I have collected 178 responses of students and their preferences for the different aspects of Higher Education. Which include Availability of the Programme, Tuition Fees, Student Union Activities etc. I have more than 50 variables to test. Each participant was asked to give a score from 1-10 for all of the variables. For some participants, most of the variables hold the same preference, and therefore they have responded with similar scores for most of the variables.
When I am conducting a Factor Analysis in the SPSS, I am getting an error that says that I have less than two cases, which is not true. I have also searched for possible problems, one of which was that I might have some missing data, which I don't.
Can someone please let me know what exactly could be the problem? Further, I need to conduct a Cluster Analysis? Can I use Cluster Analysis instead of Factor Analysis? As I would like to find out the market segments. I have tried conducting a non-hierarchical cluster analysis (K-Means Cluster Analysis) and I am getting some errors.
Hello Aadil,
Perhaps if you would post: (a) a screen shot of your data worksheet; and (b) the syntax corresponding to the (failed) factor analysis attempt, someone could offer a concrete recommendation as to why the program concludes you have fewer than two cases.
Good luck with your work.
@Aadil Below threads might be helpful to your cause.
https://www.ibm.com/support/pages/factor-analysis-returns-error-there-are-fewer-2-cases-variable-0-variance-or-only-1-variable
https://www.researchgate.net/post/When_I_try_to_run_factor_analysis_in_SPSS_I_get_a_warning_message_there_are_fewer_than_two_cases
Hello Hushye,
I reckon you are trying to identify the underlying factors that explain 50 variables through an Exploratory Factor Analysis (EFA).
Please note that there are different routes and decision points in conducting an EFA.
a. Decide on the extraction method. Popular extraction methods are (1) Max Likelihood when data is normally distributed (2) Principal Axis Factoring and (3) Principal Component Analysis (PCA) if we need to drive a minimum number of factors and explain the maximum portion of the variance. Try with each one.
b. Decided on rotation method: There are two popular classes of methods (1) Orthogonal Rotation, used in cases when multiple factors do not correlate and (2) Oblique Rotation, used in cases when multiple factors correlate. Methods like Varimax, Quartimax, Equimax belongs to orthogonal whereas, Oblimin, Promax belongs to Oblique rotation. Try the one appropriate in your case.
If none of the above works, you may increase the Maximum number of iterations to converge. This option is available in the Extraction ribbon in SPSS.
Furthermore, you may even specify the number of factors to be extracted in the same window.
Do check the box corresponding to the Scree plot to get an idea of the extracted factors.
Hope the above information helps.
-Vishnu
Hi Hushye,
I was talking about similar problems with Mohammed Habes who might have an answer as I wasn't happy with the outcomes of FCA.
Let's know if you solve the puzzle.
regards
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size