3 Questions for U.S. Census Director Robert Groves
On Tuesday’s NewsHour, Judy Woodruff discusses the new numbers with Census Bureau Director Robert Groves. Beforehand, he spoke with the Rundown about some of the lessons from this year’s Census and how the next one might take shape.
Obviously, much has changed about society as the information age evolves. So what did we learn about the Census process this year that will help next time?
ROBERT GROVES, U.S. Census Bureau director: I think we learned a lot. First of all, the American public this year stepped up in a way we didn’t expect. They returned their questionnaires at a rate that none of the experts, including me, thought they would. This is a good thing.
We also learned that different ways of measuring fit different populations. And so thinking forward, we have to find ways to continually adapt to the diverse lifestyles in the country. So looking forward, we need an Internet option for the Census — as many countries have already. And we’ll do that, because some of the folks who didn’t return the paper questionnaire could return an Internet questionnaire. And when we tried to contact them we had a lot of difficulty, they weren’t home many hours of the day.
I think we also learned that language is an important attribute of doing a good Census. As people have said, the L.A. school system has kids in it that speak over 130 different languages. We did for this Census an advertising campaign in 28 languages, and we had 57 languages covered in terms of language assistance. That will continue in the future. We have to have methods of measuring the population that reflect the fact that we are a very diverse country in terms of language.
And then I think we, the Census Bureau, needs to continue to find ways to do this more economically. We spent a lot of money on this Census. This is a Census that’s very expensive relative to other country’s costs. I’m concerned about that. I want to — especially given these tough economic times — find ways to become more cost-efficient in the future. And that will probably be along the themes of adapting methods to different lifestyles.
The population has grown, but the participation rate stayed around 74 percent. Why was that?
ROBERT GROVES: Well that was a win from my perspective. It’s not obvious — you compare the two numbers and say, “Gee, you were only able to do as well as you did last time.” But as a survey methodologist, I know that every survey that we do has been experiencing declines in participation rates throughout the decade. This is true in Western Europe as well. So there’s an overall secular decline in the proportion of people who respond to surveys and censuses. The fact that we got the same level that we did in 2000 is like a gigantic win. Most of my colleagues would have thought that was impossible.
That’s the result, I think of a couple of things. One this was a short-form only Census. In prior years, about a sixth of the households got a longer questionnaire. That didn’t happen. That longer questionnaire generally generates lower participation rates. So that was a good thing.
Secondly, we had a bilingual questionnaire that was sent disproportionately to Spanish-speaking areas. That helped.
Thirdly, we learned something from survey. We stole an idea from survey methodology that has been used for over 40 years. And that is, a lot of us don’t respond to questionnaires not because we are refusing — we get the questionnaire, we open it, we look at it, we say “Oh, I’ll do that over the weekend.” You put it on your desk and within a couple of days, it’s buried under a bunch of bills and you forget about it. I do this all the time.
A replacement questionnaire sent to those kind of people usually works to spur their participation. We did that and it worked exactly as it should have worked, we saved tons of money, because people returned that questionnaire at levels we didn’t think they would.
You mentioned that an Internet Census will eventually happen. How would that take shape?
ROBERT GROVES: This is happening in many countries and there are some surprising results, first of all. The most surprising result is by adding the Internet as an option along with other options of responding you don’t increase the overall participation rate that dramatically. What tends to happen people who would have after some prompting returned a paper questionnaire they will do it online. They will prefer to do it online, but it doesn’t get rid of the problem of having to follow up some people who do not participate. That is the first thing.
The method that is often done is to use the mail first to send out a letter that has basically a PIN number for you to go online enter that PIN number, enter your data that way. The Internet software design and hardware design has to take into account that these data must be kept secure and confidential. So security is a big issue in the design of an Internet study as is true of the other modes as well, but it is a new kind of security.
At the end of this process all of us around the world involved in censuses and surveys anticipate that the proportion of people who prefer an Internet option will just rise year by year and this will be a very natural thing for us to do. The final thing to note about an Internet design is that we have a test Internet Census right now we have been playing with. So committing to an Internet design is something very easy for me to do as director of the Census Bureau. Knowing what the Internet of 2020 will look like, however, is really hard and I know that if we used our Internet questionnaire from 2010 in 2020 people will say, “Oh, that is so 2010 of you.” So we have to stay nimble in designing Internet software and hardware solutions to make sure we are using all of the power of the internet of 2018 or 2019 when we actually do it.
Photo by Travis Daub. Lea Winerman and Mike Melia contributed to this post.