Sorry for the repeat - looks like my last message was truncated before it went out!

Sam,

Your point is well taken - but one of the really interesting thing about doing CTA research in Chicago is that our market share is so large that I have no trouble finding CTA customers -- even in the suburbs (about 20% of our customers are suburban residents!). This makes it much easier to find users of our products than when I was doing commercial market research before coming to CTA. However, for the TBAS project, the primary output of the survey was are market share, so we were just as interested in surveying former customers and never customers as frequent (5+ days per week) infrequent (1-4 days per week) and occasional customers (<1 per week, but at least once in the last year)

There is geographic variation in our market penetration. My neighborhood in Chicago is so well served by Metra - and travel time to the nearest CTA rail station is so long by bus - that my immediate neighbors are less likely to ride CTA frequently unless they transfer once downtown. However, for much of the rest of the South Side and for many other parts of the service area it takes no time at all to find a frequent (5+ days a week) CTA customer. Whenever I meet people here, the minute I mention working for CTA I am often deluged with comments from CTA customers - If I am at a party on the North or Northwest Side 80 - 90% might be frequent CTA customers. This makes it really easy for me to ask questions that give me insights into whether our survey data collection has been skewed somehow. And so I do. It also means that there is a great deal of self-identification of our customers if I mention working for CTA. Happily, due to the success of our Customer Satisfaction work, I do a lot less referring of customers to our call center and a lot more explanation of new features and programs.

I do not mean to diminish the importance of reality checks in analytical work, especially when engaged in primary data collection. I find it an essential practice to double check even marginally questionable survey results with some kind of testing. In the case of TBAS that included remote monitoring of survey calls, checking key results against other sources of internal and external data, and other checks of the data.

I'll give you an example of a valuable lesson I learned in checking up on concerns about data collection methodology. One of my standard checks with new surveyors for an on-board survey is to make sure - through field observation and examination of surveys returned - that they are surveying all kinds of riders without regard to gender, race, age, or likelihood of being willing to agree do the survey - because I know that every once and a while I will run into problems with these potential sources of bias. For example, a male surveyor in their 20's who comes back with surveys from mostly women in their 20's or 30's (or a female surveyor who does the reverse and turns in many more men than can be expected) - is an obvious potential source of bias. I watch for that even with experienced surveyors and I check for this and other problems during every on-board project, I also try to have a good idea of what demographics to expect on our routes through field observations and pre-testing when I design a project. Given that, I was more than a little concerned when a relatively new male surveyor on a North Lakeshore route turned in far more surveys of women (80-85% female rather than the 55 - 65% which is more typical for bus.. The rail split tends to be 54% female/46% male) than I had expected. Since I had not noticed figures quite that high on my field checks and pre-test, I immediately went out into the field and did a quick gender count on 30 or 40 buses in that corridor. Lo and behold, the surveyor was right and my expectations were wrong. On the particular route he had been assigned there were many more women riding than I had expected.

So, you are right to ask in essence, do I believe that my survey results match reality? I believe they do. A very large portion of the TBAS project was spent on methodological development so that the survey was as accurate as it could be. One known sources of bias may be important. Households without phones are not represented. This is an instance where the census has a clear advantage over a telephone survey. Since those households are more likely to be transit dependent, that suggests that the TBAS survey results could understate our market penetration, especially given the population growth among Hispanic residents in Chicago since this group may be a little less likely to own a phone.

No survey is perfect. Budget requirements often force tradeoff's. If I had the money (probably $2M+ in 2002 dollars) to do a market penetration study with a budget like the Urban Family Life Project that I worked on for NORC in the early 80's, I would use many of the same methods that they used for that study: random sampling of census tract blocks, definition of households (what is a household - do 10 single locking rooms and 1 shared kitchen make 1 household or 10), listing of household residences on sampled blocks including checking the alleys for additional units, random selection of households, listing of the members of the selected household, random selection of the household member to be interviewed with 10 repeat refusals required before they can be dropped from the sample, a budget for paying reluctant respondents, and the survey conducted as an oral interview with extremely clear definitions by highly skilled interviewers who have undergone extensive training to avoid any leading responses, coding of numerous open-ended questions by highly trained and experienced coders, and a big budget for weighting, analysis and reporting, I would do it. Or more likely hire a contractor to do it. The results would be incredibly useful and might create a wealth of innovative ideas and transportation planning data. However, for that kind of money I could do 8-10 system level telephone surveys like TBAS that will yield repeatable, verifiable, sensible results that are a good match for other data and provide meaningful and very useful actionable information.

I realize that our market share information and electronic fare acceptance has been higher than in other systems and so they may seem to be skewed when you compare it to your own experiences in Ohio. One thing that helped believability of TBAS was that we were not just looking for CTA customers. There was no benefit by being either a customer or non-customer with the survey in terms of survey length, for example.

I've avoided sending the survey report to the whole list because of its file size (752kb), but if you would like a copy of the report which includes extensive detail on methodology and the full questionnaire I would be happy to forward it to you. Substantive methodological suggestions are most welcome.

Since you brought up the share of CTA pass and transit card users: our AFC data can provide us very detailed information and very useful data about usage and revenue by fare type that goes far beyond what was available with a single "pass" key on a bus farebox or a press of a pass button at the ticket agent's window. Our system now not only provides detailed reporting of electronic fare use - but - validates the legitimacy of the pass in question - removing a source of arguments for bus riders - and - giving us the ability to offer many more - and more flexible - types of fare media designed to fit niche markets - and - know exactly how much use each farecard or pass type is getting. Having actual data whole categories of problems with interpreting fare transaction data. Our "store" at www.transitchicago.com/store/index.html allows us to reach many more customers - including those who are coming in from out of town - and has increased the use of electronic fare media among bus customers.

I haven't looked farecard usage data in awhile but here's some data on AFC usage at CTA that I happen to have close at hand:

Our AFC system was turned on over the summer of 1997. By October of 97, cash as a share of revenue had dropped from 62.1% to 39.4% of revenue compared to October of the previous year.

When unlinked trips are examined, pass use (excluding Transit Cards) grew from 8% of bus transactions (10/86) to 25% of bus transactions (10/97) after the AFC system had been fully implemented.

Originating (first ride) transit card use on CTA bus grew from 0% of bus transactions to 24% by 10/97, and 27% by 10/99.

On rail, pass use grew from 8% (10/86) to 25% by 10/99. Originating Transit Card Use on rail jumped immediately to 48% of all rail transactions and was at 50% by 10/99.

Cash entries on bus dropped from 27% (10/96) of all bus transactions to 18% (10/99).

On CTA rail cash use dropped from 30% of all entries (10/96) to 6% (10/99).

If you are interested in more information on current Transit Card, pass and cash ratios at CTA I can probably get updated results to you. A lot more detail on the impacts of our AFC system is available in Impacts of Transit Fare Policy Initiatives Under and Automated Fare System, Foote and Stuart, Transportation Quarterly, Volume 54, Number 3, Summer 2000. There is also an earlier article in the Journal of Public Transportation.

Putting the South branch of the Red Line in the median of the Dan Ryan (I-90/I-94) and the Congress Branch of the Blue Line in the median of the Congress Expressway (I-90) was an innovative idea. While it was great in concept, I'm not sure that it had as many benefits as liabilities however.

While it is true that rush hour commuters do see the trains whizzing by, the lack of Park and Ride lots on the Red Line meant that those stuck in traffic could not pull off the expressway and use CTA.

Worse yet, highway planners failed to build a NB exit from either 94 or 57 at 95th street so that drivers headed to the Loop from south east and south west of the station had no easy method to access even on-street parking at the 95th street terminal. Its almost as if highway planners didn't want auto drivers coming from the south suburbs and Indiana to switch to CTA!

Another problem is that while CTA trains can get up to 50 or 55 mph between stations, auto drivers who are willing to break the speed limit in "non-rush" conditions, are able to speed past CTA trains instead of the other way around. Granted, the Dan Ryan is so crowded that non-rush conditions tend to happen around noon and after 8pm on weekdays, this effect somewhat neutralizes the rush hour effect.

The Congress branch does have some park and ride lots and worse traffic problems, so the benefits you describe are better there.

On both branches however, there is one other negative impact. Putting the station in the median isolates the station from the surrounding communities. So, very little access to the station is by walking, much more is by bus. There is a longer walk to nearby stores which reduces some of the beneficial impact of the lines (dry cleaning, small grocery stores etc) on the community. Also walkers must cross very busy intersections to get to the station which often forces a transfer from bus that might not otherwise be needed. Expressway accidents (snow plows pushing through concrete barriers, toxic spills, autos flipping onto the tracks) can occasionally impact service.

The Orange Line routing seems to get the benefits of both which could be a key to its success. While much of the route can be seen from the expressway, the station entrances are at street level with plenty of walking traffic. The stations seemed to be much more integral to the surrounding communities. Access still may involve crossing a city street, but drivers are not gearing up to enter an expressway.

If I was building the south branch of the Red Line today, I probably would have put it next to the expressway rather than in it - but all hindsight is 20/20. Another key to Orange Line success was the Stevenson Expressway reconstruction which followed the opening of the Orange Line.

All the best!

Peter J. Foote
PFoote@TransitChicago.com

Market Research/Resource Planning
Planning Division
Chicago Transit Authority

120 N. Racine Avenue
Chicago, IL 60607

(312) 733-7000 x 6840 (Voice)
(312) 432 - 7108 (Fax)

****************************************************************************
****************************************************************************
*******************************************
Unless otherwise stated, the views expressed are those of the author and not
those of the Chicago Transit Authority.
****************************************************************************
****************************************************************************
*******************************************
http://www.TransitChicago.com

CTA CUSTOMER SERVICE HOTLINE
1-888-YOUR-CTA

RTA Travel Information
836-7000

----- Original Message -----

From: Sam.Granato@dot.state.oh.us

To: Gardner, John F

Cc: ctpp-news@chrispy.net ; owner-ctpp-news@chrispy.net

Sent: Wednesday, July 10, 2002 9:30 AM

Subject: RE: [CTPP] Reconciling Census Transit Commuters...Impact of Rail Lines etc

Try this out to guage the difference between survey answers and reality - ask a group of people if any of them take public transit, then ask them to show you their transit passes (daily, monthly, whatever).

Someone in Chicago at least had the foresight to put the rail system in the freeway medians, so commuters stuck in rush-hour traffic can watch the trains at least appear to get to their destination faster than their car is. So no out-of-sight-out-of-mind problem that transit has most everywhere else.

Sam Granato
Ohio DOT, Office of Technical Services
1980 W. Broad Street, Columbus, OH 43223
Phone: 614-644-6796, Fax: 614-752-8646
"Men and nations will all eventually do the right thing - after first exhausting all the other alternatives." - Abba Eban

"Gardner, John F" <GardnerJF@dot.state.sc.us>
Sent by: owner-ctpp-news@chrispy.net
07/10/02 09:05 AM

To: "'Peter J. Foote'" <pfoote@transitchicago.com>, ctpp-news@chrispy.net
cc:
Subject: RE: [CTPP] Reconciling Census Transit Commuters...Impact of Rail Lines etc

Peter Foote's responses are excellent. Although I suspect the smaller
metros have had such an increase in choice riders, because the congestion
and parking cost incentives are not as strong.

Very good point about the impact of prepaid fare media -- day passes,
monthly passes, etc. -- boosting off-peak ridership. Prepaid fares help
level the playing field between cars and transit. Drivers behave as if the
marginal cost of an additonal trip or mile is zero; unlimited-ride passes
induce the same behavior in transit users.

-----Original Message-----
From: Peter J. Foote [mailto:pfoote@transitchicago.com]
Sent: Tuesday, July 09, 2002 8:34 PM
To: Mark Schlappi; 'Chuck Purvis'; ctpp-news@chrispy.net
Cc: dstuart@transitchicago.com; mpatzloff@transitchicago.com
Subject: Re: [CTPP] Reconciling Census Transit Commuters...Impact of
Rail Lines etc

The impact of a new rail line, branch or extension on a system's unlinked
trip count depends upon many factors.

In general terms, the implementation of rail systems should generate new
unlinked trips because they tend by nature to draw new customers, including
suburban men, out of their cars. This is especially true when parking fees
in the CBD are high. Links can be lost however, when rail route design
allows former bus riders to save one or more transfers previously made on
their commute.

CTA's Orange line (running diagonally across the grid SW from the Chicago's
Loop) which opened 10/31/93, was a bonanza for Pace suburban bus ridership
for routes connecting to CTA at Midway airport. The line exceeded CTA
ridership expectations in part due to high suburban ridership, but also
resulted in some lost linked trips (not individuals) between bus routes -
and of course a substantial shift from bus to rail.

Limited stop express services can add individuals overall but eliminate the
need for some transfers when they cross over route ends. The X49 for
example, eliminated some transfers (unlinked trips) between routes 49A and
49, however ridership increases and customer satisfaction rose far more than
trips lost.

In Chicago, it is more likely that increases in off-peak discretionary
riding by electronic pass users (30 and 7 day rolling pass markets have
grown substantially since the were re-introduced in the Winter of 1998) are
responsible for a portion of this shift. New York of course had tremendous
ridership growth when the eliminated the 2nd fare required to move from bus
to rail when their Metro Card was used. Check with them, but my guess would
be that they have also had big boosts in off-peak riding, etc stimulated by
having pre-paid media available. DC's Smart Card probably generates many
new rides as well. Could the impact of just these 3 systems be enough to
cause an increase in trips with no new net gain in riders? Once you have a
farecard, even a money card rather than a pass, you are much more inclined
to hop on a bus or train to take a trip. Discounting and the impact of
lowered average fares on ridership through the use of these fare instruments
may also be a factor.

The issue of whether census data questions are complete enough to extract
true mode splits (raised in another thread) is a valid one, especially when
respondents are asked about an entire previous week is a valid one. BUT, I
also think that it is unrealistic to expect more from the Census that can
reasonably be expected.

Reconfiguration of a rail line can cause losses. When the Dan Ryan/Lake and
Howard/Englewood/Jackson Park were realigned 2/1993 to connect the busy
South and North branches in the Red Line and less busy South and West
branches as the Green Line many transfers which had been recorded as
unlinked trips were eliminated with no real reduction in persons (average
trip lengths expanded). Reconstruction of a rail line can also cause
losses. The decision to close the Green line for reconstruction rather than
single track in 1/1994 resulted in bus and offset gains made by the new
Orange Line rail.

The impacts of service changes are often complex. It likely that changes in
auto ownership and transit riding frequency among fairly frequent customers
over the decade are also an important of the story. It is also very
difficult to make assessments of persons using a system from unlinked trip
counts. This is part of why NTD average trip length calculations and
passenger miles estimations are important.

Peter J. Foote
PFoote@TransitChicago.com

Market Research/Resource Planning
Planning Division
Chicago Transit Authority

120 N. Racine Avenue
Chicago, IL 60607

(312) 733-7000 x 6840 (Voice)
(312) 432 - 7108 (Fax)

****************************************************************************
****************************************************************************
*******************************************
Unless otherwise stated, the views expressed are those of the author and not
those of the Chicago Transit Authority.
****************************************************************************
****************************************************************************
*******************************************
http://www.TransitChicago.com

CTA CUSTOMER SERVICE HOTLINE
1-888-YOUR-CTA

RTA Travel Information
836-7000

----- Original Message -----
From: "Mark Schlappi" <Schlappi@mag.maricopa.gov>
To: "'Chuck Purvis'" <CPurvis@mtc.ca.gov>; <ctpp-news@chrispy.net>
Sent: Wednesday, July 03, 2002 6:01 PM
Subject: RE: [CTPP] Reconciling Census Transit Commuters with Ridership
Statistics

> Is there a possibility that changing route structures, with more rail,
have
> caused more unlinked trips?
>
> -----Original Message-----
> From: Chuck Purvis [mailto:CPurvis@mtc.ca.gov]
> Sent: Wednesday, July 03, 2002 3:36 PM
> To: ctpp-news@chrispy.net
> Subject: [CTPP] Reconciling Census Transit Commuters with Ridership
> Statistics
>
>
> To: CTPP-News
>
> One of the interesting journey-to-work results is the lack of change, at
the
> NATIONAL level, in the total number of transit commuters. The US had
> 6,069,589 transit commuters according to the 1990 Census, and 6,067,703
> transit commuters according to the 2000 Census, a 0.03 percent decrease.
(On
> the other hand, the US transit commute SHARE declined from 5.3 percent in
> 1990 to 4.7 percent in 2000.)
>
> This compares to national transit ridership statistics which show a 6.4
> percent increase in annual unlinked passenger trips comparing 1990 to
2000.
> (Source is APTA's 2002 Public Transportation Fact Book, Table 26.) The
APTA
> book (I would assume based on FTA-collected ridership statistics) shows
> annual unlinked public transit trips increasing from 8,799 million trips
in
> 1990 to 9,363 million trips in 2000 (the 2000 numbers are preliminary,
> according to the 2002 APTA Fact Book).
>
> So, an issue is how to reconcile a 0.0 percent change in national transit
> commuters with a 6.4 percent increase in national transit ridership.
>
> A plausible explanation is that the work trip share of public
transportation
> trips has declined since 1990. According to the 1990 NPTS, 42.6 percent of
> public transportation person trips are for the purpose of "earning a
living"
> (NPTS Databook, Vol. 1, Table 4.40).
>
> So, I can calculate that about 3,748 million unlinked transit boardings
(in
> 1990) are "work trips" and that this might decline to about 3,747 million
> unlinked transit boarding "work trips", in 2000. This means that perhaps
40
> percent of year 2000 transit boardings are work trips (3,747 / 9,363),
which
> is quite plausible at the national level.
>
> The story might be that the national number of transit work trips, 1990 to
> 2000, has remained fairly stable, and that, at least at the national
level,
> the growth in transit is attributable to non-work travel.
>
> The data question is: is information available from either the 1995 NPTS
or
> the 2001 NHTS that can corroborate this possible trend - - a decline in
the
> work purpose share for public transportation trips?
>
> Also, who has attempted to reconcile their change in regional transit
> commuters with their own transit ridership statistics? What would be most
> helpful is any comparisons of on-board surveys or household travel surveys
> that show any changes in the trip purpose mix for transit trips.
>
> (Other larger issues still loom in terms of the plausibility/fixability of
> the Census data. We are very concerned about the overall LOW numbers of
> TOTAL commuters and employed residents we're seeing in the 2000 Census in
> our region....)
>
> Wishing a Safe & Sane Happy Fourth of July to All!
>
> Chuck Purvis
>
>
> ***********************************************
> Charles L. Purvis, AICP
> Senior Transportation Planner/Analyst
> Metropolitan Transportation Commission
> 101 Eighth Street
> Oakland, CA 94607-4700
> (510) 464-7731 (office)
> (510) 464-7848 (fax)
> www: http://www.mtc.ca.gov/
> Census WWW: http://census.mtc.ca.gov/
> ***********************************************
>