Exploratory plots

Total crashes

For the control variables, takeaways are:

  • crashes are comparatively rare in C2, high in C4 and C5
  • crashes appear to occur more frequently at intermediate speeds
  • a strong positive relationship between VMT and crashes

For the equity variables, takeaways are:

  • a generally negative relationship between income and crashes
  • an apparent negative relationship between education and crashes, though a less strong relationship than that of income
  • a generally negative relationship between average vehicles and crashes
  • a generally positive relationship between zero-vehicle density and crashes

Severe crashes

For the control variables, takeaways are:

  • minimal discernible impact of context class, and any observed impacts are inconsistent
  • crashes appear to occur more frequently at intermediate speeds, but less notably so than for total crashes
  • a strong positive relationship between VMT and crashes

For the equity variables, takeaways are:

  • a generally negative relationship between income and crashes
  • a generally negative relationship between education and crashes
  • a generally negative relationship between average vehicles and crashes
  • a generally positive relationship between zero-vehicle density and crashes

Pedestrian/bike crashes

For the control variables, takeaways are:

  • crashes appear most common in light-urban contexts (C3C, C4)
  • crashes appear to occur more frequently at intermediate speeds
  • a strong positive relationship between VMT and crashes

For the equity variables, takeaways are:

  • a generally negative relationship between income and crashes
  • an apparent negative relationship between education and crashes, though a less strong relationship than that of income
  • a generally negative relationship between average vehicles and crashes
  • a generally positive relationship between zero-vehicle density and crashes

Modeling

Total crashes

The total crashes model includes all levels of both context class and posted speed. The base levels are set to C1 and 20mph for context class and posted speed, respectively.

The equity coefficients are interpreted as follows:

Holding all other variables constant…

  • Median household income has a significant impact on 5-year crash totals. For every $1 increase in median income, we would expect the 5-year crash total to drop by a factor of 0.9999891. For reference, this equates to a ~10% reduction in 5-year crash totals for every ~$9,500 increase in median income.
  • Education has a marginal impact on 5-year crash totals. For every +1 increase in average education level, we would expect the 5-year crash total to rise by a factor of 1.295375.
  • Average vehicles has an insignificant impact on 5-year crash totals.
  • Zero-vehicle density has a significant impact on 5-year crash totals. For every +1 increase in zero-vehicle density, we would expect the 5-year crash total to rise by a factor of 1.000892. For reference, this equates to a ~10% increase for every ~105 increase in zero-vehicle density.
Variable n Estimate Lower bound Upper bound p-value sig
Median household income -1.093e-05 -1.656e-05 -5.302e-06 <0.001 ***
Average level of education 0.259 -0.007 0.524 0.056 \(\cdot\)
Average vehicles per household 0.214 -0.127 0.554 0.219
Density of zero-vehicle households 8.913e-04 3.414e-04 1.441e-03 0.001 **
(Intercept) 0.17 -1.946 2.285 0.875
VMT 4.753e-05 4.315e-05 5.191e-05 <0.001 ***
Context class = C2 6 -0.865 -1.747 0.017 0.055 \(\cdot\)
Context class = C3R 110 -0.082 -0.536 0.372 0.724
Context class = C3C 112 0.398 -0.056 0.852 0.086 \(\cdot\)
Context class = C4 386 0.591 0.152 1.029 0.008 **
Context class = C5 19 1.013 0.43 1.596 <0.001 ***
Context class = C6 10 0.929 0.235 1.622 0.009 **
Context class = SD 4 -1.527 -2.568 -0.486 0.004 **
Posted speed = 25 26 1.93 -0.089 3.949 0.061 \(\cdot\)
Posted speed = 30 61 2.48 0.479 4.481 0.015 *
Posted speed = 35 173 2.782 0.789 4.776 0.006 **
Posted speed = 40 101 2.985 0.988 4.982 0.003 **
Posted speed = 45 253 2.808 0.813 4.802 0.006 **
Posted speed = 50 20 2.937 0.911 4.964 0.005 **
Posted speed = 55 19 2.01 -0.025 4.045 0.053 \(\cdot\)
Posted speed = 60 7 2.052 -0.065 4.17 0.058 \(\cdot\)
Posted speed = 65 6 0.786 -1.383 2.956 0.477

Severe crashes

The severe crashes model excludes the Special District context class level of the context class factor and the 20mph level of the posted speed factor due to insufficient data. The base levels are set to C1 and 25mph for context class and posted speed, respectively.

The equity coefficients are interpreted as follows:

Holding all other variables constant…

  • Median household income has a marginal impact on 5-year severe crash totals. For every $1 increase in median income, we would expect the 5-year severe crash total to drop by a factor of 0.9999941. For reference, this equates to a ~10% reduction in 5-year severe crash totals for every ~$17,500 increase in median income.
  • Education has no impact on 5-year severe crash totals.
  • Average vehicles has no impact on 5-year severe crash totals.
  • Zero-vehicle density has a significant impact on 5-year severe crash totals. For every +1 increase in zero-vehicle density, we would expect the 5-year severe crash total to rise by a factor of 1.001088. For reference, this equates to a ~10% increase for every ~90 increase in zero-vehicle density.
Variable n Estimate Lower bound Upper bound p-value sig
Median household income -5.930e-06 -1.258e-05 7.239e-07 0.081 \(\cdot\)
Average level of education 0.058 -0.239 0.355 0.703
Average vehicles per household 0.071 -0.318 0.459 0.722
Density of zero-vehicle households 0.001 0.001 0.002 <0.001 ***
(Intercept) -1.244 -2.272 -0.215 0.018 *
VMT 4.071e-05 3.644e-05 4.498e-05 <0.001 ***
Context class = C2 6 -1.068 -3.196 1.061 0.326
Context class = C3R 110 0.552 -0.114 1.217 0.105
Context class = C3C 112 0.585 -0.077 1.246 0.083 \(\cdot\)
Context class = C4 386 0.973 0.327 1.62 0.003 **
Context class = C5 19 0.77 -0.008 1.548 0.052 \(\cdot\)
Context class = C6 10 1.098 0.215 1.982 0.015 *
Posted speed = 30 61 0.234 -0.34 0.808 0.424
Posted speed = 35 173 0.591 0.067 1.115 0.027 *
Posted speed = 40 101 0.899 0.358 1.439 0.001 **
Posted speed = 45 253 0.757 0.226 1.288 0.005 **
Posted speed = 50 20 0.093 -0.672 0.858 0.812
Posted speed = 55 19 0.392 -0.427 1.211 0.348
Posted speed = 60 7 0.56 -0.605 1.725 0.346
Posted speed = 65 6 0.478 -0.639 1.595 0.402

Pedestrian/bike crashes

The severe crashes model excludes the 20mph level of the posted speed factor due to insufficient data. The base levels are set to C1 and 25mph for context class and posted speed, respectively.

The equity coefficients are interpreted as follows:

  • Median household income has a significant impact on 5-year ped/bike crash totals. For every $1 increase in median income, we would expect the 5-year ped/bike crash total to drop by a factor of 0.9999884. For reference, this equates to a ~10% reduction in 5-year ped/bike crash totals for every ~$9,000 increase in median income.
  • Education has a marginal impact on 5-year crash totals. For every +1 increase in average education level, we would expect the 5-year ped/bike crash total to rise by a factor of 1.42561.
  • Average vehicles has an significant impact on 5-year ped/bike crash totals. For every +1 increase in average vehicles, we would expect the 5-year ped/bike crash total drop by a factor of 0.553884 (or, roughly half as many ped/bike crashes for every 1 extra vehicle per household on average)
  • Zero-vehicle density has a significant impact on 5-year crash totals. For every +1 increase in zero-vehicle density, we would expect the 5-year crash total to rise by a factor of 1.00148. For reference, this equates to a ~10% increase for every ~65 increase in zero-vehicle density.
Variable n Estimate Lower bound Upper bound p-value sig
Median household income -1.163e-05 -1.995e-05 -3.308e-06 0.006 **
Average level of education 0.355 -0.009 0.718 0.056 \(\cdot\)
Average vehicles per household -0.591 -1.074 -0.108 0.016 *
Density of zero-vehicle households 0.001 0.001 0.002 <0.001 ***
(Intercept) -0.382 -1.607 0.843 0.541
VMT 4.271e-05 3.739e-05 4.802e-05 <0.001 ***
Context class = C2 6 -0.423 -2.625 1.78 0.707
Context class = C3R 110 0.164 -0.575 0.904 0.663
Context class = C3C 112 0.378 -0.349 1.105 0.309
Context class = C4 386 0.837 0.131 1.542 0.02 *
Context class = C5 19 1.191 0.364 2.017 0.005 **
Context class = C6 10 1.033 0.085 1.981 0.033 *
Context class = SD 4 -1.126 -3.376 1.124 0.327
Posted speed = 30 61 1.014 0.381 1.647 0.002 **
Posted speed = 35 173 1.014 0.419 1.61 <0.001 ***
Posted speed = 40 101 1.163 0.542 1.783 <0.001 ***
Posted speed = 45 253 0.71 0.097 1.323 0.023 *
Posted speed = 50 20 0.959 0.145 1.773 0.021 *
Posted speed = 55 19 0.388 -0.587 1.364 0.435
Posted speed = 60 7 0.27 -1.321 1.862 0.739
Posted speed = 65 6 -1.826 -3.563 -0.089 0.039 *