2 Reliability
Reliability is measured by ontime performance of buses, which is calculated as the percentage of the time that a bus arrives at key stops (time points) at scheduled times. This data can be downloaded from MBTA’s Performance Dashboard at http://www.mbtabackontrack.com/performance/index.html#/home.
In 2016, the systemwide ontime performance of buses was 67.1%. In 2017, it improved slightly to 68.3%. Across the two years of available data, the systemwide ontime performance was 67.8%.
There is significant variation in ontime performance across the system, ranging from a low of 38.4% to a maximum of 100% ontime performance. Route 111 had 77.6% ontime performance over this time period, indicating its performance exceeded that of three-quarters of all other routes. Table 2.1 below shows the 10 worst routes in terms of ontime performance for 2016, 2017, and averaged across the two years 2016-2017.
| Worst Routes 2016 | Ontime % 2016 | Worst Routes 2017 | Ontime % 2017 | Worst Routes 2016-17 | Ontime % 2016-17 |
|---|---|---|---|---|---|
| 434 | 34.9 | 9703 | 33.2 | 434 | 38.4 |
| 449 | 37.5 | 434 | 40.7 | 9703 | 38.9 |
| 459 | 40.7 | 41 | 42.2 | 449 | 41.3 |
| 448 | 40.8 | 448 | 43.6 | 41 | 42.1 |
| 41 | 42.0 | 449 | 43.7 | 448 | 42.5 |
| CT2 | 42.9 | 459 | 44.3 | 459 | 42.9 |
| 9703 | 45.7 | 195 | 44.4 | CT2 | 45.6 |
| 08 | 46.7 | 105 | 45.8 | 105 | 46.6 |
| 558 | 47.1 | 325 | 46.0 | 137 | 47.4 |
| 14 | 47.2 | 171 | 46.3 | 411 | 47.8 |
Figure 2.1 below is an interactive map of bus lines colored by ontime peformance. Green shaded lines are above the systemwide ontime percentage, while red lines are below the systemwide ontime percentage. Hover cursor over lines to see specific route numbers. Click on a line for more detail. Pan (click and drag) or zoom (+/-) on the map to explore.
Figure 2.1: Map of ontime percentages by bus route for 2016-17. Red lines indicate ontime percentages below systemwide percentage. Green are above systemwide ontime percentage.
2.1 Reliability vs Demographics
2.1.1 Minority vs Ontime Performance
One of the key questions in analyzing transit equity is whether vulnerable groups face disparate treatment or access to resources. The first thing to look at is whether there is a systematic relationship between route demographics and the level of service. Figure 2.2 below is a scatterplot of percent minority of each bus route vs percent on time of the same bus routes. The blue line is a linear trend line. The slighly negative slope of the trend line suggests a weak negative or inverse relationship between percent minority and percent ontime. In other words, the higher the percentage minority of a line, the lower the percentage of ontime performance. Hover your cursor over individual points on the graph to see the Route number and ontime percentage.
Figure 2.2: Scatter plot of percent minority vs percent ontime. Red dot is Route 111.
While there is no simple linear relationship between percent minority and percent ontime bus service, there is a fairly consistent difference in ontime performance between minority lines and non-minority lines in aggregate. Figure 2.3 below shows boxplots of the range of ontime performance for all minority routes and for all non-minority routes using different minority thresholds. Because the definition of what constitutes a “minority” line is debatable, the boxplots show the impact of choosing the MBTA’s systemwide percentage threshold for defining a minority line versus the states’s EJ policy threshold. The difference between minority and non-minority ontime performance is apparent either way, but it is less apparent under MBTA’s higher threshold for defining minority lines. In the aggregate, non-minority bus riders enjoy more reliabe - better ontime performance - bus service than minority bus riders. No minority bus line experienced better than 79% ontime performance over the period 2016 - 2017.
Figure 2.3: Boxplots of bus route ontime percentages for minority and non-minority lines when the minority is defined by the state EJ threshold of 25% or MBTA threshold of 34.3%.
2.1.2 Low Income vs Ontime Performance
The relationship between ontime performance and percent of riders who are low income is slighlty more pronounced. Figure 2.4 below is a scatterplot of percent low income of each bus route vs percent on time of the same bus routes. The blue line is a linear trend line. The slighly negative slope of the trend line suggests a negative or inverse relationship between percent low income and percent ontime. In other words, the higher the percentage of low income, the lower the percentage of ontime performance. This is not a strong relationship, but it is more pronounced than for minority percentages. Hover your cursor over individual points to see the Route number and ontime percentage.
Figure 2.4: Scatter plot of percent minority vs percent ontime. Red dot is Route 111.
Figure 2.5 below shows boxplots of the range of ontime performance for all low income routes and for all non-low income routes using different low income thresholds. Because the definition of what constitutes a “low income” line is debatable, the boxplots show the impact of choosing the MBTA’s systemwide percentage threshold for defining a low income line versus the states’s EJ policy threshold. Similar to minority vs non-minority lines, the difference between low income and non-low income ontime performance is apparent either way. In this case, but it is less apparent under MBTA’s higher threshold for defining low income lines. In the aggregate, non-low income bus riders enjoy more reliabe - better ontime performance - bus service than low income bus riders. No low income bus line experienced better than 79% ontime performance over the period 2016 - 2017.
Figure 2.5: Boxplots of bus route ontime percentages for low income and non-low income lines when the low income line is defined by the threshold of 25% or MBTA threshold of 28.8%.
2.2 Reliability Ratios
While there are clear differences in ontime performance for minority and low income riders, these differences do not exceed the MBTA’s threshold for determining disparate or dispropotionate impacts. The MBTA uses a 20-percent threshold for identifying potential disparate impacts for service monitoring, although this threshold is not codified in its most recent 2016 Disparate Impact and Disproportionate Burden Policy. In other words, a disparate or disproportionate impact exists only when there is a 20-percent or greater difference in service, amenities, or impacts between minority or low income riders and non-minority or non-low income riders.
Table 2.2 below shows ratios of ontime performance percentages between minority and non-minority lines and between minority lines and the systemwide ontime percentage. Only the latter ratio is used by the MBTA to determine whether a disparate impact exists. For a disparate impact to exist according to the MBTA, minority lines would need to have a ratio of 0.8 or lower relative to the systemwide percentage. Although there is a consistent difference in performance, it does not exceed this 0.8 ratio threshold.
Note that the ratio of difference is affected by both the threshold used to define minority lines (as described above) and also by whether the minority ontime percentages are compared to non-minority lines or the systemwide percentage.
| Minority Threshold | Minority Ontime % | Non-Minority Ontime % | Ratio Minority/Non-Minority | Ratio Minority/System |
|---|---|---|---|---|
| 25% | 66.70 | 73.14 | 0.91 | 0.98 |
| 34.3% | 67.13 | 69.33 | 0.97 | 0.99 |
Table 2.3 below shows ratios of ontime performance percentages between low income and non-low income lines and between low income lines and the systemwide ontime percentage.
| Low Income Threshold | Low Income Ontime % | Non-Low Income Ontime % | Ratio Low Income/Non-Low Income | Ratio Low Income/System |
|---|---|---|---|---|
| 25% | 67.02 | 71.38 | 0.94 | 0.99 |
| 28.8% | 67.14 | 69.68 | 0.96 | 0.99 |