Page 2 of 2

Census, Uncensored

Former US Census director Kenneth Prewitt discusses gay marriage counting, Michele Bachmann's comments, and other mysteries of the 2010 Census.

| Fri Jun. 26, 2009 5:30 PM EDT

MJ: So how would gay households know that they count?

KP: Well, that's something that I can't answer because I haven't seen anything yet. There will be some serious effort by Census Bureau personnel to create an operational plan to make it work. I think they will go about this very seriously to make sure there aren't any errors in the data. They want to get it right. This is who they are and what they do.

MJ: Do you think that it's surprising that here we are, in the year 2009, and we're conducting the 2010 census without using technology for all parts of the operation?

KP: Let me put it as follows. I think the Census Bureau has been a technical innovator, certainly since the start of the 20th century. After all, it was the first agency to use the Hollerith Card, the old punch card which was married to an adding machine which became IBM. It invented sampling theory in the 1930s. It was the first federal agency to use a major mainframe in the 1950s. It was extremely adept in 2000 in doing intelligent character recognition and data capture using very, very high-tech processes. And I think you can say about 2010 that it was technically innovative in using [GPS] handhelds in address canvassing. I don't know the results yet, but I hope we've come out of this with a much better address list than we had in 2000.

MJ: But do these computers from the Harris Corporation even work properly? And were they designed properly?

KP: Well, I'm not in the position to judge that. They're doing quality studies on that right now.

MJ: I don't know if you've read on and other blogs where people have complained about the functionality of these devices.

KP: When you payroll 140,000 people, it's not hard to find people who are disgruntled. And I'm not saying they're wrong. I actually talked to people who did the address canvassing work, people I know extremely well, people who had no reason whatsoever not to tell me what was going on, and they had some problems, but they are convinced that they ended the process by producing a much better address list than when they went into it. And that's the test. Of course there's always somebody saying, "This didn't work" or "That didn't work." But the test is to ask, "Do we have a better address list?" And that's quality control judgment. I'm not trying to slam [], but you have to appreciate that you are getting a self-selected group of correspondents. But I'm not passing judgment on you, and you shouldn't pass judgment on the technology until the data are in. What I'm saying is that I don't think the people who are writing to bloggers are the people producing systematic data.

MJ: I understand that. So what are the greatest obstacles that remain for the 2010 census operation?

KP: The old ones. They haven't changed. One, you've got to start with a good address list. If it's a good one, good. If it's not, then you can't have a good census because that's your frame. And when you send nonresponse follow-up people, there's that. Secondly, you've got to hope for a decent mail-back response rate, because the workload goes way up and the costs go way up if you don't have a good mail-back response rate. And we simply do not yet know what the response rate is because we haven't done it yet. But if it's not in the mid-60s, it's going to be both budgetarily and operationally very difficult for the Census Bureau. And in 2000, we had expected to be in the low 60s, and we got into the mid- to high 60s and that was an enormous boost, and we did it with a first-rate effort. And I think that the people running the advertising campaign right now and all of the outreach are very good professionals, and I am hoping they are successful. But it's up to the American people to do it. What can the Census Bureau do, other than put it in their mailboxes on schedule? And if they [the American people] don't send it back in, they have to start knocking on the door. So there is the mail-back response rate and then there's the willingness to cooperate in nonresponse follow-up, and then there's the startling problem of the enormous number of undocumented in the country, who will have every reason not to want to cooperate with the federal government.

MJ: I know it's early, but what will people say when they look back on the 2010 census?

KP: I hope they say they pulled off a miracle. [Laughs.]

MJ: Will the 2010 census need a miracle?

KP: That's a quip, but I think it's going to be very, very difficult to do this census in the current environment. That has nothing to do with the Census Bureau's skill set. It has to do with the American people's not wanting to be bothered, not answering their mail, not having phone lines anymore because they got cell phones, and the Census Bureau wants to do follow-up on the phone, and they can't find phone numbers because they've only got the landline number. It has to do with the number of mobile people who would have changed houses between the time you did the address listing and when you knocked on the door. Houses will get torn down and houses will be built. The whole foreclosure crisis is a major crisis because whole hunks of the country are empty when they should be functioning neighborhoods. There are just a host of problems. And then there are the ones we can't predict. Who knows? Natural disasters, strikes, I can't tell you what's going to happen. I know it's going to be difficult; it's always difficult to do a serious census, especially with today's economic, political, and general cultural circumstances. Let me ask you a question. Let's say there are 12 million undocumented immigrants in this country. What percentage of those people do you think will mail a questionnaire back in?

MJ: Ten?

KP: Whatever it is, it's a low number. So let us say it's 50 percent. That's a 3 percent undercount before you start. Let us say it's 25 percent, and then you've got an even bigger undercount before you've even started.

MJ: So that's what explains the need for the use of statistical sampling?

KP: That's one of the reasons we were going to use it, but we can't. So that just means that you're doing a census knowing that you are going to miss a very significant proportion of the American population. And you're obligated to count everyone, so it's just an uphill battle.

Stephen Robert Morse is an intern at Mother Jones, and runs the website

Page 2 of 2