This quote is from Tom Downs as forwarded by Ken Orski
"The survey data, census long
form, is increasingly viewed as biased in its results (response rates of
less than 20% in many regions, language biased in an age of immigrants,
poverty biased, and generally reflective of a white middle class data
set)"
The long form is DESIGNED to reach only 1 out of 6 (or is it now 8, i am not
sure) so when compared to the universe, the response (actually, sampling)
rate is under 20% by design, not by the failure to return the form. The
Census corrects for this in all the estimates of the universe long form
statistics, with the sample results having a lower confidence (higher +/-
range) interval.
I have no idea what percentage of long forms distrubuted are returned or
completed, but I would expect and hope it is over 95% in most cases.
And, as I recall the non-response "corrections" are only prohibited when
counting the short form data for apportionment for the congress. I surely
hope that non-response adjustments are included as the detailed CTPP data is
released.
Also, from what I recall, we fought and fought to save the transportation
questions from total elimination and had low expectation of adding non-work
travel mode questions from day one to the ACS.
As to the opinions of Mr Cox, they of course could be challenged, but most
of that is a subject for another thread. None of his statistical data would
appear to be that far off, but his use of "market share data" is highly
flawed. No subway or fixed route transit system is expected to increase
market share of "all" work trips when many of the newly created trip ends
are located as to require 2 feeder bus trips of substantial length and/or
very poor headways.
Ed Herlihy
Reston, VA
----- Original Message -----
From: "C Kenneth Orski" <korski(a)erols.com>
To: <ctpp-news(a)chrispy.net>
Cc: "Cox, Wendell" <wcox(a)publicpurpose.com>
Sent: Friday, July 12, 2002 5:18 PM
Subject: [CTPP] [Fwd: Re: "The Truth About Transit", Baltimore Sun, July 12,
2002]
Here is a colleague of mine challenging the validity
of the 2000
Census data on transit ridership. I am not a statistician and have no
way of assessing the soundness of his claims. Can anybody help me?
Ken Orski
-------- Original Message --------
Subject: Re: "The Truth About Transit", Baltimore Sun, July 12, 2002
Date: Fri, 12 Jul 2002 15:48:00 -0400
From:"Tom Downs" <tdowns(a)ursp.umd.edu>
To: <korski(a)erols.com>
C Kenneth Orski wrote:
You should read this... and ponder
http://www.baltimoresun.com/news/opinion/oped/bal-op.transit12jul12.story?co
ll=bal%2Doped%2Dheadlines
=======================
I would be glad to ponder the following:The survey data, census long
form, is increasingly viewed as biased in its results (response rates of
less than 20% in many regions, language biased in an age of immigrants,
poverty biased, and generally reflective of a white middle class data
set)The data does not reflect the fact that journey to work is about 20%
of trips and declining. Do we care about the explosion in "other trips"
and transit's role in them? We also have no clue as to what is happening
to pedestrian trips...Tom Downs
#################################################################
#################################################################
#################################################################
#####
#####
#####
#################################################################
#################################################################
#################################################################