SPSS: The 'Old Faithful' of Statistical Packages
March 01, 2017
SPSS: The 'Old Faithful' of Statistical Packages
Score 7 out of 10
Vetted Review
Verified User
Modules Used
- IBM SPSS Statistics
Overall Satisfaction with IBM SPSS
Working at a university, we have certain analytic products that are available to us for free. Last year, we had an open license to SPSS. That changed as of this year and now only one person on our team may have it downloaded, but SPSS has been used multiple ways by us. We use it for internal reliability and validity analysis of survey data. SPSS is fairly easy to use on this front: writing and running code in its Syntax window is straightforward and user-friendly.
In terms of data cleaning (which is needed for our type of analysis) SPSS is not ideal. Handling and cleaning large datasets is better done in Excel and then imported into SPSS. This is because SPSS recognizes a row of data as one entity (or in our case "respondent"). Once the data is imported in, you are constrained to what SPSS can't "see" or understand about your data that other analytic programs might be able to.
SPSS is a great starting point for statistical analysis, and it is used across the university in most basic statistics courses.
In terms of data cleaning (which is needed for our type of analysis) SPSS is not ideal. Handling and cleaning large datasets is better done in Excel and then imported into SPSS. This is because SPSS recognizes a row of data as one entity (or in our case "respondent"). Once the data is imported in, you are constrained to what SPSS can't "see" or understand about your data that other analytic programs might be able to.
SPSS is a great starting point for statistical analysis, and it is used across the university in most basic statistics courses.
- Variable manipulating & recoding is fairly easy with SPSS. Not only does SPSS have a recoding function, you can also just write the syntax for this and save it for future use.
- SPSS is mostly user-friendly and gives budding statisticians and analysts empowerment with its easy to use menus and easy to read messages.
- SPSS outputs are easy to make sense of and follow, no matter the skill level. If using the tables generate to put into another report, it is easy to do so as well. By right-clicking a table, SPSS gives you options of how you would want to copy it and in what format. This is very useful when generating reports with this data, which I often did.
- SPSS does not update frequently, new versions seldom come out. There have not been many changes to the interface (if any) and because of this, you are stuck using older statistical methods of analysis. When searching for user guides and tips online, many have not been updated in years, because of this.
- Data cleaning capabilities could use some work. It is tough to build a data set in SPSS because it is limited to what you import, and if the variables aren't imported correctly, there is a great deal of correcting that needs to be done and this can be time consuming. It adds a number of extra steps that could otherwise be avoided.
- It can be difficult to handle large files, or robust data. Text inputs are also cumbersome for SPSS if they are longer than a certain number of characters. Since we often analyze survey data that can contain open ended questions, this is problematic.
- In terms of simple analysis for internal reliability statistics (e.g. Cronbach's Alpha) SPSS was very helpful.
- SPSS has been found easier to use by some colleagues than Excel, (these colleagues are on the lower end of technological capability) and using SPSS alongside RStudio and Excel has given us a variety of methods to accomplish one task.
- The lack of readable user guides and online info hindered our ability to breach into the areas of internal validity testing that we were not familiar with. Updated guides are necessary.
I have also used RStudio and SAS previously. In fact, I'm currently using RStudio since our SPSS license has expired. SPSS lacks the capabilities of these other two programs and it is far less intuitive. Larger data sets can be analyzed with R and SAS, but using these programs also involves coding and understanding the errors one gets with these programs. However, more training is needed to use these programs, and they are not as user-friendly. RStudio is free, and constantly updating, with great graphing packages. SPSS is hindered by its price, stagnant interface, and limited graphing capabilities.
Using IBM SPSS
| Pros | Cons |
|---|---|
Like to use Relatively simple Easy to use Technical support not required Well integrated Consistent Quick to learn Convenient Feel confident using Familiar | None |
- Running frequencies
- Creating histograms
- Writing and running syntax/code
- Bootstrapping
- Importing large data files
- Data cleaning/running errors