2001 Riders' Survey Results

2001 Riders' Survey Results

Is Muni getting better? This is the question we ask every year when we conduct our annual Muni Riders' Survey. This year, we are once again pleased to report that Muni service quality is improving, although at a slower pace than last year. Muni's on-time performance improved again, with 17.3 percent of our riders experiencing delays on average, earning Muni another B-minus for on-time service; Muni's service quality also became more consistent across the various lines and modes.

Rescue Muni conducted its fifth annual survey (sixth total) in February of 2001. This year, 65 volunteers recorded 1773 bus and streetcar rides, noting how long they waited and how long the trip took. Some volunteers also stood at fixed points and recorded all vehicles that came by; 350 of the 1773 trips were recorded that way. As we have done since 1997, we compared riders' actual waiting times with the headways advertised on Muni's Street and Transit Map (posted in all bus shelters) and service bulletins. Riders who waited more than the full headway are considered delayed - thisis a fairly liberal measure, but one that is consistent with riders' expectations. We also asked riders to record crowding on a scale of 1 to 5, and (for the first time) whether a vehicle was clean or not.

(Note that this year, service was added on the 12-Folsom line as part of the South of Market service realignment during the survey period; all 12 trips were taken after this change, so the new timetable was used.)

Table 1: Best and worst lines; systemwide performance

Route % riders delayed Grade Change
01-00 2000
% late 1999
% late 1998
% late Total
responses

Total 17.3% B -1.6% 18.9% 25% 28% 1773

Best five lines:

2 0% A 19% 9% 28

108 0% A 0% 0% 0% 21

43 3% A -8% 12% 26% 23% 49

6 4% A -14% 18% 21% 21% 49

27 5% A 2% 22

Worst five lines:

21 29% C +14% 14% 26% 30% 43

67 29% C 24

M 29% C +5% 25% 26% 31% 24

J 30% C +4% 25% 36% 42% 61

38 32% D -19% 52% 33% 26% 37

On-time service improved systemwide as measured in our survey. As noted above, Muni's overall rating improved to 17.3 percent of riders delayed (B-minus grade), a small but significant improvement from 2000. Compared to Muni's nadir in 1998, however, this is more impressive: 28% of riders were delayed then, so Muni has reduced delays by more than a third. Normalized waiting time got slightly worse systemwide; customers waited an average of 76% of posted frequency (a 50% average is optimal).

On-time service was more consistent this year across the modes, with the F-Market historic streetcar continuing to provide excellent service (graded A, 9% late) while other modes showed improvement as well. Diesel and trolley coach lines both reduced delays a small amount from 2000, with 13% and 17% riders delayed respectively this year. (Both earned a B grade.) Muni Metro service did not get better, however: as in 2000, 20% of riders were delayed. A chart of performance by mode is at right.

Service was also more consistent across various lines. For the first time this year, no line was graded F and only one (the 38-Geary) was graded D, with 32% of riders delayed. Of course, many lines were graded C, with delays in the 20-30 percent range, including many high-ridership lines (J, L, M, 22, 42, and more); but the huge variations in quality we saw in 1997-99 are not present here. Muni deserves significant credit for this progress - although no line should delay even 20% of its customers, the appalling behavior of lines like the "L-Terrible" in 1998 (53% late!) seems to be no more.

8 of the 30 lines with sufficient data to report (20 points or more) were graded A, including four diesel lines that use new equipment: 2-Clement, 108-Treasure Island, 43-Masonic, and 27-Bryant. Clearly fleet replacement is making a difference for on-time service.

Crowding is an area where service is not improving. In 2001, more vehicles were crush-loaded (17%, up from 12% last year) and many more were standing room only (59%, up from 53% last year). Our crowding chart on page 19 shows this clearly. Some of this is the result of increased ridership (a good thing!) but Muni needs to monitor crowding closely and make sure that sufficient equipment is provided. The L-Taraval and 108-Treasure Island both had average crowding over 4 on a 1-5 scale - clearly too many of these trips were overcrowded.

Finally, we measured cleanliness for the first time. Here Muni has some room for improvement. 79% of Muni vehicles were clean systemwide, but several lines had a problem staying clean; the 108 was the worst, with over half of its vehicles dirty. JKLMN service from Castro to Market was also a problem here. The F was one of the best again; this is in part due to the hard work done by the Market Street Railway to keep the historic cars in shape. Maintenance crews of the other lines should take note - and Muni staff should consider asking operators of lines where cleaners are not available during the day to "pitch in" and collect the garbage when customers don't do their part. Enforcing a zero-tolerance policy against litter (as on BART) could help as well.

Table 2: Cleanest and dirtiest lines

line % clean grade

Best:

2 100% A

6 100% A

1 100% A

42 100% A

F 97% A

Worst:

JKLMN 65% D

33 62% D

K 62% D

15 60% D

108 48% F

Total 79% C

This year, we noted another drop in participation in the survey. Our volunteers were as dedicated as in previous years, but there were fewer of them; our 1773 data points for a full month of service were the fewest for a systemwide survey since the first survey in 1997. This unfortunately means that we do not have good data for as many lines - the "B" grade given the 14-Mission, for example, is only based on 13 data points, clearly not a sufficient sample. For Rescue Muni, of course this is bad news; but it may be reflective of higher confidence in Muni, as customers feel that they don't have to submit a survey is service is "okay." We will study the comments submitted to try to understand this better before next year's survey.

So is Muni getting better? Yes, it is. This survey shows real progress since last year, and big improvements since Muni's worst days in 1998. But there is still more to be done; in transit-first San Francisco, we should not declare victory until all of Muni's customers get service that earns an "A" grade. Service expansion will help here (see our detailed report in this newsletter), and continued fleet replacement should also make a big difference, but the hard work by Muni employees to plan and provide quality service will, as always, be the key to continued progress.

2001 Muni Riders' Survey: Complete Results

route

Total

resps

%

late

grade

chg

01-00

avg

wait

norm

wait

avg

crowd

% clean

clean

grade

2000 %late

1999

%late

1998

%late

*

2 28 0% A 0:01 15% 2.21 100% A 19% 9%

108 21 0% A 0% 0:04 9% 4.10 48% F 0% 0%

43 49 3% A -8% 0:10 89% 1.75 88% B 12% 26% 23%

6 49 4% A -14% 0:11 91% 2.29 100% A 18% 21% 21%

27 22 5% A 0:03 25% 3.91 82% B 2%

44 53 8% A -2% 0:12 83% 2.71 92% A 10% 25% 9%

F 33 9% A -12% 0:04 37% 2.25 97% A 21% 11% 13%

JKLMN 52 10% A -5% 0:01 46% 2.40 65% D 15% 20%

1 64 13% B -11% 0:05 65% 2.59 100% A 24% 28% 23%

33 43 14% B 6% 0:09 50% 2.64 62% D 8% 12%

9 21 14% B 9% 0:06 57% 3.33 85% B 5% 31% 27%

24 79 14% B -2% 0:06 62% 2.72 80% B 17% 22% 30%

30 36 15% B -35% 0:04 67% 2.42 81% B 50% 26% 21%

N 348 16% B -5% 0:07 81% 3.11 74% C 20% 23% 42%

15 29 17% B 0:08 92% 2.36 60% D 19% 31%

47 24 17% B 2% 0:02 56% 3.17 84% B 14% 40%

K 31 19% B 16% 0:08 69% 2.31 62% D 4% 32% 41%

26 28 20% C 0:17 94% 2.25 92% A 36%

49 48 21% C -8% 0:11 93% 2.57 87% B 29% 23% 29%

30X 24 21% C 21% 0:04 71% 2.92 93% A 0% 20%

5 39 26% C 14% 0:05 63% 2.71 86% B 12% 16% 28%

22 34 26% C 5% 0:06 83% 3.00 66% D 21% 22% 29%

42 27 27% C -3% 0:08 124% 2.94 100% A 30% 36% 25%

KLM 165 28% C 10% 0:05 100% 3.38 69% D 18% 22% 14%

L 43 28% C -0% 0:06 78% 4.20 69% D 28% 26% 53%

21 43 29% C 14% 0:10 105% 3.17 77% C 14% 26% 30%

67 24 29% C 0:25 128% 2.83 67% D

M 24 29% C 5% 0:10 92% 2.71 89% B 25% 26% 31%

J 61 30% C 4% 0:06 103% 2.65 90% A 25% 36% 42%

38 37 32% D -19% 0:07 107% 3.03 79% C 52% 33% 26%

18 13 0% A -11% 0:07 43% 2.77 75% C 11% 10% *

45 17 2% A -33% 0:05 46% 1.63 63% D 36% 23% 16% *

16BX 14 7% A 1% 0:04 37% 3.57 100% A 6% 19% *

71L 11 9% A 1% 0:04 47% 3.91 100% A 8% *

29 10 10% A -8% 0:09 59% 1.80 70% C 18% 40% *

71 10 10% A -19% 0:06 54% 2.70 90% A 29% 23% 31% *

14 13 15% B -17% 0:04 68% 2.75 67% D 32% 47% 51% *

12 12 17% B 0:04 49% 1.60 100% A *

23 12 17% B 0:11 63% 3.00 80% B 13% *

Total

1773

17.3%

B

-2%

0:07

76%

2.94

79%

C

18.9%

25%

28%

Notes:
1. Routes in italics with an asterisk (*) in the right column had fewer than 20 responses; we are reporting them here for completeness, but these results should be considered less accurate than those in roman type.
2. Crowding is on a scale of 1 (empty) to 5 (jammed).
3. Routes not linked to Transit Info are either no longer in service (42, old 47) or not advertised in aggregate (KLM, JKLMN).

[ RM Home Page ]

On-time service improved systemwide as measured in our survey. As noted above, Muni's overall rating improved to 17.3 percent of riders delayed (B-minus grade), a small but significant improvement from 2000. Compared to Muni's nadir in 1998, however, this is more impressive: 28% of riders were delayed then, so Muni has reduced delays by more than a third. Normalized waiting time got slightly worse systemwide; customers waited an average of 76% of posted frequency (a 50% average is optimal).
On-time service was more consistent this year across the modes, with the F-Market historic streetcar continuing to provide excellent service (graded A, 9% late) while other modes showed improvement as well. Diesel and trolley coach lines both reduced delays a small amount from 2000, with 13% and 17% riders delayed respectively this year. (Both earned a B grade.) Muni Metro service did not get better, however: as in 2000, 20% of riders were delayed. A chart of performance by mode is at right.
Service was also more consistent across various lines. For the first time this year, no line was graded F and only one (the 38-Geary) was graded D, with 32% of riders delayed. Of course, many lines were graded C, with delays in the 20-30 percent range, including many high-ridership lines (J, L, M, 22, 42, and more); but the huge variations in quality we saw in 1997-99 are not present here. Muni deserves significant credit for this progress - although no line should delay even 20% of its customers, the appalling behavior of lines like the "L-Terrible" in 1998 (53% late!) seems to be no more. 8 of the 30 lines with sufficient data to report (20 points or more) were graded A, including four diesel lines that use new equipment: 2-Clement, 108-Treasure Island, 43-Masonic, and 27-Bryant. Clearly fleet replacement is making a difference for on-time service. Crowding is an area where service is not improving. In 2001, more vehicles were crush-loaded (17%, up from 12% last year) and many more were standing room only (59%, up from 53% last year). Our crowding chart on page 19 shows this clearly. Some of this is the result of increased ridership (a good thing!) but Muni needs to monitor crowding closely and make sure that sufficient equipment is provided. The L-Taraval and 108-Treasure Island both had average crowding over 4 on a 1-5 scale - clearly too many of these trips were overcrowded.
Finally, we measured cleanliness for the first time. Here Muni has some room for improvement. 79% of Muni vehicles were clean systemwide, but several lines had a problem staying clean; the 108 was the worst, with over half of its vehicles dirty. JKLMN service from Castro to Market was also a problem here. The F was one of the best again; this is in part due to the hard work done by the Market Street Railway to keep the historic cars in shape. Maintenance crews of the other lines should take note - and Muni staff should consider asking operators of lines where cleaners are not available during the day to "pitch in" and collect the garbage when customers don't do their part. Enforcing a zero-tolerance policy against litter (as on BART) could help as well.