The departure of Martha Farnsworth Riche as Director of the U.S. Bureau of the Census occurred in the midst of her dispute with Republican Congressional leaders over the use of sampling. On January 28, a few days before she was to leave office, I met with Marty to get her personal take on her experiences and her plans for the future. I am pleased to share with readers her candid observations and thoughtful reflections.
Marty began her career in 1961 as an economist with the Bureau of Labor Statistics. She spent the next 15 years working to add meaning to the data that the BLS produces, and to make the agency's information both better known and more accessible to the public. She took her communications skills to the private sector in 1978 as a founding editor of American Demographics magazine, the nation's first magazine devoted to interpreting demographic and economic data for corporate and public executives. In 1991, she became Director of Policy Studies for the Population Reference Bureau. At PRB, she focused on building a bridge between the members of the professional statistical community, the academic community, and the federal policy-making community. Dr. Riche holds a BA and an MA (with honors) in economics from the University of Michigan, and a PhD in literature and linguistics from Georgetown.
You have said that your main objective as Director "was to position the Bureau as the nation's pre-eminent collector of demographic and economic data for the Information Age and the 21st Century." As such, you presented a plan for Census 2000 that includes scientific sampling--a plan that Republicans continue to oppose. What is your take on the conflict between science and politics?
From the point of view of a demographer, I'm interested in high quality data. I think of all of our [Census Bureau] data programs as somewhat akin to cars driving down the street. Most of our "cars" are high quality--they're Tauruses and Saturns and Camrys and BMWs. But we've got this Model T that's struggling along and that's the census. A friend said to me recently, "I don't understand all the fuss over sampling. I remember as a student in the 1950s being told that everybody uses sampling to produce statistics because it's both more accurate and less costly. Only the census doesn't use sampling and that's because Congress won't allow it." And she was astonished that we were still, 40 years later, fighting the same battle. So I think the conflict is where science meets belief systems. It's always a difficult battle. And I would say that we [the profession] have moved the belief system forward considerably in the current battle, but the battle is still being waged. At a minimum, I would like people's expectations to become more realistic. If the Congress prohibits the census from using statistical sampling, I do not want any professional to allow anyone to expect that the Census will count 100% of the people, because it won't.
What can researchers do to make science clearer to policymakers?
We have to rectify our failure to communicate. We're accustomed to talking to one another, not to policymakers. So when we talk to policymakers, we talk to them in the same way as we talk to one other but they hear something quite different than our colleagues hear. That's not communication.
There are two things that members of the research community can do. The first is to be mindful of the differences between a research audience and a broader audience. For instance, three years ago at PAA Margo Anderson presented a paper that she coauthored with Steve Feinberg about the 1990 census adjustment debate. They concluded, "The alignment of statistical experts on both sides of the dispute ironically had theeffect of vitiating the statistical arguments..." In other words, we removed ourselves from the debate by not understanding that it was not about questions of statistical methods but was about questions of law. The result? The lawyers and policy makers concluded that sampling was not a proper census methodology. That's not what any of the statisticians were arguing--they were just arguing about the best methods for sampling--but that's how they were heard.
When you go to Congress, don't give an academic paper but understand what is at stake and what they are going to hear. People who testified when the [census] plan first came out thought they could use their testimony to get us to change the plan in some direction that they cared about. But their criticism was heard by Congress to be saying the Census Bureau's innovations are bad so let's not change at all. They were seen as not supporting our plan when what they were trying to do was to improve it. To be effective, the research community has to understand what is on the table, who is in the room, and how they will use what you say. Anne Harrison Clark is very knowledgeable about these distinctions and has, I know, communicated them very ably to the Public Affairs Committee.
Who exactly went to Congress?
Some PAA members asked to testify on behalf of the research community. We offered to brief them about the hearing but they went up there worrying about particular aspects of the census plan, when Congress was really looking for a yea/nay vote. So their worries were perceived as rejection of the entire census plan, even though they said good things about parts of it. That was unfortunate, because these individuals' concerns will not be met by our doing the census the old way. So they did an enormous amount of damage to changes that would benefit many, without furthering their own interests effectively.
Can one person single-handedly do that?
Yes, because they stand, in Congress' view, for the whole research community. Members do not make a distinction between a witness who represents an entire group, like a NAS panel, and one who just represents him or herself. It's easy but ineffective to propose modifications that cost more since critics don't have responsibility for keeping costs down. But the NAS panel had that responsibility, and fulfilled it.
The other thing we have to do is to understand what people do want to hear. Congress does not want to hear that we are being more scientific. When we give them a scientific answer, they think we are rude and arrogant, or disrespectful and arrogant. They want to hear an answer to their question, they want to hear something that addresses their concerns in terms they understand. That means we have to understand them and understand what they're looking for, which usually means we have to find a new way to talk about what we are doing.
Who should take the lead on this?
I would say that any principal investigator or research grant-maker or research program director--whoever has accountability for a project or program--has got to have a communication plan. They don't have to do it themselves, but they have to make sure that someone does it. It is very important to have the goals and results of your research understood. If you do not manage the perception of what you're doing, you run the risk of being misunderstood or at best, simply not being understood.
What are your feelings about leaving Census before this issue is resolved? Were you trying to send a signal, to force the issue on sampling by leaving at this juncture?
No, I have done what I came to do. I got out the list of goals that I set myself when I went to meet with Secretary Brown and I have done all of them. The only one where I would not give myself a good mark would be educating the public, the media, and the Congress on the benefits of "estimation." I give myself 2 out of 3 on that. The public understands, the media understands, and half of Congress understands (the other half doesn't want to). That is as far as I can go. So I've done everything that was on my agenda. And I'm longing to get back to being a demographer. It has been very frustrating to be here at the most important source of demographic information, and not have the time to even read the press releases. So I'm looking forward to learning how to use census data on the internet, and going back to my real love which is reading and interpreting demographic data so the public can make choices that will help them in the future.
Under your tenure, the Bureau was moving toward the development of data delivery systems that would revolutionize our ability to gain access to and analyze information. For example, the Data Access Dissemination System (DADS) that would deliver the Census 2000 data online to researchers and students. What do you think will happen to these efforts now that you're leaving?
I think that train has left the station. The public is very excited about the idea of having data that are accessible to everyone, not just professionals. DADS will really allow us to meet the needs of people who have only an occasional question about census data down to people who live and breathe the data and that is what is so exciting. With publications we never could do that. Publications were very unwieldy because we had to guess in advance what someone was going to want to know. We can only print the tip of the iceberg. I used to get frustrated if the Census Bureau had decided on an age break that wasn't appropriate for what I was looking at. So as a demographer I've been pushing very hard to get the data out to let the user, whether the person is a researcher or an interested citizen or a kid working on a term paper, to get the data they need, not what the Census Bureau.
And you're referring to individual level records in some form?
No, we have to maintain confidentiality. But whether it's summarized data or a sample of individual records, people will be able to ask the data questions and not have to take it predigested. To the extent that we can recover our costs, we want to make more data available to more people. That's been my overriding concern.
If you had it to do over, would you have done anything differently in your tenure here?
I think I would have gotten coaching earlier. Any demographer suddenly thrust into a senior executive position is required to develop a lot of new skills. Washington is full of very very good people who can coach you in a variety of areas. They have been my salvation. I've used a coach, for instance, for testimony and one for the media. I used a strategic planning consultant and a team-building consultant. I think I would have added to that roster a management consultant or a management coach because I suddenly went from PRB with an overall staff of 35 and an [immediate] staff of only a two --I've always kept my staff small; I like to be lean and move fast--to a staff of 10,000 and a budget to match. I've done a massive amount of reading on how to move large organizations. I think [I would have benefited by] somebody who could have coached me in managing within a bureaucracy, managing both up and down within a bureaucracy. The Census Bureau director, and it's probably true of directors of other statistical agencies, is really where science meets policy. And I realized just the week before I took this job how awkward this position is. I've got this unmovable science and this unmovable policy/politics and I've got to find a way to make them join and it's very uncomfortable. And I had to learn how to do the political part of it.
People who've known you have described you as dynamic, interesting and innovative. On the other hand, the Census Bureau is perceived by many to be rigid, bureaucratic, and slow-moving. How did you enjoy working at the Census Bureau?
I enjoyed making the Census Bureau work to my pace.
Do you think it has?
It has. But what it took, of course, was finding the people willing to do that, inside the Bureau or outside and bringing them in. An example of what the Bureau can do was last summer when Congress gave us only 30 days to write a report that we all thought would take 90 days. I had told the Under Secretary that we needed 60 days minimum. And we only got 30 days. We got it done and it was good. That was a demonstration that the Bureau can work fast. The problem is that it's disheartening when the Bureau moves fast and then the next stage of the bureaucracy works at the old pace.
Commerce Secretary Daley has characterized you as "an extremely dedicated, committed and effective public servant." What is your view about public service? Do you have anything to say to demographers interested in working for the government?
I think everybody should do it at some time in their career. It is an extraordinary learning experience. We can all complain about the bureaucracy, the oversight, the obstacles to getting what we conceive as our "work" done, but it is very helpful to learn why those obstacles are there. Overcoming them is part of the public servant's work too, because the public has other requirements for our work besides science. So it is both humbling and enriching to have a better understanding of the context in which our work is demanded, critiqued or received.
What are your future plans?
I am giving myself a nine-month sabbatical. I call it the Urban Farnsworth sabbatical after my father who died last year. I'm going to use his money to fund it, and I know he would approve of it. In other words, I'm taking a piece of my retirement right now to reflect and regenerate before I go back to work, because this has been a very wearing experience. I want to take time to do my favorite things. I want to garden. I want to go to horse races. I want to lecture, and write about demographic issues and demographic trends.
And less about your experiences at the Census Bureau?
No, only to generalize them in a way that might be useful to the profession. While I've been here I've had a chance to think about the fundamental meaning of the activities that we carry on at the Census Bureau. That will add a new perspective to my work as a demographer. And during the course of doing my favorite things, I expect to find out what I want to do next.
Do you see yourself after the nine months with another "real" job?
I need another real job-I can't afford to retire yet. I look forward to taking another leadership position but I really look forward right now to doing my favorite things.
Will you be at PAA in Chicago?
Yes, I'm looking forward to going to sessions.