Livable Transit Corridors: Methods, Metrics, and Strategies (2016)

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.

117 Table G-1 provides factor analysis results for 10 of the 12 metrics used in this Handbook. The two factor variables produced in this process were then used (along with the remaining 2 of 12 total metrics) as independent variables in an ordinary least squares (OLS) linear regression model to predict a proxy quality-of-life indicator variable: the corridor non-auto internal trip capture rate. Linear regression model results using the variables (factors) produced from the factor analysis model run plus the two remaining metrics—corridor pedestrian environment and corridor pedes- trian collisions per daily 100,000 pedestrians—are shown in Table G-2. Linear regression results suggest the collection of livability metrics (the independent variables) is a good predictor of transit corridor non-auto internal trip capture rates, and by inference, transit corridor quality of life. These findings helped the Handbook’s research team validate the metrics. Table G-3 provides analysis of variance (ANOVA) statistical analysis results suggesting that for the 250 outside-of-CBD corridors analyzed, the metrics used in this Handbook (including Housing Unaffordability and Income Diversity) have average values for each typology category that are significantly different from each other. These significant differences are consistent with the theoretical hypotheses posed prior to analysis and the values shown in Table 10 (see Step 3 for additional discussion of these results). These findings indicate that as a group, the 12 metrics used in this study are useful for distinguishing one typology category from another. Furthermore, since the linear regression model (see Table G-3) predicting corridor non-auto internal capture rates provided validation of the metrics, these ANOVA results also help validate the typology. A P P E N D I X G Statistical Analysis of Metrics and Typology Categories

118 Livable Transit Corridors: Methods, Metrics, and Strategies Rotated Component Matrixa Metric Factor 1 2 Transit employment accessibility 0.839 0.439 Corridor transit service coverage 0.795 0.486 Corridor housing unaffordability -0.137 -0.845 Corridor income diversity 0.680 Corridor jobs density 0.925 Corridor retail jobs density 0.897 -0.157 Corridor health care opportunities 0.705 0.522 Corridor density (population/acre) 0.754 0.463 Access to culture & arts 0.918 Ridership balance 0.332 0.101 Notes: N = 250 U.S. Transit corridors. Extraction method: Principal component analysis. Rotation method: Varimax with kaiser normalization. aRotation converged in 3 iterations. Total Variation Explained: 71.69%. Table G-1. Factor (loadings) analysis results summary table for outside-of-CBD corridors. Variable (Factor or Metric) Coefficient Significance Factor 1: Transportation/Land Use/ Livable Opportunities Integration 0.043 *** Factor 2: Housing Affordability & Income Diversity 0.164 *** Pedestrian environment (intersection density) 0.355 *** Pedestrian collisions per 100,000 pedestrians -0.256 *** Constant 0.093 ** Model Fit N1 31 R Square 0.914 Notes: 1Non-auto Internal Capture Rates developed for 31 transit corridors from California, Texas, and Florida National Household Transportation Survey Supplementary datasets. ** = p < 0.05 *** = p < 0.01 Table G-2. OLS linear regression results predicting outside-of-CBD corridor non-auto internal capture rates.

Statistical Analysis of Metrics and Typology Categories 119 Variable (Metric) Sum of Squares Degrees of Freedom Mean Square F- Statistic P-Value Transit Jobs Accessibility Between Groups 4.121E+10 2 2.061E+10 193.215 0.000 Within Groups 2.655E+10 249 106644971 Total 6.777E+10 251 Transit service coverage (aggregate frequency of transit service per square mile) Between Groups 1182198429 2 591099214 128.986 0.000 Within Groups 1136495917 248 4582645 Total 2318694345 250 Housing unaffordability (percent of income spent for housing) Between Groups 234.932 2 117.466 5.219 0.006 Within Groups 5648.862 251 22.505 Total 5883.794 253 Income diversity (variance from regional median household income) Between Groups .060 2 .030 6.161 0.002 Within Groups 1.218 251 .005 Total 1.278 253 Jobs density (employees/acre) Between Groups 15694.037 2 7847.019 165.450 0.000 Within Groups 11904.508 251 47.428 Total 27598.545 253 Retail jobs density (retail employees/acre) Between Groups 88.899 2 44.449 134.779 0.000 Within Groups 82.778 251 .330 Total 171.677 253 Transit balance of ridership flows Between Groups .350 2 .175 2.512 0.089 Within Groups 4.732 68 .070 Total 5.082 70 Health care opportunities (health care employees/ acre) Between Groups 332.513 2 166.257 80.229 0.000 Within Groups 520.143 251 2.072 Total 852.656 253 Population density (population/acre) Between Groups 23995.677 2 11997.839 191.116 0.000 Within Groups 15757.190 251 62.778 Total 39752.868 253 Access to culture and arts (corridor entertainment employees/acre) Between Groups 459.856 2 229.928 157.497 0.000 Within Groups 366.433 251 1.460 Total 826.288 253 Pedestrian environment (intersection density) Between Groups 174445.130 2 87222.565 97.357 0.000 Within Groups 224871.842 251 895.904 Total 399316.972 253 Pedestrian collisions per 100,000 pedestrians Between Groups 43.936 2 21.968 4.630 0.013 Within Groups 313.167 66 4.745 Total 357.103 68 Notes: P-Values less than 0.100 are considered statistically significant, indicating there are significant differences between the average values of each typology group for that variable (metric). Table G-3. ANOVA results comparing average metric scores for each outside-of-CBD transit corridor typology category.