Table 3.
GDP per capita | Diff | GDP per capita × Diff | Cum | GDP per capita × Cum | Cum_diff | AIC | BIC | |
---|---|---|---|---|---|---|---|---|
A | 304.8 | 321.4 | ||||||
B | 148.9 | 168.3 | ||||||
C | 0.07 | 136.9 | 159.1 | |||||
D | 0.09 | 150 | 172.1 | |||||
E | 0.074 | 0.004 | 138.9 | 163.8 | ||||
F | 0.025 | 148.7 | 170.9 | |||||
G | 0.065 | 0.009 | 138.6 | 163.5 | ||||
H | 0.042 | 148.4 | 170.6 | |||||
I | 0.068 | 0.013 | 138.5 | 163.4 | ||||
J | 0.029 | 145.9 | 165.3 | |||||
K | 0.086 | −0.087 | −0.067 | −0.237 | −0.1 | 0.184 | 138.8 | 172 |
‘Diff’ stands for horizontal diffusion (the sum of productions of neighbouring countries divided by the squares distance between capital cities). ‘Cum’ stands for the cumulative estimated scientific production of a country before that time – i.e. vertical diffusion. ‘Cum_diff’ stands for the cumulative horizontal diffusion. The models A and B are excluded, as they use no covariate. The coefficients of model K are difficult to interpret, some covariates like horizontal diffusion having a negative effect. This is due to high covariance between covariates, especially when interactions are included.