Table 3.3.
RPBPHM model estimation results and marginal effects for public transport travel intentions.
| Variable category |
Bus Coefficient |
Bus t-stata |
Train Coefficient |
Train t-stata |
Marginal effects on the likelihood of reducing bus use [y = 1] |
Marginal effects on the likelihood of reducing train use [y = 1] |
|---|---|---|---|---|---|---|
| Constant | −0.702 | 6.60 | −0.689 | −7.05 | – | – |
| Car use prior to COVID-19 | 0.365 | 3.20 | 0.392 | 3.38 | 0.0640 | 0.0817 |
| COVID-19 bus risk perceptions | 0.195 | 2.59 | – | – | 0.0312 | – |
| COVID-19 train risk perceptions | – | – | 0.225 | 3.01 | – | 0.0449 |
| Household size | – | – | −0.196 | −2.40 | – | −0.0422 |
| Random Parameters | ||||||
| Age | – | – | −0.213 | −1.60 | – | −0.0502 |
| Standard deviation of parameter density function | – | – | 0.315 | 2.62 | – | – |
| Unable to work | −0.580 | −1.96 | – | – | −0.1364 | – |
| Standard deviation of parameter density function | 0.502 | 2.06 | – | – | – | – |
| Resides in Lothian | −0.225 | −2.29 | – | – | −0.0413 | – |
| Standard deviation of parameter density function | 0.302 | 3.41 | – | – | – | – |
| PT use prior to COVID-19 | 0.148 | 1.46 | – | – | 0.0190 | – |
| Standard deviation of parameter density function | 0.447 | 5.21 | – | – | – | – |
| Heterogeneity in the mean of RPthe term preceding “:” is the random parameter and the succeeding term is the exogenous influence. | ||||||
| Resides in Lothian (Lothian) : Highest Education level (Higher, Non-degree tertiary level) | 0.387 | 2.22 | – | – | – | – |
| PT use prior to COVID-19 (>= 1 day per week) : Social media COVID-19 information (frequent) | 0.277 | 2.01 | – | – | – | – |
| PT use prior to COVID-19 (>= 1 day per week): Resides in Central Belt (Lothian/Greater Glasgow/Clyde) | −0.364 | 2.42 | – | – | – | – |
| Unable to work (long-term illness or disabled) : Websites or online news pages COVID-19 information (frequent) | 0.820 | 1.75 | – | – | – | – |
| Cross-equation correlation coefficient (t-stat in parentheses) | 0.988(180.50) | |||||
| Number of observations | 615 | |||||
| Halton draws | 500 | |||||
| LL(0) | −762.12 | |||||
| LL(βFP), fixed parameters bivariate probit model (FPBP) | −561.84 | |||||
| LL(βRP), random parameters bivariate probit model (RPBP) | −555.48 | |||||
| LL(βRPHM), random parameters model with heterogeneity in the means of random parameters (RPBPHM) | −548.65 | |||||
PT = public transport; RP = Random Parameter; -=not applicable; LL (0) = log-likelihood at zero; LL(β) = log-likelihood at convergence.
Likelihood Ratio Test (I): RPBP > FPBP with > 95% level of confidence.
Likelihood Ratio Test (II): RPBPHM > RPBP with > 99% level of confidence.
a
Variables with |t-stat| > 1.65 are significant at > 90% level of confidence, those with and |t-stat| > 1.96 are significant at > 95% level of confidence.