 |
Hector Maletta...contd. | |
|
||
|
||
|
||
|
Hector Maletta (contd): On the other hand, SPSS provides many sophisticated tools for statistical analysis, but the results are often too complex for presentation to others who are not so well versed in statistics. On most occasions, I find it necessary or convenient to do whatever analysis I want, then export the results, say, into Excel and prepare an easy-to-understand table by myself. It is often just this simplified output that goes into my report, albeit oftentimes I include the detailed results in a statistical appendix or in an academic paper about the same subject. It all depends on the audience intended, of course. Besides this question about the level of statistical sophistication in the audience, certain analyses deliver too much content in their output, or deliver it in a less than crystal-clear fashion, thus necessitating editing. I seldom use raw SPSS output in most of my written work, except for simple tables, though I know many people do. King Douglas: What makes a problem difficult? Hector Maletta: It varies with the problem. The difficult problems sometimes involve not so much figuring out how to do it in SPSS, but they arise when dealing with a new problem that demands an innovative statistical approach or one I am not familiar with. King Douglas: Many problems can be solved in a variety of ways. Do you value parsimonious code and work to attain it? Or is parsimony irrelevant given that a computer is going to do the work? Hector Maletta: I look for parsimonious code, of course. But if developing the most parsimonious code will take me two days, I prefer, then, a less parsimonious way because it is faster on the whole, though perhaps not faster on the machine, especially when I’m not likely to be facing the same situation twice. King Douglas: Have you ever seen a solution to an SPSS problem that you thought was particularly elegant, poetic or beautiful? Hector Maletta: Many times. I have even congratulated the author many times. There are quite a few people able to do that in our list, many of whom you are interviewing in this project. King Douglas: Say you have creatively solved a particularly difficult problem, with even a flash of brilliance. Whom do you tell? Hector Maletta: Normally I would try to write a paper and publish it, if I see it as brilliant enough to be worth it. Sometimes I mistrust my own brilliance, saying to myself that I may be mistaken. So I consult someone on the SPSS list or with a colleague, “Take a look at this and tell me if I’m an idiot, or what.” I have a network of colleagues in my field who are experts in food security, nutrition, poverty and other things, so I can consult with many people about my work. For statistical issues I usually go to the SPSSX-L list. King Douglas: Describe the way in which you attack a difficult problem. Do you work with pencil and paper first? stare at the wall or ceiling? work directly in the syntax editor? Hector Maletta: First I may go to the corner café and over a cup of coffee write what I need to do, most often on a paper napkin. I write the structure of the problem…not the details. I need an understanding of the whole first, either when writing or reading. When I browse a book, for instance, a technical one I mean, I like to understand the whole book at once, looking at the table of contents or the introduction, before getting into particular chapters or pages. When you are faced with a long syntax file—1,000 lines long—you often don’t see the whole, nor are you able to pinpoint the tricky or brilliant parts. This is mostly because SPSS syntax files are not structured, they are just text files. For that, I usually first prepare a scheme of the whole thing, the main phases to go through. Then I write a title for each section in the form of asterisked and UPPERCASED comments before writing the syntax. |
||
|
||