Transportation Systems Analysis
C&EE185/285-1 Winter 2026
Assignment 1 Solutions
Q1:
Several factors should be considered to choose the best model, including the coefficients' magnitudes,
their associated t-ratios, predictor variables’ correlation, and the goodness of fit (R-squared value) of
each model.
Model 1:
Pros:
o High R-squared value (0.9) indicates a good fit.
o Both coefficients have high t-ratios, indicating statistical significance.
Cons:
o Only one explanatory variable (total employment) is considered, which may not fully
capture the variation in work trips across different types of employment.
o The intercept is relatively high compared to the coefficient of the explanatory variable. If
the intercept is much larger than the coefficient of the explanatory variable, it might
suggest that the model predicts a substantial number of work trips even when there is
no employment in the zone (which may not make intuitive sense).
Model 2:
Pros:
o Higher R-squared value (0.925) compared to Model 1, indicating a better fit.
o Three explanatory variables are included, which might capture more of the variation in
work trips.
o All coefficients have moderately high t-ratios, indicating statistical significance.
Cons:
o One of the coefficients (X4) has a t-ratio just below the common threshold of 1.96,
suggesting it might not be statistically significant.
Model 3:
Pros:
o Very high R-squared value (0.996) indicates an excellent fit.
o Both explanatory variables have extremely high t-ratios, indicating strong statistical
significance.
Cons:
o The intercept is negative, which might not make sense in the context of the problem.
o Only two explanatory variables are considered, potentially overlooking other factors
influencing work trips.
C&EE185/285-1 Winter 2026
Assignment 1 Solutions
Q1:
Several factors should be considered to choose the best model, including the coefficients' magnitudes,
their associated t-ratios, predictor variables’ correlation, and the goodness of fit (R-squared value) of
each model.
Model 1:
Pros:
o High R-squared value (0.9) indicates a good fit.
o Both coefficients have high t-ratios, indicating statistical significance.
Cons:
o Only one explanatory variable (total employment) is considered, which may not fully
capture the variation in work trips across different types of employment.
o The intercept is relatively high compared to the coefficient of the explanatory variable. If
the intercept is much larger than the coefficient of the explanatory variable, it might
suggest that the model predicts a substantial number of work trips even when there is
no employment in the zone (which may not make intuitive sense).
Model 2:
Pros:
o Higher R-squared value (0.925) compared to Model 1, indicating a better fit.
o Three explanatory variables are included, which might capture more of the variation in
work trips.
o All coefficients have moderately high t-ratios, indicating statistical significance.
Cons:
o One of the coefficients (X4) has a t-ratio just below the common threshold of 1.96,
suggesting it might not be statistically significant.
Model 3:
Pros:
o Very high R-squared value (0.996) indicates an excellent fit.
o Both explanatory variables have extremely high t-ratios, indicating strong statistical
significance.
Cons:
o The intercept is negative, which might not make sense in the context of the problem.
o Only two explanatory variables are considered, potentially overlooking other factors
influencing work trips.
