UCANEWS

P.O. BOX 69626

Kwun Tong, Kowloon

Hong Kong

20 April 1987

XXXXXXXXXXXXXX

Psychology Department

Ateneo de Manila University

Box 154

Manila

Dear XXXXXX:

Congratulations on your assistantship at the University of Chicago and the $8,000 stipend, which, while apparently generous by Manila standards, is probably just adequate for Chicago.

Which is why my first piece of advice to you is: learn how to cook! There's nothing that will save you more money than cooking your own food rather than going to restaurants.

As for advice on how to prepare for a data analysis job, let me say that much will depend on the kind of research people at the University of Chicago are doing.

1. If it is simply a matter of getting used to working with microcomputers and mainframes, working on our department's microcomputer (even if it doesn't have a hard disk) and on UP's mainframe, should be helpful. I presume, of course, that Chicago uses IBM micros. If it uses the MacIntosh, you will have to run to Padre Joaquin.

2. But what is perhaps more important is getting used to the software you'll be working with . . . and here, I fear you may run into some problems.

a. You will probably have to work with packages like SPSS, BMDP, and SAS, each of which has its own set of operating principles . . . something akin to a grammar and syntax. I presume you have done some work with SPSS, but I doubt whether you have done anything with SAS or BMDP. You can work with SAS at UP. For BMDP, I suppose you will have to go to Los Banos (and even there, I understand that hardly anyone uses BMDP, though it is available).

b. But an even more vexing problem is the fact that many of the programs you will most probably work with at Chicago are not within easy access to you now. The SPSS program at Diliman may be "easily accessible," but alas, it is an old version. It does have factor analysis, discriminant analysis, and canonical correlation, but it does not have anything on multidimensional scaling and multiple analysis of variance . . . even the program for the repeated-measures anova is quite limited, as is that for multiple regression which gives you only the forward option on step-wise regression.

c. I am, of course, assuming that much of the data analysis being done by Chicago's psychology department is more or less traditional stuff dealing with experimental and quasi-experimental designs. But if they are doing non-experimental (survey) work, you may have to be familiar with log-linear models for nominal data . . . and for this kind of analysis, you would need the later versions of SPSS, SAS, or BMDP or some other special programs, e.g., ECTA.

d. If, on the other hand, Chicago's psych department is doing work on the simulation of cognitive processes, we are talking of a totally different set of computer programs which you will have to master there.

I would suggest asking Chicago exactly what kind of data analysis is envisioned. If you are expected to do analyses that you have never done in the past, then do some reading.

The market research methodology texts available in Manila should give you a good introduction to many of these data analysis models. You might want to ask Leni de Jesus to lend you the book on multivariate analysis which I gave her before I left. This book shows how different multivariate analyses may be run on SPSS and BMDP (I forget now whether SAS is included).

Then, there's Berenson, M. L., Levine, D. M., and Goldstein, M., Intermediate Statistical Methods and Applications (Englewood Cliffs: Prentice-Hall, 1983) which should be available in the Rizal Library. I say "should be," because I checked it out so I could have a xerox copy made of the entire book . . . but other people I have recommended this book to haven't been able to find it. This book, a more advanced statistics text, assumes students will use SPSS, BMDP, or SAS to run their data.

3. Inquire whether it is expected (or desirable) that you have your own personal micocomputer. If so, you may have to consider the option of buying one in Chicago or getting a clone in Manila (where it would much cheaper). In the latter case, you would have to be inquire about the likelihood (remote, I believe) of getting into trouble with customs. Also in the latter case, you may want to bring certain computer programs with you, e.g., Wordstar, rather than buying or duplicating them over there.

4. Finally, don't forget that part of our job will include preparing or editing data for computer analysis. Hence, you should be familiar with DOS operations and some IBM utility programs. It may also be desirable to have some knowledge of Lotus 1-2-3 or Symphony since you may want to use these programs simply as a device for inputting data (here, I must confess I am extrapolating from my own experience with DBASE III) or for producing graphics.

Please don't think I'm trying to frighten you . . . it's just that your question is so open-ended that I have to answer in the same way. It could well be that initially, all that is expected of you is to be able to enter data and to run more or less simple analyses, with the expectation that you pick up the more complex stuff later on.

Finally . . . and purely out of a deep academic interest in alternative ways of reducing cognitive inconsistency, how many copies did Padre Joaquin's newest vice-president buy of the Asia magazine issue in which she was quoted but at the same time misidentified as [snip]