Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
cating a stronger negative effect on travel decisions for WH trips during the congested periods. ⢠Shift variablesâTwo kinds of shift variables are computed, namely, shift early (SE) and shift later (SL), which measure the difference between the time period indicator (on a scale from 1 through 24 with 0.5 incre- ments) and the midpoint of the first three time periods (a.m., midday, and p.m. peak periods). SE is used when the time period indicator is less than the midpoint whereas SL is used when it is greater. The square of these variables is also used in the models to see the impact of very short and very long delays on temporal choice behavior. During model estimation, these shift variables are multiplied by the delay variable as well as other vari- ables to see the combined effect on time-of-day choice. The coefficients for the delay variables multiplied by SE and SL are significant and positive, while these are nega- tive when multiplied by the square of SE and SL. This indicates that travelers are more likely to switch their time choice when undertaking trips that may gen erate either very short or very long delays. The model statistics demonstrate that the rho-squared with respect to 0 is reasonable (0.191 for WH and 0.188 for HW), but the rho-squared with respect to the con- stants (0.003 for WH and 0.014 HW) shows that the con- stants account for nearly all the variation in time-of-day choices. While it may be desirable for the variables in the models to account for more of the time-of-day choices, the primary objective of the model is to provide sensitivity to trip characteristics, which is achieved by these models. Additional steps are carried out for the time-of-day models: ⢠The models were estimated using the full set of variables listed above, with additional testing for the best specification of the shift variables. The estimation results by trip pur pose are shown in Tables 2, 3, and 4. 146 INNOVATIONS IN TRAVEL DEMAND MODELING, VOLUME 2 TABLE 2 Home-Based Work Time-of-Day Choice Model Home to Work Work to Home Observations 6,931 6,076 Final log L â19,500.8 â17,032.7 Rho-sq. (0) 0.188 0.191 Rho-sq. (constant) 0.014 0.003 Alternatives Variable Definition Coefficient tâStat Coefficient tâStat AM1âAM10 AM Delay max(0, AM GC â NI GC) â0.06172 â4.7 â0.4277 â3.2 MD1âMD10 MD Delay max(0, MD GC â NI GC) â0.2834 â6.0 â0.3935 â9.9 PM1âPM10 PM Delay max(0, PM GC â NI GC) â0.1747 â4.2 â0.100 constant EV EV Delay max(0, EV GC â NI GC) â0.1714 â5.7 AM1âAM5 AM Shift Early AM Delay x (7.5ât) 0.1121 7.7 AM1âAM5 AM Shift Early2 AM Delay x (7.5ât)2 â0.01914 â3.4 AM6âAM10 AM Shift Later AM Delay x (tâ7.5) 0.01842 2.3 AM6âAM10 AM Shift Later2 AM Delay x (tâ7.5)2 MD1âMD5 MD Shift Early MD Delay x (12.5ât) 0.1063 4.4 MD1âMD5 MD Shift Later MD Delay x (tâ12.5) 0.09548 2.8 0.1144 4.6 PM1âPM5 PM Shift Early PM Delay x (17.0ât) 0.0766 2.8 0.09523 7.0 PM1âPM5 PM Shift Early2 PM Delay x (17.0ât)2 0 â0.03593 â4.8 PM6âPM10 PM Shift Later PM Delay x (tâ17.0) 0.05933 1.8 0.1056 9.4 PM6âPM10 PM Shift Later2 PM Delay x (tâ17.0)2 0 â0.03027 â6.0 AM1âAM10 AM HH size min(HH size,4) â0.3419 â7.6 AM1âAM10 AM Low Income HH income <$45K â0.5176 â5.9 AM1âAM10 AM High Income HH income >$75K 0.515 4.3 AM1âAM10 AM Crossing dummy(Bridge_N > 0) 0.3545 2.3 MD1âMD10 MD HH size min(HH size,4) â0.3427 â6.6 MD1âMD10 MD High Income HH income>$75K 0.461 3.4 0.6694 3.7 MD1âMD10 MD Shared ride dummy(car occ.>1) 0.479 3.8 â0.5917 â3.5 MD1âMD10 MD Crossing dummy(Bridge_N > 0) 0.618 2.6 PM1âPM10 PM HH size min(HH size,4) â0.05966 â2.1 PM1âPM10 PM High Income HH income >$75K 0.9454 5.6 PM1âPM10 PM Shared ride dummy(car occ.>1) 0.6686 4.2 â0.5694 â4.6 PM1âPM10 PM Crossing dummy(Bridge_N > 0) 0.6383 3.7 EV EV High Income HH income >$75K 0.5285 2.7 AM1âAM5 AM HS Shift Early AM HH Size x (7.5ât) 0.0722 4.0 AM1âAM5 AM HI Shift Early AM Low Inc x (7.5ât) 0.1194 2.3 AM1âAM5 AM HI Shift Early AM High Inc x (7.5ât) â0.1216 â2.8 AM1âAM5 AM BR Shift Early AM Crossing x (7.5ât) â0.4295 â5.8 AM6âAM10 AM LI Shift Late AM Low Inc x (tâ7.5) 0.2483 3.8 PM1âPM5 PM HI Shift Early PM High Inc x (17.0ât) â0.2217 â4.1