Dave A Dickey is a SAS Book Author hailing from Cincinnati, Ohio. Along with Professor Wayne Fuller, he developed the Dickey-Fuller test, which has been cited 45,000 times in scientific literature. He has worked at both the NC State University Department of Statistics and the U.S. Census Bureau, and he is a member of the American Statistical Association. His suggestions for upcoming SAS programmers include those that have helped him tremendously, such as learning about program data vector. He also speaks about the value of networking and talking to others in the field, sharing your knowledge and learning from them in return. Finally, he talks about the importance of just playing around with SAS and experimenting, and attending SAS meetings around the world. He also goes into detail about PROC ARIMA, his favorite, as it has the Dickey-Fuller test in it, and how it can be used for mathematical forecasting.

 

Interviewer

Can you please tell us more a bit about your career path and any publications or professional communications?

 

Prof. DAVE A DICKEY

I started out in Ohio I grew up in southwestern Ohio near Cincinnati, and I attended Miami University in Ohio, Miami of Ohio some people call it, and I was a math major, so this was in the 60s the early 60s, so some people will remember that the Vietnam War was raging at that time, and I had a low draft number, and I would have been drafted had I not stayed in school. So, I got a master’s degree in mathematics  at Miami, and then after that they were also giving me gave what was called deferment, so I’ll have to say that war had an influence on my career path, at that point after I graduated with my master’s degree in math I continued on as a teacher for a while I taught at Randolph-Macon College for two years that’s just north of Richmond, and I taught at the College of William and Mary for a couple years;William and  Mary was actually the first one, and they were offering these temporary positions, and at that point by the time that year rolled around which I guess was 1972 the war was winding down, and I had almost aged out of the window where they were taking soldiers. So, I continued my education, and I went to Iowa State University where I was  mentored by advisor Wayne Fuller, Professor fuller, and I worked out something called the Dickey-Fuller test, and that I just looked this morning I that the papers that we wrote out of that, out of that research had been cited 45,000 times in the scientific literature, and it is incredible, and that if you sort of think about it, scientific papers have been written citing at work every day since I graduated on average, so that’s something I’m happy about. I’ve written maybe seventy some papers I guess over my career but the two or three that we wrote at the beginning that’s really where my reputation came from, then I joined the faculty at North Carolina State University, one of my students John Brocklebank he’s done very well at SAS, he asked me in his early career there at SAS Institute would I help write a course on time series we did that, and we taught it together for a while, and then another colleague Herbert Kirk who was at NC State when I came to interview, and then joined the group at SAS, he suggested that we write a book about time series which is exactly what we did, and that was a very nice experience, and as, you know, we are in the process of finishing up the third edition of that one more sort of interesting thing maybe for the people who are listening when I came to interview I met a guy named Jim goodnight in the department, and another guy named John Sall in the department, and so those were in the group the research group that formed SAS, and of course, you know, Dr. Goodnight, you know, John Sall there was another person Tony Bar that I met when I came here to interview, by the time I joined NC State University Department of Statistics, those fellows had moved across the street to a building that they were renting, and they started SAS Institute, and the rest is history they’ve done incredibly well, and here we’re very proud of the folks that came from you I’ve had several students that went out to SAS, and ended up working there, the only other thing I guess I might mention is I had a sabbatical leave one year, and I worked at the US Census Bureau that was a good experience as well, I’ve been teaching, I teach all I only teach graduate students, and I do research I’ve had maybe 16 or 17 Ph.D. students that finished up under my leadership.

I did have a couple of nice things happen to me along the way, I am a fellow of the American Statistical Association which is a nice honor, and at one point I don’t know if it’s still true but I was the most highly cited researcher in our entire University, the library did a study about that. So, it’s been a very nice career for me I recently just last year went on what’s called phased retirement, so I’m a half-time person now, I teach one course in data mining, and I do consulting with our agricultural, and life science leaves here, and kind of winding down toward the end of my career I turned 71 last year, so even growing social security now.

 

Interviewer

Can you propose 3 valuable tips or strategies that are necessary to become a Top SAS Programmer?

 

Prof. DAVE A DICKEY

One thing that helped me tremendously in understanding how this works, and I would advise anyone starting out with SAS, and wanting to do some programming I would advise them to learn about the program data vector, this idea of a program data vector which not everyone is familiar with that it was a I believe it was actually Jim bar in the early days they came up with that program data vector but it enables SAS to be very efficient in terms of data manipulation, and also if you learn about the program data vector you’ll understand much better what they retain, and drop, and keep statements do, and about how datasets are merged and concatenated how they can be read into themselves in a way that creates new variables that weren’t in there to start with. So, I would certainly advise people to get up to speed on the program data vector the other thing or one other thing that I would suggest doing if you are lucky enough to be in an environment like I am where there are other experienced users is to talk to each other, simply talking to each other finding out what people are doing, asking for help when you have problems I’ve learned a tremendous amount from a couple of people here whose job was just doing SAS programming for clients Joyce Smith, and Sandi Donaghy were our two main SAS programmers when I came to work here, and I’ve worked with them for a long time I’ve learned a tremendous amount from them, and then the third thing that I would suggest that people do is to play around with SAS in other words you can learn a lot by just experimenting, and of course a020s a statistician I like experimenting, and I think most people who are who are curious, and are interested in learning things like to experiment around, and see what they can find in the software.

 

Interviewer

I would say that that would be one of the key things to successes experimentation.

 

Prof. DAVE A DICKEY:

Alright one more thing I guess maybe I should mention if you don’t mind is if you’re not in an environment where you have other SAS programmers, and people that know SAS to work with, I would suggest going to the meetings we have regional SAS meetings, we have local SAS meetings, and then, of course, there’s SAS Global Forum, and those are venues where you will meet a lot of other people that are interested in SAS, and typically very enthusiastic about SAS, and I think if you go to some of those talks, and chat with people in your spare time there you’ll learn a lot.

 

Interviewer

Which SAS procedure do you find most useful (do you use most often)?

Prof. DAVE A DICKEY

Well I’ve got to say proc arena it has the dickey-fuller test in it, so there’s there’s not much competition there, and in my mind, and maybe not anybody else’s mind but actually this proc arena is what people use for mathematical forecasting and statistical forecasting. So, I think a lot of people that use SAS, and I’ve been teaching courses for SAS since 1981 I guess, that’s another thing I’m a contract instructor for a process, and often in talking to folks from other classes at lunch or after a break or something like that, often what they want to do his forecasts, he doesn’t want to be able to predict what’s gonna happen in the future. So, I think that proc ARIMA is a very strong procedure, and it’s in my opinion it has, it’s kind of the Cadillac of forecasting procedures it’s been around for a while, there are some new things that are out there that basically automate some things, and make them maybe a little easier to access there’s forecast studio, and some other goodies that are out there, and some new procedures like the proc ESM the unobserved components models those kinds of things but really if you look at them take a little deep dive into what they’re doing what you find is that this kind of ARIMA concepts arena stands for autoregressive integrated moving average. I don’t know why we come up with these names that are so sort of complicated in, and obscure but anyway that’s  arena but for many of these kind of buoys, and things really arena kinds of models are going on under the covers when you look into, so that’s when I like a lot I also like to write to drive a standard shift car where most people would with like an automatic but I like to have that degree of control sometimes with a GUI, you perhaps lose just a little bit that that sort of hands-on control that you have with, maybe some of the other procedures that are not quite as automated.

Interviewer

Right. So, it just makes it all the more interesting.

Prof. DAVE A DICKEY

Yeah I think so, and another thing if I can mention a couple other things that I think our favorite procedures of mine I teach design of experiments, and it’s kind of amazing how many concepts of design of experiments like blocking, and incomplete blocks, and things like that, some of your statistician listeners will will understand, and recognize so many of those things involve exactly what happens in mixed models, so praat mix which by the way was initially written by one of our NC State graduates is another one on my of my favorites, and I think that mixed really was a game changer in terms of statistical analysis especially of experimental data in other words  when you run an experiment like a clinical trial or something like that it often involves a mixture of what we call fixed, and random effects, and the place where you can handle that correctly, and nicely, and rigorously is  indeed mixed, and one more if you have time for one more, I’ve explored a little bit of the SG graphics, and they’re so nice so much better than the graphics that we had before. So,  I like proc SG plot for example as very nice or making plots, and I think that, some of the folks at SAS Bob on Regas being one of them made it an effort also to have graphics associated with most of the procedures the ones that really would logically support graphics, and that’s a tremendous boon I think to the to the researchers you can run Gregg which is one of the basic statistical procedures, and there are lots of diagnostic plots that come out of here other plots that are optional and, you know, when you have an old expression there’s probably a reason that it’s an old expression, and one of the ones that would be relevant here is that expression a picture is worth a thousand words, so  there’s a lot to be gained by making a graph in fact I try to tell my students if you try to analyze data without plotting it I hope you have your statistical malpractice insurance data because that’s exactly what it is to analyze data without bruising just taking a look a visual look at what’s going on.

Interviewer

Any closing words?

Prof. DAVE A DICKEY

Well I’ve enjoyed teaching, and using SAS over the years I think the fact that I use SAS in the classroom has been something that the students have appreciated there is a lot more competition out there now, that certainly saps has a lot more competition than when I first came here,  I do a lot of different things I do research, and then all sometimes extremely large simulations SAS is good for that if, you know, what you’re doing, and it’s never failed me, and I think one of the things that is nice about SAS is it there is a little bit of a steep learning curve to get started but really most of what you want to do is in there. So, you don’t have to use one package for analysis, and another one for graphics, and another one for tables, and another one for reports, so  it’s a nice complete package it’s very well vetted in the sense that they have a Tenniel support department that helps you, and also they have a group that carefully studies what’s going on in quality control group, I guess you could say they pour a lot of money into research and making sense better year after year, and I’m very proud of what they’ve done a lot of our graduates work there, we’re proud of them we’re proud of the people of SAS, and so it’s been a pleasure for me really over the years too to have a chance to teach SAS to work with SAS, to teach for SAS, I teach as I said short courses expires I’ve enjoyed that program got a lot of help from the folks at SAS Institute in writing books, and  that’s been very nice for me, I got a very nice honor from Dr. Goode not a couple years ago it says well before I was selected as the SAS distinguished pr

(Visited 35 times, 1 visits today)
Facebooktwittergoogle_plusredditpinterestlinkedinmailby feather

Subscribe To Our Newsletter

Please Join Us.

You have Successfully Subscribed!