The Data-to-Statistics Chain

Print This Post Print This Post

Written by: Director Robert Groves

Federal statistical agencies face a set of common problems about which I’ve blogged in the past (The Future of Producing Social and Economic Statistical Information, Part I) — declining response rates producing cost inflation, meeting the demand for more timely statistics on smaller and smaller groups, harnessing new technologies, integrating new data sources into traditional survey statistical processes, and doing this all with declining budgets.  I’m convinced that the talent among the Census Bureau staff and its partners is up to the task of transforming the organization to meet future challenges.

However, not all challenges are technical or scientific.  One is ubiquitous to many common-good functions in a society – we take for granted essential facilitators of our day-to-day lives.  We count on electricity to power our electronic devices; we rely on clean water coming from our faucets at home; we assume our telephone calls will go through when we make them; we expect the bridges we drive over will support our cars.

These features of our lives are so central to our daily routines that we have trouble assigning a value to them.  Without them, our lives would be incomparable to our current ones.  What’s it worth to you to have clean water?  If the bridge didn’t exist, what would you be willing to spend to build it?  How central to our lives are these common good features?  Indeed, could we live without them, as a cost-saving strategy?

I have a cousin raised in an urban area who in her youth answered the question of “Where do peas come from?”  She said, “Peas come from the can my mother opens in the kitchen.”  She was correct, of course, but ignorant about the ultimate source of the vegetables, the farmer’s field.

Statistical information is a bit like that.  Every day we see in the paper statistical information purporting to describe our world.  To many people, that paper is the can of peas.  The media provided the information.  The faucet produced the clean water.  The outlet produced the electricity.

One of the key challenges to statistical agencies is to communicate their benefits to the larger society.  However, to fulfill the challenge, they need to describe a long chain of events.  The chain begins with residents of the society, who give freely their answers to surveys and censuses trusting in a pledge of confidentiality.  It ends with them using statistical aggregates of all those answers, to make important personal, community, and national decisions.  The statistical agency starts the whole chain of events, but there are many independent intermediaries.  If you don’t know the chain, you may misjudge the value of different steps in the chain.

When people understand the chain and appreciate the statistical information it produces, better informed judgments are made about the value of agencies to the society.  Clear, plain English, simple explanations of the data-to-statistics chain are a continual challenge.  We all need to do our part.

Note: This will be the last blog I write as Census Bureau director.  I have been honored to serve as director, and I hope future appointments for this position will be made from among scientists who are committed to technical innovations in the production of statistical information.  The Census Bureau must be a scientific organization to fulfill its mission because science is devoted to innovation and continuous improvement.

This entry was posted in About the Agency, Measuring America. Bookmark the permalink.

4 Responses to The Data-to-Statistics Chain

  1. JW says:

    Well if you’ve realized that the public thinks of the data the census collects like the can of peas envisioned by your cousin – then why doesn’t the census bureau advertise the importance of what it does? Any company worth its salt knows that to stay in business advertising is critical and reaching your core audience is the key to its success. It’s interesting – closing 6 offices and turning peoples lives upside down – when a simple cost effective advertising campaign would have been very, very effective. There are even free and low cost advertising alternatives that could easily be implemented to educate the public about the work we do and how if affects them.

  2. Chris Stuber says:

    Quick comment somewhat related to your blog. I would like to see a more efficent, cost effective government. Reduce paper consumption, scan all important stuff online and telecommute to save fuel consumption. Same for schools, which I envision children doing their work from a computer. Get those buses of the streets… its a mess. Gas is $3.50 a gallon so why do we commute to workplace when the tools are there to bring the working wherever you are. Open that laptop and get the job done. The policys need to change with evolving the technology. Getting instant stats, mapping, field work. If every gov worker could a little saving, every little bit help US to be a more responsible nation. Very good blog, lead by example.

  3. Terry says:

    I’ve worked in local census offices for two running censuses. If you are an ethical person who wants to see our culture succeed, I recommend such a job. This article reflects the difficulty of the promotion of a high quality product. Data is extremely cheap today, it is very easy to produce, it can be produced in volume and can be distributed with an inexpensive computer. Finding good and reliable data in a morass of data has never been more difficult. Good data needs to be both easy to understand and difficult to misunderstand. I, for one, appreciate the Census Bureau’s efforts in recent times.

  4. Chris Stuber says:

    I am a former Census employee relocated South. DC Should realize there is Census is about outreach to people where they are. I don’t want to move back, so why wont they allow employee to work as Tellecommuters? I don’t get it, How does govt. intend to get the job done. It’s part of the data chain.

Leave a Reply

Your email address will not be published. Required fields are marked *