(2) The average variable cost per student from RTVUs at the prefecture and city level
The findings of the regression model show that the regression coefficient for number of enrolled students is 0.046015 (national), 0.029315 (Fujian), 0.079082 (Hunan), and 0.137426 (Xinjiang), which indicates that the average variable cost per student for RTVUs at the prefecture and city level is 460 RMB (national), 293 RMB (Fujian), 791 RMB (Hunan), and 1374 RMB (Xinjiang). (See Table 3) When compared with Table 1, we notice a large gap between the average variable cost per student and the average cost per student. It shows that there is still room for RTVUs at the prefecture and city level to reduce the average student cost.

Table 3 shows that the average variable cost per student varies according to the economic development level of the region. The basic trend is that the variable cost per student in the east is lower than in the central part, and highest in the west.

Table 4 is the results of regression on the variable cost per student in RTVUs at the prefecture and city level after adjusting the data by coefficient. The results are similar to that of Table 3. The average variable cost per student is as follows: National 345 RMB, Fujian 220 RMB, Hunan 593 RMB, and Xinjiang 1031 RMB.

Table 3. Regression result for average variable cost per student for RTVUs at the
prefecture and city level

 

Regression coefficient (t- Statistic)

Variables

National

Fujian

Hunan

Xinjiang

Intercept

346.3628(4.544102)***

265.8287(3.892756)***

468.7471(2.551921)*

239.6621(2.387278)*

Number of enrolled students

0.046015(3.936580)***

0.029315(3.147089)**

0.079082(2.723158)*

0.137426(2.438192)*

R-squared

0.158196

0.209822

0.254204

0.792192

Adjusted
R-squared

0.147930

0.189028

0.220304

0.717974

F-statistic

15.40989

10.09043

7.498671

10.67396

Prob (F-statistic)

0.000179

0.002956

0.011997

0.000221

Hausman Test

Chi-Sq. Statistic=1.461752
Prob.= 0.2267

Chi-Sq. Statistic=0.298538
Prob.= 0.5848.

Chi-Sq. Statistic=0.756249
Prob.= 0.385

Chi-Sq. Statistic=5.366253
Prob.= 0.0205

Method

random

random

Random

fixed

Notes: a. * significance< .05; **significance< .01; ***significance< .001
b. unit cost in the regression is 10,000 RMB
c. original cost data
d. Unchanged price= 100×current year cost /local CPI

Table 4. Regression result for average variable cost per student for RTVUs
at the prefecture and city level

Variables

Regression coefficient (t- Statistic)

National

Fujian

Hunan

Xinjiang

Intercept

259.7721(4.544102)
***

199.3715(3.892756)
***

351.5603(2.551921)
*

179.7466(2.38278)
*

Number of enrolled students

0.034512(3.936580) ***

0.021986(3.147089) **

0.059311(2.723158)
*

0.103069(2.438192)
*

R-squared

0.158196

0.209822

0.254204

0.792192

Adjusted R-squared

0.147930

0.189028

0.220304

0.717974

F-statistic

15.40989

10.09043

7.498671

10.67396

Prob (F-statistic)

0.000179

0.002956

0.011997

0.000221

Hausman Test

Chi-Sq. Statistic=1.461752
Prob.= 0.2267

Chi-Sq. Statistic=0.298538
Prob.= 0.5848.

Chi-Sq. Statistic=0.756249
Prob.= 0.385

Chi-Sq. Statistic=5.366253
Prob.= 0.0205

Method

random

random

random

fixed

Note: a. * significance< .05; **significance< .01; ***significance< .001
b. unit cost in the regression is 10,000 RMB
c. Cost data adjusted by coefficient (0.75)
d. Unchanged price= 100×current year cost /local CPI

(3) The analysis on the scale economy for RTVUs at the prefecture and city level and the optimal scale in student number
When other conditions have been controlled, there is scale economy for ODL in RTVUs at the prefecture and city level, i.e., with the increasing number of students, the average cost per student declines. The conclusion applies to both national and provincial samples. See Table 5 (original data) and Table 6 (original data adjusted by coefficient).

According to the regression model results, the number of enrolled students yielding the lowest average cost per student has been further computed using the U-shaped curve. The results indicate:

  1. Original data: From a nationwide perspective, for RTVUs at the prefecture and city level, the average cost per student is lowest when the number of enrolled students is 7837; when the number of students increases past 7837, the average cost per student goes up. That is to say, the optimal scale in student number for RTVUs at the prefecture and city level is 7837. From the analysis results of the provincial samples, the optimal scale in student number for Fujian, Hunan and Xinjiang are: 9129, 3855, and 2212 respectively. When compared with Table 1 showing the average scale in enrolled students, there is considerable room for increasing the enrollment of students in RTVUs at the prefecture and city level in the east, while in the central and western regions, there is also room for expansion, but to a smaller degree than in the east.

    We notice from Table 5, that from a nationwide perspective, the ratio between student and teachers has a significant inverse effect on the average cost per student, i.e., if the student to teacher ratio increases, the average cost per student decreases. The tendency reflects the teaching mode in RTVUs at the prefecture and city level—primarily media-based self-access learning. However, for the provincial samples, the impact of the student-teacher ratio on average cost per student is not significant. This phenomenon may be due to the sample set of provincial samples being relatively small.
     
  2. Original cost data adjusted by coefficient: The results of adjusting the cost by coefficient demonstrate that the nationwide optimal scale in enrolled students in RTVUs at the prefecture and city level is 7833. The analysis results of the provincial samples show that the optimal scale in enrolled students for Fujian is 9129, 3852 for Hunan, and 2212 for Xinjiang, The above results are very close to the results for the original data.

Table 5. Estimated result for the analysis on the scale economy of RTVUs at the
prefecture and city level

Variable

U-Shaped (t-statistics)

Nationwide

Fujian

Hunan

Xinjiang

Number of enrolled students

-0.779092
(-4.178833)
***

-0.555048
(-5.445197)
***

-4.479582
(-4.201364)
***

-9.394680
(-5.920911)
***

Number of enrolled
students squared

4.97E-05
(3.240568)
**

3.04E-05
(3.667262)
***

0.000581
(3.390468)
**

0.002123
(4.898520)
***

Ratio between
students and teachers

-14.99015
(-2.128598)*

-8.249869
(-2.001070)

2.612661
(0.125987)

-5.613314
(-0.188096)

Dummy variable for
East China

-519.8220(-1.188961)

Dummy variable for
Central China

897.0043(1.923173)

Intercept

4553.341
(11.26444)***

3370.077
(13.97083)***

9517.423
(7.301773)***

12299.17
(10.85228)***

R-squared

0.493311

0.726069

0.855525

0.834257

Adjusted R-squared

0.460831

0.703241

0.778472

0.803180

F-statistic

15.18813

31.80664

11.10302

26.84503

Prob (F-statistic)

0.000000

0.000000

0.000047

0.000002

Method

random

random

fixed

Random

Notes: a. * Significance < .05; **significance< .01; ***significance< .001
b. Method: determined by Hausman Test Prob.
c. original cost data
d. Unchanged price = 100*current year cost /local CPI

Table 6. Estimated result for the analysis on the scale economy of RTVUs at the
prefecture and city level

Variable

U-shaped regression coefficient (t-statistics)

Nationwide

Fujian

Hunan

Xinjiang

Number of enrolled students

-0.584319(-4.178833)
***

-0.416286(-5.445197)
***

-3.359686
(-4.201364)***

-7.046010
(-5.920911)***

Number of enrolled
students squared

3.73E-05(3.240568) **

2.28E-05(3.667262)***

0.000436
(3.390468)**

0.001592
(4.898520)***

Ratio between
students and teachers

-11.24261(-2.128598)*

-6.187402(-2.001070)

1.959495
(0.125987)

-4.209985
(-0.188096)

Dummy variable for
East China

-389.8665(-1.188961)

Dummy variable for
Central China

672.7532(1.923173)

Intercept

3415.006(11.26444)

2527.557(13.97083)***

7138.068
(7.301773)***

9224.379
(10.85288)***

R-squared

0.493311

0.726069

0.855525

0.834275

Adjusted R-squared

0.460831

0.703241

0.778472

0.803180

F-statistic

15.18813

31.80664

11.10302

26.84503

Prob (F-statistic)

0.000000

0.000000

0.000047

0.000002

Method

random

random

fixed

random

Notes: a. * Significance < .05; **significance< .01; ***significance< .001
b. Method: determined by Hausman Test Prob.
c. original cost data adjusted by coefficient (coefficient=0.75)
d. Unchanged price = 100*current year cost /local CPI