A systematization of statistical data management/manipulation tasks and a comparison of statistical data management/manipulation capabilities of SAS and SPSS base programs

摘要

The general purpose of this study was to systematize statistical data management/manipulation (SDMM) tasks and to evaluate the SDMM capabilities of two popular statistical program packages, the Statistical Analysis System (SAS) and the Statistical Package for the Social Sciences (SPSS). The objectives of this study were to: (1) identify commonly used SDMM tasks and construct an SDMM task system; (2) comparatively review and evaluate SDMM capabilities of SAS and SPSS base programs; (3) produce a comparative and concise list of SDMM commands/statements for the two packages; and (4) make suggestions about the improvement of SDMM capabilities of the statistical analysis packages.;The considerations of definition and classification techniques for SDMM systematization were from a logical angle. The SDMM task system was conceptualized and constructed by identifying and defining its individual tasks and by systematizing and explaining the general intrarelations among these tasks within the SDMM system and the interrelations between the SDMM system and the statistical data analysis (SDA) system.;Three basic criteria (i.e., availability, flexibility, and simplicity) were used in the evaluation of the SDMM capabilities of the SAS and SPSS packages. Similarities and differences as well as advantages and disadvantages for each specific capability of SAS and SPSS were addressed.;The results indicated that SDMM can be perceived as a system. The SDMM task system consists of five blocks (i.e., control, input, process, output, and database) and many tasks/subtasks. The objects of these SDMM tasks have three basic levels: variables, records, and files. The SDMM blocks, tasks, and subtasks are interrelated, by having the same general purpose, acting on the same operation object, being hierarchically structured, being interdependently organized, and/or being dynamically executed.;Four items of comparison in a more general setting were found. (1) Similarities in overall SDMM capabilities between the two packages overweigh dissimilarities. (2) Each package has its own distinguishing features. (3) These two packages are not without shortcomings. (4) Advantages and disadvantages are relative concepts, and are also changeable.