Page 148 - DLIS401_METHODOLOGY_OF_RESEARCH_AND_STATISTICAL_TECHNIQUES
P. 148

Unit 10: Data Analysis Using Software Packages




          research, SPSS is a far better tool for some quick and painless analysis. When you need to get  Notes
          some numbers flying around and are doing transformations like a maniac you had better
          know how to do some SAS. In terms of creating easily exportable, attractive output SPSS
          blows SAS away. It is not even close. SAS/Graph is well, not so good. The reason SAS is more
          valuable is because SPSS cannot really handle large databases very well. If I need to manipulate
          thousands and thousands of cases, SAS has the power and reliability to get you there. It also
          exists in a UNIX/mainframe environment. SPSS does not.

          10.3   STATA

          Some features of STATA:

          (a)  Interactive, very fast performance. Analysis can be done iteratively or through programming
               (called ado files). STATA loads the entire data set into RAM (which makes it much faster
               than SPSS in the normal case) which means the performance degrades considerably
               when data set size exceeds available memory. Whether this matters depends on the exact
               size of the data sets, and how much RAM you have, and how effectively you can subset
               the data. You may find that you can extract subsets with programs like stat/transfer
               (cheap and very effective) and avoid the problem.
          (b)  Fast performance requires large memory, would need memory and possibly system upgrade.
               Model could be constrained by lack of memory.
          (c)  Offers “NetCourses”; less expensive training accomplished via E-mail.
          (d)  Some love the analysis tools and support, both from the STATA company and from
               other users on the STATALIST (many of whom are STATA employees who answer
               questions and take input back into the company. Even the president of the company
               participates in their listserv.).
          (e)  STATA does not play as nicely with other programs as SAS does. Also, STATA does not
               have many of the modules available in SAS.
          (f)  STATA can handle extremely large datasets, both in terms of number of cases and variables.
               We can take a table of results from STATA and using the Copy Table command, paste
               results directly into Excel, which we can then very easily format into a Word document.
          (g)  Updates are easy to install, and there are many very good user-written plug-ins available.

          (h)  STATA’s documentation is superior to those available from SAS.
          (i)  STATA is much more competitively priced considering you get the entire package rather
               than getting pieces of it. They even have a deal where students can get a 160 page
               manual and a one year license for a negligible amount of money (Small STATA).

          STATA and SPSS

          STATA has many advantages from a statistics point of view, depending on what you are
          doing. They have excellent support for complex samples (cluster, stratified, weighted) and
          offer lots of statistics you can’t get from SPSS—especially for limited dependent variables.
          SPSS is ahead of STATA in terms of the friendly interface and we think it is easier to do data
          management with SPSS. If you’re looking for a single program to handle big data sets and do
          relatively simple analyses, SPSS is certainly up to the job.








                                           LOVELY PROFESSIONAL UNIVERSITY                                   143
   143   144   145   146   147   148   149   150   151   152   153