnover intention. Among others, leadership
management satisfaction had most negative correlation with turnover intention, while individual development satisfaction had
least negative
correlation with turnover intention. Yang Xiuwei et al. (
2005) took teachers in universities and colleges as the objects of study and pointed
out that three dimensions of job satisfaction, namely, the work itself, leadership management and interpersonal relationship, had significan
t-
ly negative influences on turnover intention. Cheng Yu (2010) proved that job satisfaction of IT employees had significantly negative corr
e-
lation with their turnover intention. Among all dimensions accounting for job satisfaction, satisfaction with the working con
ditions, leaders
behaviors and external reward had significant influences on their turnover intention; He Xiongwei
(2009) found that work payoff and work
tasks of the work satisfaction had significant influences on the turnover intention of university and college teachers. Abou
t the relationship
between age and job satisfaction, some studies have found that age is positively correlated with job satisfaction, and the hi
gher one’s age,
the higher job satisfaction he/she has (Bluedorn, 1982; Martin, 1979), negatively correlated (Muchinsky et al
., 1978), and not significantly
correlated (Ronen, 1978; Nassab, 2008). And some other research has proved that the correlation between age and overall job s
atisfaction is
shown as a U-shaped model (Herzberg, Mausner, 1957), inverted U-shaped (Wang Zhigang, Jiang Huiming, 2004), and J-
shaped model
(Saleh, 1964); as for the correlation between gender and job satisfaction, some experts believe that no simple conclusion can
be drawn on
the distinction of job satisfaction of men and women (Herzberg et al
., 1957), while Chinese data has shown that female employees have
higher job satisfaction than males (Wang Zhigang, Jiang Huiming, 2004), and Ismael & Richard (1991) has discovered that femal
e teachers
have higher job satisfaction than male teachers; the correlation between years of working and job satisfaction remains inconclusive. A
c-
cording to some research, the longer the length of service is, the higher the employees’ job satisfaction (Martin, 1979), alt
hough some others
have proved the job satisfaction of the employees with more years of working is lower than that of those with shorter length of service
(Lin
Zheng, 1999; Gibson & Klein, 1970). There is also a study which has found that age only has significant influence on the working relatio
n-
ship of women (Shott, Albright & Glennon, 1963); the correlation between the degree of education and job satisfaction is uncertain
, because
experts have argued that they are positively correlated (Shi Pu, 1991), negatively correlated (Blegen & Muller, 1987) an
d not significantly
correlated (Ding Hong, 1987).

J. A. Chen
20
fore the research conclusion will further enrich the empirical research results of the relationship between the job
satisfaction and turnover intention of Chinese knowledge staff.
2. Measuring Tools, Sampling Methods and the Demographic Variables of Samples
2.1. The Measuring Tools and Credibility
The Job Satisfaction Survey is based on modified Employee Satisfaction Questionnaire [2] of Tsui (1992) and
the Minnesota Satisfaction Questionnaire (MSQ). Likert’s Five-Scaling Method is adopted to score the subjects’
satisfaction and make evaluation. “1” represents total disagreement, “5” strong agreement, the higher the score,
the higher the job satisfaction. The Job satisfaction scale overall Cronbach a consistency coefficient is 0.778,
meeting the usually recognized requirement of internal consistency reliability above 0.6 [3].
The turnover intention falls into far from intense, not very intense, common, relatively intense, very intense,
and the assignments of the five points with Likert Scale are 1 - 5.
2.2. Sampling Method
The Questionnaire takes “Directory of the First Batch of Key Cultural Enterprises of ** Province” as a sample
frame, covering such seven industries as cultural tourism industry, publication and distribution industry, film
and television industry, entertainment and leisure industry, printing industry, animation game industry and crea-
tive design industry. About the definition of knowledge employees, in the actual survey, the data of employees
with a junior college degree and above have been collected. 1000 questionnaires have been issued and 546 of
them taken back, with a questionnaire recovery rate of 54.6%. The number of effective questionnaires is 465 and
the effective questionnaire recovery rate 46.5%.
2.3. The Demographic Variables of the Samples
The general information of the demographic statistics of the questionnaires is as follows.
(1) Gender. Male 53.2%, female 53.2%.
(2) Age. 25 years old and under constituting 3.9%, 26 - 30 years old 11.2%, 31 - 35 years old 15.1%, 36 - 40
years old14.1%, 41 to 50 years old 34.6%, 50 years old above 21.0%.
(3) Degree. The subjects with a Doctoral Degree accounting for 1.0%, Masters 15.3%, Bachelors 57.1% and
junior college graduates 26.6%.
(4) Professional title. The subjects with a senior title accounting for 23.8%, intermediate title 33.7%, primary
title 9.9%, title to be conferred 32.7%.
(5) Monthly income level. 2000 Yuan and below accounting for 2.6%, 2001 - 4000 Yuan 29.4%, 4001 - 6000
Yuan 21.1%, 6001 - 8000 Yuan 14.9%, 8001 - 10,000 Yuan 8.2%, 10001 - 15,000 Yuan 17.0%, 15,001 - 20,000
Yuan 5.7%, 20,000 Yuan above 1.0%.
(6). Position. Marketing 10.0%, Operating Management28.0%, Professional Skills 39.5%, other positions
22.5%.
(7) Years of working. 5 years and below 12.3%, 6 - 10 years 14.7%, 11 - 20 years 22.5%, 21 - 30 years 25.0%,
31 years above 25.5%.
(8) The number of previous work units. No previous work unit 37.0%, 1 (22.0%), 2 (21.5%), 3 (11.5%), 4
(3.5%), 5 and above 3.5%.
3. Factor Analysis
3.1. Factor Analysis Validity Test
KMO and Bartlett’s Test of Sphericity are used to test whether the samples are suitable for factor analysis. The
SPSS17.0 analysis results showed that the KMO measure value of the samples is 0.916, the Approximate
Chi-Square value of Bartlett’s Test of Sphericity is 2029.539, P value reaches 0.000, so the indicators show that
the samples are suitable for factor analysis (see Table 1).
3.2. Factor Extraction
According to the rule that the factor feature value is greater than the threshold value of 1, three common factors

J. A. Chen
21
have been extracted, and the cumulative resolution reached 71.402%, which shows that the overall explanatory
power of the common factors is stronger.
3.3. Factor Induction
Use Varimax to rotate the factors and determine which factor each item belongs to according to whether the
factor loading value of each item reaches 0.5 as the threshold value (see Table 2).
The research based on the factor analysis method has found that the job satisfaction of knowledge staff in the
cultural industry includes three main factors: the first main factor is related to the job itself, including reflecting
individual professional knowledge level, the opportunities giving play to their specialty, the growth opportuni-
ties, reflecting one’s value, the creativity and degree of freedom of the job itself, etc.; the second main factor is
career development, including the status of cultural industry in the regional industries, the prospects of the in-
dustry itself, the prospects of the unit, etc.; the third main factor is the pay, including the competitiveness of the
pay, income rationality and pay fairness and so on.
4. Multiple Logistic Regression Analysis
On the basis of the above factor analysis results, multiple logistic regression analysis has been made with the
work itself, career development and pay factor as explanatory variables, and gender, age, education, income, etc.
as control variables, in order to determine the probability of job satisfaction affecting turnover intention.
Table 1. KMO and Bartlett’s Test.
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.916
Bartlett’s Test of Sphericity
Approx. Chi-Square 2029.539
df 105
Sig. 0.000
Table 2. Factor loading matrix after rotation.
Factor
Item 1 2 3
You are satisfied with the work environment of your unit. 0.592 0.048 0.389
You are satisfied with the degree of freedom and creativity of your work. 0.722 0.102 0.448
The post enables you to give play to your operational capability and realize your personal value. 0.721 0.412 0.274
Your job can give play to the professional knowledge you have learned. 0.666 0.294 -0.080
You have many opportunities for personal growth. 0.735 0.238 0.351
You take pride in your work. 0.549 0.527 0.161
You plan to have long-term career development in * * Province. 0.170 0.825 0.294
You plan to have long-term career development in this unit. 0.243 0.855 0.226
You are willing to have long-term development in the field of cultural industry. 0.219 0.869 0.188
You can feel the joy from work. 0.537 0.566 0.192
You current occupation can give play to your specialty and enable you to realize your personal value. 0.610 0.530 0.223
The industry you are in is the key industry of * * Province and very promising. 0.458 0.438 0.538
You think that your income is acceptable compared with your contribution to the job. 0.170 0.255 0.840
Your pay level is higher than the average level of the industry, very competitive. 0.196 0.200 0.833
You think that the pay system of your unit is relatively fair. 0.228 0.222 0.814

J. A. Chen
22
4.1. Model Test
Establish a multiple logistic regression model with the work itself, career development and pay factor as expla-
natory variables and turnover intention as a dependent variable.
“-2LL” Likelihood Chi-Square test. From the initial model to the final model, “-2LL” value reduces by about
70, and the model has been improved significantly. The probability of the model test sig. value is .000, which
indicates that the model’s degree of fitting is good (see Table 3).
4.2. Probabilistic Forecasting of Job Satisfaction on Turnover Intention
In the multiple logistic regression model, the four groups of data of turnover intention are compared with the last
group with “very intense” turnover intention. From Table 4 “Test of the Multiple Logistic Regression Model”,
the following conclusion can be drawn.
(1) There are no significant differences in the influences of the career development factors on the five groups
of turnover intention. Compared with the group with “very intense” turnover intention, the significances of the
influence probabilities on the other four groups of turnover intention to different extents are all above 0.05. This
indicates that at the probability level of 0.05, the influences of the differences in career development satisfaction
on the turnover intention do not cause discrepancy with statistical significance. In the existing research conclu-
Table 3. Model Test.
Model −2 Log Likelihood Chi-Square df Sig.
Intercept Term 411.677
Final Model 341.901 69.776 12 .000
Table 4. Test of the multiple logistic regression model.
Turnover
Intention (a)
B Standard Error Wald Value
Degree of
Freedom
Significance
Forecasting
Probability
95% Confidence
Interval for Exp(B)
Lower
Bound
Upper
Bound
Far from Intense Intercept Term 3.453 0.755 20.912 1 0.000
The Work Itself 1.485 0.588 6.387 1 0.011 4.416 1.396 13.972
Career
Development 0.445 0.526 0.714 1 0.398 1.560 0.556 4.377
Pay 1.267 0.508 6.219 1 0.013 3.551 1.312 9.615
Not very Intense Intercept Term 2.866 0.763 14.119 1 0.000
The Work Itself 0.832 0.587 2.007 1 0.157 2.298 0.727 7.268
Career
Development 0.257 0.532 0.232 1 0.630 1.293 0.455 3.671
Pay 1.097 0.517 4.509 1 0.034 2.995 1.088 8.246
Common Intercept Term 2.378 0.775 9.406 1 0.002
The Work Itself −0.269 0.560 .232 1 0.630 0.764 0.255 2.288
Career
Development −0.450 0.521 .745 1 0.388 0.638 0.230 1.771
Pay 0.412 0.494 .698 1 0.404 1.510 0.574 3.973
Very Intense Intercept Term 1.035 0.863 1.440 1 0.230
The Work Itself 0.229 0.697 0.108 1 0.742 1.257 0.321 4.928
Career
Development
−0.175 0.623 0.079 1 0.778 0.839 0.247 2.846
Pay 0.854 0.639 1.786 1 0.181 2.349 0.671 8.221

J. A. Chen
23
sions about the relationship between the job satisfaction of knowledge staff and their turnover intention, whether
the career development factors have impact on the turnover intention have not come to a consistent conclusion:
some research has concluded that career development has no effect on turnover intention, while others think ca-
reer development has certain predictive effect on the turnover intention. Based on the samples of the knowledge
staff of the cultural industry of West China, the research shows that different Knowledge workers in the western
region culture industry as samples of research shows that different career development satisfactions has little in-
fluence on the probability of the turnover intention, that is, the turnover intention of the knowledge employees of
the cultural industry in the West region cannot be predicted by career development satisfaction.
(2) The influences of the work itself on the group with “very intense” turnover intention and that with “far
from intense” group are significantly different. Compared with the group with “very intense” turnover intention,
the significance of the influence probability of the satisfaction with the job itself on the group with “far from in-
tense” turnover intention is 0.011, which indicates that under the significance level of 0.05, the influences of the
satisfaction with the work itself on the turnover intention have statistically significant difference. According to
the model, a predictor formula can be written: the number of the employees who have lower job satisfaction and
choose “far from intense” turnover intention is 4.416 times as large as that of those with higher job satisfaction.
In other words, in the cultural industry of the western region, the turnover intention of the knowledge staff with
lower satisfaction with the work itself is not higher than that of the knowledge workers with higher satisfaction
with the work itself. Some empirical research of our country indicates that the satisfaction with the work itself of
knowledge staff and the turnover intention has significantly negative correlation. Nevertheless, the result of this
research that adopts multiple logistic regression analysis to analyze the influence of the satisfaction with the
work itself on the probability of the turnover intention is not consistent with that. On the contrary, the turnover
intention of the knowledge employees in the cultural industry of the western region who have lower satisfaction
with the work itself is lower than that of those with higher job satisfaction. The analysis of possible reasons will
be done later.
(3) Among the groups with “very intense” turnover intention, “far from intense” turnover intention and “not
very intense” turnover intention, there were significant differences in the influence of pay satisfaction. Com-
pared with the group with “very intense” turnover intention, the significances of the influence probabilities of
pay satisfaction on the groups with “far from intense” and “not very intense” turnover intention are 0.013 and
0.034 respectively. This shows that: under 0.05 significant level, compared with the group with “very intense”
turnover intention, the influences of pay satisfaction on the groups with “far from intense” turnover intention
and “not very intense” turnover intention and the influence on the group with “very intense” turnover intention
are statistically significantly different. According to the model: the number of employees with lower pay satis-
faction and “far from intense” turnover intention is 3.551 times as large as that of those with higher pay satisfac-
tion; the “not very intense” turnover intention of the employees with lower pay satisfaction is 2.995 times than
that of those with higher pay satisfaction. That is to say, the turnover intention of the employees with low pay
satisfaction is far lower than that of those with high pay satisfaction, or in other words, the knowledge em-
ployees with high pay satisfaction have correspondingly higher probability of turnover intention than those with
lower pay satisfaction. This is different from the existing research conclusion in China.
4.3. Probabilistic Forecasting with the Demographic Variables
Gender, age, educational background, professional title, monthly income level, position, years of working, the
number of previous work units being considered as the covariates, establish a multiple logistic regression analy-
sis model with the three main factors of satisfaction such as the work itself, career development and pay as ex-
planatory variables in order to predict the probabilities of the turnover intention of different extents. The estab-
lished model shows that when the above demographic variables as the covariates, only the monthly income level
variable can pass the significance testing, therefore the below model building will take monthly income level as
a covariate and job satisfaction factor as the explanatory variable, to forecast the probability of different turnover
intentions.
“−2LL” likelihood Chi-Square value from 403.096 in initial model to 321.392 of the final model, “−2LL”
value reduces by 81.704，which indicates that the model improvement is obvious. Meanwhile, the probability
sig. value of model test is 0.000 (see Table 5), which indicates that the degree of fitting of the model is good.
The model regression results are as follows:

J. A. Chen
24
In the multiple logistic regression model with monthly income level as the covariate, the group with “very in-
tense” turnover intention is a control group. That is, the four groups with different levels of turnover intention
are contrasted with the group of “very intense” turnover intention, to determine whether the turnover intention
levels of the two groups are significant different. From the model of Table 6, the following conclusion can be
drawn.
(1) The influences of the work itself on the groups with “not very intense” and “very intense” turnover inten-
tion are significantly different. Compared to the influence on the group with “very intense” turnover intention,
the probability significance of the influence of the work itself on the group with “far from intense” turnover in-
Table 5. Model test.
Model −2 Log Likelihood Chi-Square df Sig.
Intercept Term 403.096
Final Model 321.392 81.704 16 0.000
Table 6. Test of the multiple logistic regression model with monthly income level added.
Turnover
Intention(a) B Std. Error Wald df Sig. Exp(B) 95% Confidence Interval for Exp(B)
Lower Bound
Upper Bound
Far from Intense Intercept Term −3.133 2.984 1.103 1 0.294
The Work Itself 1.834 0.725 6.399 1 00.011 6.259 1.511 25.917
Career
Development 0.078 0.601 0.017 1 0.897 1.081 0.333 3.508
Pay 0.360 0.583 0.381 1 0.537 1.433 0.457 4.496
Monthly Income
Level 2.842 1.542 3.397 1 0.065 17.146 0.835 351.981
Not very Intense Intercept Term −4.781 3.004 2.532 1 0.112
The Work Itself 1.171 0.730 2.574 1 0.109 3.226 0.771 13.496
Career
Development −0.119 0.610 0.038 1 0.845 .888 0.268 2.936
Pay 0.004 0.597 0.000 1 0.994 1.004 0.312 3.236
Monthly Income
Level 3.104 1.543 4.049 1 0.044 22.283 1.084 458.088
Common Intercept Term -4.201 2.978 1.991 1 0.158
The Work Itself .114 0.707 .026 1 0.872 1.121 0.280 4.481
Career
Development −0.795 0.596 1.778 1 0.182 0.452 0.140 1.453
Pay −0.512 0.565 0.823 1 0.364 0.599 0.198 1.812
Monthly Income
Level 2.843 1.539 3.413 1 0.065 17.173 0.841 350.612
Very Intense Intercept Term −4.798 3.113 2.376 1 0.123
The Work Itself 0.652 0.813 0.643 1 0.423 1.919 0.390 9.452
Career
Development −0.484 0.686 0.497 1 0.481 0.60.7 0.161 2.365
Pay 0.054 0.700 0.006 1 0.938 1.056 0.268 4.160
Monthly Income
Level 2.639 1.560 2.860 1 0.091 13.994 0.658 297.799

J. A. Chen
25
tention is 0.011, which shows that: under 0.05 significance level, the impact probability of the work itself on the
group with “far from intense” turnover intention reaches statistical significance. The prediction formula, written
according to the results of the logistics model: the number of the employees with low satisfaction with the work
itself who choose “far from intense” turnover intention is 6.259 times than that of those with high satisfaction
with the work itself. In other words, after gender, age, educational background, professional title, monthly in-
come level, position, years of working and the number of previous work units being considered as the covariates,
none of the factors in the covariates has significantly different influences on the turnover intentions of the two
groups, i.e., the probability of the knowledge employees with low satisfaction with the work itself having “far
from intense” turnover intention is much higher than those with high satisfaction with the work itself.
(2) Compared with the group with “very intense” turnover intention, the probability of the influence of the
monthly income level on the group with “not very intense” turnover intention is significantly different, while the
influence probability of the work itself, career development and compensation on the turnover intention of this
group is not differential. The influence significance value of monthly income level on the group with “not in-
tense” turnover intention is 0.044, which shows that: under significance level of 0.05, the impact probability of
monthly income level on the group with “not very intense” turnover intention is very significant. The prediction
formula can be written: the number of employees with low monthly income level who choose “not very intense”
turnover intention is 22.283 times that of those with high monthly income level. That is to way, the intention of
the employees with low monthly income level in the cultural industry of the western region is much higher than
that of those with high income level.
(3) such covariates as gender, age, educational background, professional title, income level, position, working
years, the number of previous work units being added, the influences of the work itself, career development and
compensation satisfaction on turnover intention, compared to the group with “very intense” turnover intention,
make little difference to the groups with “common ” and “very intense” turnover intention. That is to say, after
the above covariates being added, the work itself, career development and compensation satisfaction had no
predictability on the turnover intention of the groups with “common”, “relatively intense” and “very intense”.
5. Result Analysis and Inspiration
5.1. Result Analysis
(1) This study has found that, for the knowledge-type employees in the culture industry of the western region,
career development cannot predict the turnover intention. This empirical result is different from the existing re-
search results on the correlation between the job satisfaction and turnover intention of knowledge employees. It
is probably because that the respondents of this paper are different from previous objects of study. Generally, the
objects of study who could lead to the conclusion that the career development and turnover intention are nega-
tively correlated are teachers in universities and colleges as well as staff in science and technology institutions.
With higher professional knowledge level, higher requirements on specialty accuracy, and more support from
organizational and external environment for their career development, they have stronger pursuit of career de-
velopment, compared to the knowledge employees of cultural industry. As it were, career development is the
lifeline of this group of knowledge staff. Comparatively speaking, eagerness degree of the knowledge staff in
cultural industry for career development is lower. Therefore, we should be very prudent while using satisfaction
of career development to predict the turnover intention of the knowledge staff in the culture industry of the
western areas.
(2) Another conclusion of this study is also very interesting: the turnover intention of the knowledge em-
ployees with low satisfaction with the job itself is lower that with high satisfaction with the work itself; and
those with high monthly income level have stronger turnover intention. Two reasons may explain this pheno-
menon. First, the objects of study work in the cultural industry of western region where the market development
level is relatively low. Most knowledge employees studied by existing research of China are from eastern region
or IT industry with higher marketization level. In our country, however, the reform in the fields of cultural un-
dertakings is in progress and the marketization levels of various factors in the fields are under way. The Western
China, the degree of development of the human capital market of relatively lagging behind, has fewer available
work opportunities and lower human capital flow smoothness, which causes that some employees have low sa-
tisfaction with the work itself and compensation but are unwilling to quit. Second, the high income group with
abundant human capital in the cultural industry of western region has the ability to seek the return of value

J. A. Chen
26
matching that; once the external environment is suitable, they can get better job opportunities, so they tend to
have higher turnover intention. In recent years, the booming cultural industry in East China has provided more
and better job opportunities, which has contributed stronger turnover intention of the knowledge staff with high-
er income in the cultural industry of West China. This requires further in-depth discussion in future research;
besides job satisfaction, the demission behaviors need to be discussed from such variables as external job op-
portunities.
(3) In the probability model of job satisfaction affecting turnover intention, after the 8 covariates being added,
all statistical characteristics but monthly income level do not have significant influences on turnover intention.
This conclusion is also different from some existing analysis results. The value of this research conclusion lies
in that it does not simply discuss the differences of a certain characteristic in terms of work satisfaction, but
combine various characteristics with job satisfaction to build an integrated model, improving the synthetic pre-
diction accuracy of various characteristics and able to “target” at the group with stronger turnover intention
more accurately. If we studied the differences of one of the variables with job satisfaction Under the probability
level of 95%, the synthetic judgment accuracy of the 8 covariates is only 0.66 (0.95 × 0.95 × 0.95 × 0.95 × 0.95
× 0.95 × 0.95 × 0.95 = 0.66), while the overall judgment accuracy of the model we built on the 8 characteristics
as covariates is 95%. It may also be the reason why this research conclusion is different from some existing
analysis results. Moreover, this analysis conclusion is more valuable in practical operation. In practice, it is ne-
cessary to accurately adopt the characteristics provided by the analysis conclusion to judge the probability of the
turnover intention; if the analysis conclusion based on a singular factor is used to an individual with different
characteristics in one, it is difficult to reach judgment with high accuracy. The comprehensive model this paper
has built can accurately point out that income level is an important indicator to judge the probability of turnover
intention, while other characteristics are not significant.
5.2. Inspiration
(1) With the knowledge workers in the culture industry of Western China as the objects of study, the empiri-
cal research conclusion on the relationship between job satisfaction and turnover intention is not exactly the
same as the existing analysis conclusions. This suggests the complexity of the correlation between the job satis-
faction and turnover intention of knowledge-based employees, which can’t be simply predicted with the relation
model of between work satisfaction and turnover intention of the knowledge employees from other countries,
industries or regions. Instead, the characteristics of knowledge employees in cultural industry different from
those of employees in universities, colleges, research institution and IT industry must be considered and further
researched, so as to improve the accuracy of predicting the turnover intention with work satisfaction in practical
operation.
(2) The turnover intention of the knowledge employees with low satisfaction with the work itself is not higher
than that of the knowledge workers with high satisfaction with the work itself, which is against the existing con-
clusion in the academic field. This reminds us that there are some “zombie” knowledge employees in the cultur-
al industry of western region, and their low work satisfaction will adversely affect the creativity at work. This
group of knowledge employees is not the focus of turnover intention forewarning management, but the main in-
centive objects.
(3) In current practice, the preferential tool of preventing the drain of knowledge staff in the cultural industry
of western region is high salary, and turnover forecasting management of the group with higher pay is neglected.
Our research has found that the knowledge employees with high pay satisfaction and income level have higher
probability of turnover intention than those with low pay satisfaction and income level. This indicates that the
latter group should be the main objects of turnover forecasting management in the cultural industry of western
region. Meanwhile, tools other than “high pay” should be discovered to effectively forecast the turnover inten-
tion of the group with high income level and pay satisfaction.
References
[1] Porter, L.W. and Steers, R.M. (1973) Organational Work and Personal Factors in Turnover and Absenteeism. Psycho-
logical Bulletin, 80, 151-176. http://dx.doi.org/10.1037/h0034829
[2] Tsui, A.S., Egan, T.D. and O’Reilly III, C.A. (1992) Being Different: Relational Demography and Organizational At-
tachment. Administrative Science Quarterly, 37, 549-579. http://dx.doi.org/10.2307/2393472

J. A. Chen
27
[3] Peterson, R. (1994) A Me ta-Analysis of Cronbach’s Coefficient Alpha. Journal of Consumer Research, 21, 381-391.
http://dx.doi.org/10.1086/209405