Hector Maletta: Yes. The Compaq was pretty heavy, and had a red and white plasma screen. Next I bought a black-and-white Toshiba laptop.

King Douglas: What SPSS modules to you use on a regular basis?

Hector Maletta: I use the Base module, of course, plus Tables, Regression Models, Advanced Statistics and Trends…they keep changing the names. Sometimes I also use Amos for multivariate models.

King Douglas: Do you consider yourself a specialist in any area of SPSS programming?

Hector Maletta: Mostly in data management. Before I do any statistical analysis I very often have to restructure the data files. I use a number of SPSS procedures to open and organize multi-level files, to combine and split files and to redefine the units of analysis using Aggregate or Merge Files or whatever.

I don’t do primary research normally because I am a consultant for the United Nations. They send me to some country and I’m faced with a survey already done and files already organized in a certain fashion. Normally, they are a mess, so I have to reorganize the entire database before starting work. The principles around which these files are organized vary from one place to the next, so I rarely find the same problem twice. You find new problems with every new survey you encounter. So on the spot, in some god-forsaken place, in two days, I have to reorganize an entire data file in order to get some results in time.

Besides, some of my work involves creative applications of statistical procedures. Recently, for instance, I developed a multivariate index of living conditions for which I used Factor Analysis in a novel way. Also recently, and for the same work (a Human Development study on Bolivia), I developed a method based on Cox Regression combined with other procedures to analyze a child’s life chances and risk factors, and used it to classify Bolivia’s yet unborn children (those expected to be born in the next five years) on a number of risks such as dying before being 5, reaching that age with chronic malnutrition, or not attending school.

This was shown to be a powerful tool to identify vulnerable families and at-risk communities for targeting social programs in that country. This application took me into three different population censuses for Bolivia, with several million cases each, something I couldn’t have done a few years before with just a notebook.

King Douglas: Do you write and/or employ SPSS macros in your work?

Hector Maletta: I employ them sometimes, but I don’t develop them because most of my work is not repetitive. I’m facing new tasks all the time.

King Douglas: Do you write and/or employ SPSS scripts in your work?

Hector Maletta: No. In particular, I do not use scripts to reformat SPSS output, because I almost never use SPSS output directly: I’d rather export it into Excel or Word for my reports.

King Douglas: What is your comfort level with statistics?

Hector Maletta: I’m very comfortable with statistics…it’s my work. I can manage rather sophisticated levels. I’m not a mathematical statistician by trade—I’m a sociologist and an economist, and at most an applied statistician. I work with statistics and quantitative methods all the time, and am always looking for ways to exploit existing information in a meaningful way.

King Douglas: What are your thoughts about "Instant Statisticians," as made possible by desktop statistical packages?

Hector Maletta: In fact, the phrase was mine. I wrote a message to the SPSSX mailing list called “Instant Statisticians” complaining about it. I have mixed feelings on this. On the one hand, I think it is wonderful that everyone can use a sophisticated statistical package like SPSS to do complex statistical work even if they have neither a full knowledge of the computing algorithms nor a complete knowledge of the mathematics beneath the procedure. I compare this to driving a car. You don’t need to be a mechanical engineer or a physicist to drive a car…you only have to know the basics. I don’t think you must be a mathematical statistician to apply some procedure like linear regression.