Linda J. Young is Chief Mathematical Statistician and Director of Research and Development of USDA's National Agricultural Statistics Service . She oversees efforts to continually improve the methodology underpinning the Agency's collection and dissemination of data on every facet of U.S. agriculture. She works on the surveys designed to characterize agricultural activity in the US.

(Background music plays)

John Bailer : I'd like to welcome you to today's Stats and Short Stories episode. Stats and Short Stories is a partnership between Miami University and the American Statistical Association. Today's guest is Dr. Linda Young, Chief mathematical statistician and director of research and development of U.S.D.A.'s National Agricultural Statistics Service. I'm John Bailer. I'm Chair of the department of statistics at Miami University and I'm joined by my colleague Rosemary Pennington, professor in the department of media, journalism and film. We're delighted to be speaking with Linda on our short episode today. Linda, welcome!

Linda Young : Thank you.

Bailer : Oh! Certainly! One thing that I'm just curious to hear your thoughts on is some of the challenges in doing survey research. There are terms that we hear a lot - the idea of undercoverage, non-response, adjustment to service…to surveys. So if you could just describe in a few sentences, what does it mean when they say that there's an under coverage problem, when trying to conduct a survey? What does it mean to say that there is non-response in a survey and how might that cause a problem in a survey and what other kinds of adjustments might you do?

Young : To conduct a survey, just think, you have to have a list or some other way in which you identify whom you want to survey. One of the primary things we have is a list that we maintain of all farms and potential farms. However, think about it. Some farms go out of business, others come into business. They don't automatically call us up and say hey, I have started a farm. Our list is incomplete and at any point in time; we struggle and work very hard to try to keep it as complete as possible. But it is an unattainable goal, so when we conduct a survey, any farm that is not on our list is in our group of under coverage. We know we don't have them and so we haven't surveyed them. And we want to adjust for that. Now, how much adjustment depends upon what kind of farm. What kind of survey we are conducting. If it is a large farm, we almost always have those. It is the small farmers that we have trouble finding. And so if it's something organics like something local foods, then undercoverage is a bigger issue.

Rosemary Pennington: We talked to the former director of the U.S. Census several months ago and he talked about, you know, the issue of testing questions and ensuring that the survey questions are measuring what they're meant to be measuring and so I wonder given the fact that you're surveying farmers and then you do this survey of farms every five years, sort of what do you guys doing between those two big surveys, sort of test your questionnaire and to make sure you're gathering the data that you want to be gathering?

Young : We have a substantial group that conducts cognitive test of our questions and since the last census we've developed a new set of questions for our demographic section. All of those have been heavily tested other questions on the census have also been tested. Just to do what you mentioned, we want to be sure that they're answering the questions we ask.

Bailer : So talk a little about how cognitive testing of a survey item is conducted?

Young : We identify farmers who are willing to work with us and we send individuals out and they have, basically, an outline of how they are to approach things. They first administer the questions and watch the person as they complete the form and then, they go back and ask questions. So for example they said you know it seemed to take you a while to answer to this question. Was there a problem with it and then they discuss the things that have come up, the concerns they have. Sometimes we will see that the question they answered isn't the one that we were trying to ask, and we try to figure out why that was a problem and then we come back and revise the questions, go back out and test them again and it's an iterative process.

Bailer : So just one last question. I mean, so if I want to prepare for a career at NAS, or if I'm helping my students get ready to work in your organization, what kind of background should they have to do this work?

Young: So we have several different career paths. For a statistician, for statistics it's good for them to have some sample survey experience, to take some classes in that, to take classes in modeling and multivariate. So it's a broad spectrum of statistical tools that we bring to bear within the agency. Another track is more of the cognitive usability testing and so that's more of the psychology-type training, as well as a foundation in some statistics that a person would need to have to do that work.

Bailer : Perfect! Thank you. It's been our pleasure to have Linda Young join us on Stats and Short Stories. Stats and Stories is a partnership between Miami University's Department of Statistics and Media, Journalism and Film and the American Statistical Association. Stay tuned and keep following us on Twitter or Apple podcasts. If you'd like to share your thoughts on our program, send your e-mails to statsandstories@miamioh.edu and be sure to listen for future episodes where we discuss the statistics behind the stories and the stories behind the statistics.

(Background music plays)