The Census Bureau has put its workers on notice about cheating. And they could be fired if caught fabricating surveys.
As part of the crackdown, field employees who gather information from the public will have to sign a “policy agreement on maintaining data integrity” once a year.
These harsh conditions were disclosed in the agreement attached to an Oct. 1 memo — both of which were obtained by The Post — from William Hatcher, associate director of field operations. The memo itself didn’t get to workers until mid-November, according to sources.
In the memo, Hatcher tells workers that Census wanted to maintain its “unblemished reputation for producing quality data.”
If Hatcher really believes that, it’s time for him to retire and let a worker in the field take over, someone who knows this view is an outright fantasy.
Census’ record is about as blemished as the face of a 19-year-old with a sugar addiction. And a guy in Hatcher’s position either should know about it or he’s got his head in the sand.
As readers of my column know, whistleblowers in five of the six Census regions in the US have alleged data were being falsified on a regular basis. Congressional investigators know the names of these whistleblowers and are probing the matter.
Hatcher wrote: “It is essential that we continue to maintain the highest standard of data quality during field-data collection so that the public has complete confidence in the products we provide, and our sponsors, (my emphasis) who fund our surveys, continue to feel they can rely on us for their data needs.”
The second bullet point of the policy agreement said explicitly that workers should “never fabricate information or record answers not provided by the respondent.”
In case any employee missed the significance of all this, Census gave them a pre-paid FedEx envelope in which to return the signed agreement. Pre-paid! Now we know Census means business when it is paying for postage for a clueless policy.
If Census isn’t allowed to cheat anymore, then maybe someone can explain the miracle that occurred while producing data for Friday’s unemployment report for November.
Census collects data for a number of government agencies including the Labor Department. They are the “sponsors” that pay for the surveys and that Hatcher referred to in his memo.
If Census were a company, these sponsors would be referred to as “customers.” And like any company, Census is always worried that customers won’t like the job it is doing and go elsewhere.
While there is no company that can go door-to-door for data collection like Census does, there are new computer technologies that could make current data-collection techniques obsolete.
That’s what Census is worried about.
A source who has proven to be very reliable in the past tells me that all six regions had a large number of unfinished interviews for the unemployment survey just a day before the deadline — and yet a huge number were finished.
So many were completed, in fact, that Census came very close to the 90 percent success rate demanded by the Labor Department. Just a day earlier, Census had completed only 84.51 percent of the surveys needed.
By Nov. 24, New York still had a whopping 1,132 interviews to complete ahead of the Nov. 25 deadline. Chicago had 889 incomplete; Atlanta 779; Denver 1,365; Los Angeles 864; and Philadelphia 1,124.
In all, 6,153 interviews were unfinished nationwide, according to the source.
Normally, when that many interviews go down to the wire, I’m told that many of them end up as what Census calls Type A’s — or non-interviews. Type A’s don’t count toward the 90 percent goal that Census is trying to reach.
So how did the 6,153 uncompleted interviews suddenly add up to the 90 percent goal?
“It’s a miracle,” quipped the source.
If Hatcher wants to keep his “unblemished” record, maybe he should pass some of those policy agreements to the bosses.
By the way, experts forecast the Labor Department will report that the unemployment rate held steady at 5.8 percent in December. Whatever number comes out, just smile because it’s really a joke.
Experts also believe 230,000 new jobs were created last Monday. That figure is not calculated by the Census.
Now everyone wants to get into the act. The HousingWire (housingwire.com) ran a Nov. 26 with the headline: “The Census Bureau is cooking the new home-sales numbers.”
Read it yourself, but the scam is this: Census keeps showing gains in monthly housing sales only because it is in the habit of revising downward the previous month, against which the growth is measured.
If the previous-month revisions weren’t made, there would be no growth. And, yes, that’s all part of the Census Bureau’s “unblemished” record.
Posts
No comments:
Post a Comment