2011 10 3861

1 2
1

98
213,700
9550 1597

(1)
(2)

(3)
(4)

38

A Study for the Model of Graduate Students

Capability Development in Taiwan
Tzu-Chin Rejoice Chou1, Shing-Chen Wu 2
1

Assistand Professor, Department of Applied Statistics & Information Science &

Graduate School of Education, Ming Chuan University
2

Teacher, Haishan Elementary School

Abstract
The purpose of this study is to explore a model for graduate students capability
development in Taiwan. The study population are 213,700 public and private graduate
students for 98-school year in Taiwan, there are 9550 questionnaires totally by using
proportional allocation random sampling sent out, and returned 1597 valid
questionnaires; then, applied descriptive statistics, 1-Way MANOVA, Pearson
correlation, and structural equation model (SEM) statistical methods to analyze
data. Results and findings are summarized as follows:(1) The graduate students
development capability status are medium to high level, they are confident most in the
ability to develop teacher-student interactions and intellectual growth; (2) There is a
significant medium positive correlation between teaching methods and students
development capability, student-teacher interaction and students development
capability, as well as teaching methods and student-teacher interaction; (3) The
graduate students development capability are different on marital status, school
district, university concentration, professional status, work salary, daily work hours,
and children numbers; (4) The model has established, teaching method had
significant positive influence graduate students development capability in intellectual
growth and learning result, while teaching method and student-teacher interaction
are high correlation. This study provides some suggestions based on the results, and
expects to serve as reference points of future research on increasing graduate students
development capability in Taiwan.

Key words: Teaching methods, student-teacher interaction, graduate students

capability, structural equation model (SEM)

39

John Naisbitt & Patricia Aburdence Megatrends 2000

Naisbitt & Aburdence, 1991

20052005

1986
16 2011 156 25
10 2009
4.78%

2005

21

2009
40

97.14% 97.10% 7
68.19
8 18

Brophy, 1988

2003

E. N. Goplerued 1980

2000
1994
1998

41

1.
2.
3.
4.

42

1997
1999

2005

J. PiagetJ. Bruner

2005

2005
2001

43

1998

2003

2
34

2009
2009

44

Leung & Kember2006Student

Experience Questionnaire for Graduates

Intellectual

Learning Outcome

Working Together

2008

98
98213,700
1597

4
98
45

10%11%25
36001
1

23546
46.9%

10265
20.4%

14342
28.5%

2103
4.2%

50256
100%

4479
1948
2722
401
9550

202 (47%)
88 (20.5%)
123 (28.6%)
18 (4.2%)
430 (100%)

26Leung Kember
200619-23

t
213

-14.655 .000

11.151 .000

.428 .000

3.430 .001

14.553 .000

-10.463 .000

Q8.

7.880 .000

Q9.

-1.345 .000

Q10.

-5.714 .000

2.410 .016

-13.931 .000

Q1.
Q2.
Q3.
Q4.
Q5.
Q6.

Q19.
Q20.
46


Q21.

8.760 .000

Q22.

1.152 .000

Q23.

-1.611 .000

Q24.

9.121 .000

Q7.

2.255 .024

Q15.

1.362 .000

-11.284 .000

Q17.

4.504 .000

Q18.

6.971 .000

5.571 .000

-2.271 .023

4.408 .000

-7.000 .000

-3.228 .001

3.228 .001

Q16.

Q11.
Q12.
Q13.
Q14.

Q25.
Q26.

Ha

H 7a

H 2a

H 3a
H 4a

H 5a

H 6a

1
47

H a1
H a2
H a3
H a4
H a5
H a6
H a7

3 n=1597
Cronbachs .7 .9
Q24
F5
3

9
9
.888
F1
4
4
.818
F2
5
5
.794
F3
6
5
.855
F4
2
3
.854
F5
.942

48

.886
.821
.790
.826
.839
.941

KMO
.9 PAF Promax
5 41.04%
7.46%5.13%
4.21%3.39%
61.23%2007

M=3.98M=3.99
M=3.85
M=3.80M=3.74

M 3 4

3.87

r .55p<.01

r=.42p<.01
49

r=.67, p<.01

r .53p<.01

r=.43p<.01
r=.54, p<.01

r=.69p<.01
r=.59p<.01
r=.46p<.01
r=.67, p<.01

1-Way MANOVA
12

6
7
2005

9
10

50

11

structural equation modeling, SEM

1
2
34
2006

LISREL MLmaximum likelihood

3 10
Kolmogorov-Smirnov Z, K-S Z(p<.05)
4

3.91
4.01
3.99
3.90
3.94
3.83
3.78
3.64
3.82
3.72
3.93
3.92
3.98

.73
.70
.72
.73
.72
.78
.70
.74
.71
.74
.74
.75
.79

-.84
-.79
-.81
-.78
-.82
-.74
-.58
-.30
-.68
-.55
-.78
-.74
-.74

51

1.43
1.21
1.19
1.29
1.27
.97
.94
-.02
1.07
.43
.95
1.18
.77

K-S Z

.202
.224
.191
.238
.191
.187
.202
.159
.228
.195
.245
.211
.231

.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00

(offending estimate)

2 1
3
5

SEM
standardized
residuals, SRmodification index, MI
2006
2

1
2
3
4
5
6
7
1
2
3
4
5
6
7
8
9

**

.27
.49**
.07
.08
-.07
.05
.87**
.71
.75
.74
.72
.78
.78
.85
.70
.88

t
**

3.11
7.07**
1.22
1.04
-1.11
1.04
15.52**
31.38**
34.71**
33.12**
32.25**
35.62**
34.90**
38.58**
29.73**
38.68**

.085
.067
.054
.077
.065
.049
.056
.023
.022
.022
.022
.022
.022
.022
.024
.023

10
11
12
13
1
2
3
4
5
6
7
8
9
10
11
12
13

*p<.05**p<.01

52

.69
.85
.86
.77
.50
.42
.46
.48
.40
.39
.28
.51
.22
.53
.27
.26
.40

**

24.13
13.63**
14.37**
21.22**
25.13**
23.92**
24.55**
24.85**
23.51**
20.59**
15.63**
22.73**
9.85**
24.13**
13.63**
14.37**
21.22**

.022
.020
.018
.019
.020
.018
.019
.019
.017
.019
.018
.022
.022
.019
.022
.022
.024

.50

.71**
2 .42
.53

.27

.69**
.85**

.27**

.75**
.74**
.72**

.08
.78**
**

.87

.46

.48

.4
0

**

.49
.26

.40

.86**
.77

**

-.07
.05

**

.78
.85**

.39

.28

.07

*p<.05**p<.01

.70**
.88**

.51

.22

SEM

2
2 2
2
1597
6 absolute fit measures

incremental fit measureparsimonious fit measures

53

GFI
AGFI
SRMR
RMSEA

***

.9
.9
.05
.05.08
ECVI< ECVIIM

417.14
.961
.936
.0317
.064
.306
.114
18.329

NFI
NNFI
IFI
CFI
RFI

.986
.983
.988
.988
.980

.9
.9
.9
.9
.9 .95

PNFI
PGFI

.695
.581
489.138

.5
.5
AIC< SAIC

182.000
29307.600
327.321

AIC< IAIC

>200

ECVI
ECVISM
ECVIIM

AIC
SAIC
IAIC

CN

ECVI< ECVISM

ECVI =expected cross-validation Index ECVI ECVIIM =ECVI for independence

model ECVI
ECVISM =ECVI for saturated model ECVI
AIC
= Model AIC AIC
IAIC =independence AIC AIC SAIC
=saturated AIC AIC

7
t 1.96
SFL .69.85
Jreskog & Srbom (1993).45

SMC (Bollen,
54

1989)SMC 6
60.8% 50.4%
72.3% 60.8%
77.4 49.0%
72.3 47.6%
72.3 59.3%
(construct reliability, CR) Bagozzi and Yi (1988
.6

.6 .748

(convergent validity
average variance extracted , AVE.548 .666 .5

(Bagozzi and Yi, 1988)

7
SFL
SE
SMC

.71
.023
.50

.75
.022
.56

F1
.74
.022
.55

.72
.022
.52

.78
.022
.61

.78
.022
.61

F2
.85
.022
.72

.70
.024
.49

F3
.88
.023
.77

.69
.022
.48

F4
.85
.020
.72

.85
.018
.72
F5

.77
.019
.59

55

CR

AVE

31.38**
34.71**
33.12**
32.25**
35.62**

.86

.55

34.90**
38.58**

.80

.67

29.73**
38.68**

.77

.63

24.13**
13.63**

.75

.60

14.37**
21.22**

.80

.66


N=1597SFL= standardized factor loading
SE = standard error ;
SMC = squared multiple correlations for structural equations
CR = construct
reliabilityAVE =average variance extracted

3.87

r=.671p<.01r=.54p<.01
r=.67p<.01

1

56

10

H a1

H a2

H a3

H a4

H a5

H a6

H a7

57

58

59

1999

2006SEM- LISREL
2007 - SPSS
2009179 134
2003

2009

1998

1997
2001

20052009.8.3
http://epaper.edu.tw/old/highedu/plan.html
20092009 7
20 http://www.edu.tw/statistics/content.aspx?site_content_sn=8956
20056322-12
2003
2000
60

2005

2005
1-10
2008

Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models.
Academic of Marketing Science, 16, 76-94.
Bollen, K. A. (1989). Structural equations with latent variables. New York:
JohnWiley & Sons.
Brophy, J. (1988). Educating teachers about about managing classrooms and students.
Teaching and Teacher Education, 4(1), 1-18.
Jreskog, K. G. & Srbom, D. (1993). Structural equation modeling with the SIMPLIS
command language. Chicago, IL: Scientific Software International.
Naisbitt, J., Aburdence, P. (1991). Megatrends 2000. New York: William Morrow
and Company, Inc.Clark, R. E. (1984). Research on student thought process
during computer-based instruction. Journal of Instructional Development, 7, 2-5.
Leung, D.Y. P., & Kember, D. (2006). The influence of teaching approach and
teacher-student interaction on the development of graduate capabilities.
Structural equation modeling, 13(2), 264286.

61