The PhilPapers SurveysFactor Analysis (Back to contents)
Caveat: There are many different ways to perform a factor analysis, and we are not experts in the area, so the following should be regarded as a suggestive illustration rather than as a serious scientific analysis. Furthermore, factor analysis results are strongly dependent on the choice of questions for the survey.
Download resulting component matrix: Excel, OpenOffice, PDF
To obtain this component matrix we used one variable for each of the thirty questions (for non-binary question we chose the variable with strongest correlations to other questions). To extract the components, we used the principal components method, stopping when eigenvalue < 1. We performed a Varimax rotation with Kaiser normalization, which converged in 8 rotations. In the attached component matrix we list only loadings with an absolute value greater than 0.25.
The interpretation of the factors is a subjective matter, but one could naturally suggest that the first four factors are strongly connected to:
- Naturalism
- Realism
- Rationalism
- Externalism