Table 2 Parameters and statistics of linear candidate models

From: Rural unemployment pushes migrants to urban areas in Jiangsu Province, China

Parameters and statistics

Linear models with different candidate driving factors

 

1

2

3

4

5

6

 

\(\mu _{RU}\sim U_R^\ell - U_U^\ell\)

\(\mu _{RU}\sim U_R^\ell\)

\(\mu _{RU}\sim 1 + V_R\)

\(\mu _{RU}\sim 1 + \eta ^\ell\)

\(\mu _{RU}\sim 1 + U_U^\ell\)

\(\mu _{RU}\sim 1 + (V_U - V_R) + \Delta V_R\)

β0 p-value)

–

–

41.31 (<0.001)

−11.251 (0.608)

32.95 (0.020)

39.86 (<0.001)

β1 (p-value)

159.76 (<0.001)

144.97 (<0.001)

−0.0010 (0.557)

82.604 (0.030)

169.01 (0.722)

−0.0012 (0.439)

β2(p-value)

–

–

–

–

–

0.0064 (0.221)

N

29

29

29

29

29

25

R

0.65

0.62

0.11

0.40

0.07

0.27

p-value regression

<0.001

<0.001

0.57

0.03

0.72

0.44

LL

−123.80

−123.49

–

−126.56

–

–

BIC

250.97

250.35

–

259.86

–

–

  1. \(\eta ^\ell\) is the relative gradient in expected income shown on the right-hand side of Eq. (11b). VU and VR, as introduced in “Unemployment-driven net migration”, are expected income in urban and rural areas with unemployment rate considered; ΔVR is a measure of volatility in past income, which is calculated as \(\Delta V_R\left( t \right) = \left| {V_R\left( t \right) - V_R\left( {t - 2} \right)} \right| + \left| {V_R\left( {t - 2} \right) - V_R\left( {t - 4} \right)} \right|\). “LL” refers to the log likelihood of each linear model, and “BIC” refers to the Bayesian information criterion. BIC amongst statistically significant models at 95% (models 1, 2, and 4) suggests that models 1 and 2 are best interpreters of rural to urban migration in Jiangsu province
Back to article page