My Two Census

Formerly the non-partisan watchdog of the 2010 US Census, and currently an opinion blog that covers all things political, media, foreign policy, globalization, and culture…but sometimes returning to its census/demographics roots.

Workers: Problems Cloud W. Kansas 2010 Census Effort

I rarely classify an article as a MUST-READ, but the article below is surely a MUST-READ. Thanks to the stellar reporting of Mary Clarkin of The Hutchinson News, we have the following piece:

Three who had roles in address check speak of technological, pay issues.

By Mary Clarkin – The Hutchinson News

Lisa Stone, Patricia Wedel and Pamela Richardson thought they knew what they were getting into.

Favorable experience working on the U.S. Census prompted them to sign up to help Kansas count its population for 2010. The process began this spring with the address canvassing phase.

It does not generate much notice because it is not the actual people-counting event, but the address check is the “cornerstone of a successful Census,” the acting director of the U.S. Census Bureau told Congress in March.

Appointed to leadership roles in their respective territories, Stone of Dodge City, Wedel of Wichita, and Richardson of Manhattan, set to work.

But within hours of one another in late May, all three women resigned in protest. In separate interviews, they described what happened – and told why they strongly suspect the data for western Kansas is deficient.

Stone’s territory

At the end of crew leader training, Stone learned she would be in charge of quality control for the western third of the state. She asked if the territory could be divided. She said she was told no, the decision had been made at the regional Kansas City, Mo., office responsible for a six-state area.

“I’m not really a quitter. I’ve taken on big jobs before,” she said, so she tackled it.

Recruiting foot soldiers to verify addresses and then sending out another wave of workers to audit the results became challenging in western Kansas. Stone said the pay was partly to blame.

In western Kansas, the hourly pay rate for Census workers was $10, plus mileage. Elsewhere in the state, some people performing the same task received at least $3 an hour more.

Another handicap was technology. For the first time, Census workers had to use handheld computers, instead of pencil and paper.

“The handheld computers had a pretty high failure rate. By my estimate, about 40 percent of them weren’t usable. They had to be sent back to Topeka and swapped out,” Stone said. “They would send us replacements, but we wasted a lot of time.”

Sprint was the telecommunications network for the Census, but the scarcity of Sprint towers in western Kansas required workers to plug the handheld computers into a phone jack to transmit data.

“That really slowed things down, and there were days when the transmit just wouldn’t go through,” Stone said.

As Census officials monitored progress across Kansas, Stone found her district compared with more urban areas with plenty of workers and better technology connections.

“They kept telling me to go faster,” she said.

Headquarters supplied more people, she said. Unfortunately, extra field workers only exacerbated problems of overloaded handheld computers.

“It did fry out the circuits on one person’s computer,” she said.

Wedel feels the heat

Westen Kansas was always behind in the process, and Patricia Wedel said she “took the heat for that.”

Wedel was field operations supervisor in quality control for about two-thirds of the state, including the western half. As Stone’s superior, Wedel was empathetic.

The territory should have been broken into more districts, in Wedel’s view.

“Topeka claimed they couldn’t get enough employees in western Kansas,” she recalled.

She also attributed part of the problem to the decision to pay workers there the lowest rate.

Wedel, too, was vexed by the technology hurdles, requiring most workers to seek out a landline to transmit data.

Census officials advised workers to use the phone jacks at city halls or police stations, Wedel said, but that had the drawback of potentially tying up the only landline at some public offices.

“It was a very frustrating experience,” Wedel said, citing pressure from both the Topeka and Kansas City offices to move quickly.

Informed by a quality-control worker that the data for Clark County was incorrect, Wedel became concerned. But, she said, officials repeatedly stressed they wanted workers to limit the workweek to 40 hours.

There was “no way” they could accomplish the job under those time constraints, Wedel said.

“They really made it impossible for Kansas to get an accurate count out west,” she said.

Wedel expressed dismay at a lack of professionalism she sometimes experienced, including the time an official in the Kansas City office urged management to “kick their butts” to speed up the productivity of workers in the field.

Richardson takes calls

As a crew leader for a north-central Kansas territory that included Russell County, Richardson also wound up with more counties – 20 in all – than she originally anticipated.

“Really from the beginning, it was like bait and switch,” Richardson said.

She initially received the wrong address for the Abilene training site – the address given belonged to a private residence.

Because of another event occurring in Abilene, the Census ended up booking rooms in a cheap motel for training class participants, Richardson said. Some rooms didn’t have working phones or temperature controls. Richardson used her own card to guarantee rooms at another motel for some of the trainees.

In a perfect world, maybe the four-day allotted training would have been enough, she said.

But handheld computers crashed, delaying training. Some trainees had never touched a computer in their life, she also said.

“I knew it was going to be nothing but chaos,” she said.

Richardson, like other crew leaders, used a handheld computer to do payroll – a daily filing requirement – and to send out regular area assignments to workers in the field.

“I sat here glued to this thing,” Richardson said of a computer that took ever longer to transmit data. Even bathroom breaks did not separate her from the handheld computer.

Phone calls poured in constantly, too, from people with questions. Luckily, she had another cell phone at home, and family used that phone to reach her.

“We were supposed to go out on personal observation,” Richardson said, “but I had nothing but phone calls, phone calls, phone calls.”

Richardson found herself living and breathing the Census.

“You can’t turn in all your hours,” she noted.

Census staff in Topeka “never had any solution to the problem,” she said. “‘Pam, just do the best you can,’” she said she heard.

Richardson knew Stone’s territory lagged behind all the others in the state. The real problem for Stone was transmission, Richardson said, but they kept “threatening her like she was inadequate.”

When Stone resigned, about only 20 percent of western Kansas was finished, but an influx of workers from eastern Kansas arrived and the job was “magically” done in a little over a week, Richardson observed.

“In my mind, it’s impossible that they did western Kansas,” she said.

‘All the nightmares’

The handheld computers achieved notoriety before they were ever placed in the hands of Census workers.

The Florida-based Harris Corp. won a $600 million contract in 2006 to supply the small computers to the 2010 Census. In a dress rehearsal in 2007, however, workers experienced trouble transmitting data. Also, the computers would freeze.

In 2008, the U.S. Government Accountability Office put the 2010 U.S. Census on the “high-risk” list. The U.S. Department of Commerce scrapped plans to use the computers for the actual census, but deemed it too late to replace the computers with paper for the 2009 address check.

Lyle VanNahmen, Spearville, regarded working on the Census as akin to “a patriotic duty.” As an auditor on Stone’s crew, he said the handheld computer was “pretty slick” in the field.

However, VanNahmen spoke of the transmission problems and the lack of a paper trail. He thought Stone should have been supplied a laptop computer, instead of the more limited handheld computer, for her role.

In VanNahmen’s opinion, western Kansas “was set up to fail.”

“It chaps me,” VanNahmen said, that workers brought in to finish up the canvassing received overtime pay. And he calls it “bizarre” that the job wrapped up in slightly over a week.

Howard Mead, a crew leader responsible for areas in Sedgwick and Sumner counties, said that territory managed to complete the job ahead of schedule.

“Western Kansas had all the nightmares,” Mead said.

State Rep. Melvin Neufeld, R-Ingalls, a former Kansas Speaker of the House whose district is in the western Census district, said concerns about this Census had reached him. He also said he knew firsthand about the challenge posed there, where some physical locations and mailing addresses aren’t the same.

Pressure all around

The address canvassing phase is essentially complete, said Nancee Torkelson, the local Census office manager in Topeka.

“I think we’re probably good,” she said.

As for the status of Clark County, Torkelson said anything that was not done correctly was redone by the quality assurance team.

“For the most part, western Kansas absolutely stepped up to the plate and did most of their own,” Torkelson said.

Asked about transmission challenges, Torkelson said there were “a few areas” that did not have good coverage, but overall, the handheld computers performed better than expected.

“I think they can improve upon the handhelds, but I think they’re on the right track,” she said.

At the mention of Census workers who were not happy with the operation, Torkelson said, “I have not had any feedback on that through this office.”

“I had not had anyone call me,” she said.

Dennis Johnson, regional director of the U.S. Census, said the unit of work was houses, so the projected workload for a territory was based on those numbers. Worker pay varied, he said, because the Census looked at Labor Department wage rates for areas.

“We don’t do that helter-skelter,” he said.

To accommodate areas outside the Sprint network, the handheld computers offered the phone jack functionality, he noted.

There’s always pressure to make sure the work gets done because there is a “very firm” deadline for the Census, Johnson said.

“If they felt pressure, I can believe that,” he said.

“We anticipated challenges,” Johnson said, but for the most part, he was “very pleased” with the work to date.

Still open

Stone and VanNahmen talked hesitantly about their Census experiences. Even though they were dissatisfied by the recent phase, they find the Census an appealing idea and do not want to close doors. They both might take advantage of the chance to work on the final phases of the 2010 Census, if given the opportunity.

It involves meeting people, using problem-solving skills and working independently – an attractive combination, in Stone’s opinion.

Wedel, too, sounded interested in the possibility. She was surprised on the afternoon that she talked to The News to receive in the mail a letter from the Census saying, in effect, that she remained on active status with the Census.

But Richardson won’t be working anymore on this Census or future counts.

“I’m done for life,” she said.

Tags: , , , , , ,

10 Responses to “Workers: Problems Cloud W. Kansas 2010 Census Effort”

  1. ex-ELCO employee Says:

    1. Our ELCO had a 5% defective rate on the HHCs. (Some were found to be defective during initial setup, before they were in the field.) If an ELCO was replacing 40% of the HHCs, they were not providing proper technical support.

    2. Sprint coverage was a huge problem in our area as well. It’s a shame that the Census couldn’t contract with multiple carriers to ease some of the coverage issues, but after a week of whining, workers learned how to get around it. Frankly it probably cost the Census a good chunk of money since some workers had to drive miles either to land lines or areas with Sprint reception.

    3. There is no way that it’s possible to “fry the circuits” on an HHC because there were too many workers. That makes absolutely no sense.

    4. The pay tables are pre-calculated. I don’t believe there is any flexibility as they are based on the government calculated COLAs. It sucks, but this is something that has to be adjusted from the top down. For example, because of this situation we had a FOS who lived outside the city who was paid less than the listers in the same FOS who lived in the city.

    5. The pressure put on the field workers did seem ridiculous, but looking back, I believe that’s just a symptom. I believe the budget was blown to hell even before the operation started, and pushing to finish as quickly as possible was just an attempt to recover as much as possible.

    6. In my experience working in the ELCO, I encountered more than a few field workers who thought they knew better and should be running things. After certain FOSs and CLs in districts with lagging productivity were replaced, the productivity went up quickly. No one wants to consider the fact that they’re not a good supervisor, but if the surrounding districts are performing okay, it might be the case.

    The big problem I have with blaming the HHCs is that Address Canvassing in 2000 actually DID take the amount of time workers were being told they’d be hired for. The reason the operation finished early and the estimates were off is because of the HHC. So yes, it wasn’t perfect, but it was clearly an improvement over paper.

  2. Anonymous CL Says:

    Good points, ex-ELCO employee!

  3. Anon Says:

    I see some valid points, but a lot of this seems silly. Saying how awful the HHCs were and then stating that the people had no previous computer experience at all. Could be a lot of user error going on there. The pay thing, while discouraging, is something that is set by cost of living in the areas. They must have been working an area where the cost of living is pretty low.

    Fry the circuits? I can’t believe something like that got into print. Come on.

    Lack of Sprint coverage could be a problem and could contribute to fewer transmissions, but it sounds more like people had problem solving skills. Really, it sounds like the team may have not been completely up to the task.

    Some valid issues to me are the fact that this seems to happen each time – too many workers hired and then the pressure to get the job done WAY ahead of schedule.
    I wonder if it really is all about the Regional Directors having bets with each other.
    They find the budget to bring in lots of people from other areas, sometimes at a huge expense, when the stated schedule could still be met. It sounds like that may have happened here too.

    I believe they could have been under a lot of pressure. It’s been pretty clear many areas dealt with that. My question is if it is all about the alleged bets going on with the higher ups. If not, how are those rumors getting started?

  4. Anon Says:

    The fact that this blog takes this article SO seriously is just about all that needs to be said about this blog.

  5. QC CL Says:

    I was told there was a limit of how many enumerators could be assigned to a crew leader — 21 or 22, something like that. If a crew leader was assigned too many people, it could have perhaps locked up the CL’s HHC. It may have required an SD card reset to get the HHC working again. That could be the source of the “fry the circuits” remark. Just guessing of course.

    The lack of Sprint coverage was a huge problem — My people sometimes had to drive 20 miles or more one way to get a wireless connection. The phone connection was often unreliable, which may have been due to the wiring in the person’s house more so than the HHC. The counties where there was widespread Sprint coverage got finished weeks ahead of the others because I could keep my people supplied with AAs. Still, my area was done several weeks ahead and we were sent to adjacent districts to help out (not across state lines or anything like that).

    As people got used to using the HHCs the rate of problems dropped off, and the software updates helped. I always wondered if the tech people knew how to fix the problem with an update, why they didn’t fix the software problems earlier.

    Some of the ELCO staff made it seem like all the HHC problems were caused by the users. A fair amount were user problems to be sure, but the lack of signals, spinning beach balls, need for resets, etc. were most certainly design/implementation problems.

    One of my districts included a portion of a county in a different LCO that hadn’t opened yet. I was told the pay rate for enumerators would be $3 per hour more than what my people made, and to hold back on sending my people into the area while they figure out what to do. I did not have any people from that county in my group. As that district fell behind, I started sending people there and nobody said anything to the contrary.

  6. Stephen Robert Morse Says:

    QC CL – Thanks so much for your comments. Your words refute the sentiments of the naysayers who use this board only to tell me that the information on here is wrong, when in reality, since you provided an explaination of the potential origin of the “fry the circuits” remark, those who doubt my words and the words of others on here might learn a thing or two.

    Stephen

  7. QC CL Says:

    It is interesting that people seem to think you wrote the Hutchinson News article and are responsible for every word.

    If the person who made the comment wasn’t very familiar with computers — and keep in mind NO one was familiar with HHC — then they might put it in those terms. It’s like when my mom says, “my hard drive crashed” when what happened was Window froze and she had to reboot.

    With the HHC there was no way to know what was going on with it, although there were cryptic error codes that I don’t think were useful even to the help desk.

    Stephen, I do appreciate your efforts on this blog. There were/are enough real problems with the census that we don’t need to manufacture straw men.

    I think census worker on the address canvassing operation, myself included, lost sight of the fact that this was not a normal job. It had one goal, and there were basically no chances to get a raise, earn a promotion or anything else you might do in a normal job. So no, the RCC wasn’t really looking for new ideas or suggestions on how to fix the software from the field.

  8. Ms Placed Democrat » Census Bureau Shows Cracks In Address Check In Western Kansas! Says:

    [...] Hat Tip goes to MyTwoCensus. [...]

  9. ex-ELCO employee Says:

    Stephen, I wasn’t criticizing you, just the article.

    And what QC CL said is a good point. We had a couple of CLs near the end who had 25+ enumerators, and their HHCs ran excessively slow. (It took 20 minutes to assign an AA.) The HHCs seem to have been designed to run to extremely tight specifications (I assume to keep manufacturing costs down), and exceeding these specifications (too many HUs, too many listers) caused the HHC to run excessively slow because of the tiny on-board memory.

    Trust me, I’m not defending the Census. But the problems come from the top down, and that’s why something like this site is needed… to find out the source of the problems. Almost any lister, CL, or ELCO employee you talk to will be able to give you a list of at least 5 or 10 problems they encountered. What’s important to find out is what caused these problems.

  10. anon Says:

    I am a Census HQ worker who observed Address Canvassing operations at Trenton, NJ, ELCO in late May. People there had plenty of complaints of various kinds, but no one had ANY complaints about the HHC. I was actually shocked at how well the damned thing worked, since I had tested earlier versions and not been very impressed. I should add that I am no fan of the Harris Corp. or the general approach of Census management to this contract. But somehow it all really did come together.

    Seriously, guys, if the HHC was really performing so horribly, why did Address Canvassing finish so early?