results of factor analysis (pca, varimax)


New Member
Hey there,

I need your help in understanding the results of a factor analysis I did.

Base level: data come from a nation wide survey (n = 1000). We asked people about meanings they assign to 7 different (typical) landscape types (a,b,c,d,e,f,g). Therefore we developed a set of 20 questions/statements (1:20). An example for the questions asked is "this landscape type represents the beauty of nature" or "this landscape type represents the impact of human being" ... Each respondent was asked to every combination of landscape type and meanings type, which makes a total of 20 questions * 7 Landscape types = 140 answers per respondent.

In a post process of the survey I did dimension reduction via factor analysis. Therefore I chose principal component analysis with varimax rotation. After analysing the screeplot I decided to use the 9 factor solution.

However, the result is for my opinion quiet weird and I do not know how to deal with it. The factors exactly represent the different landscape types. So every item asked to a special landscape type is loading to the factor of that landscape type. For a better illustration I attached a screenshot of the first few rows of the component rotation matrix. On the very left column you can see the name of the landscape type in CAPITAL letters (right after "meanings". And right after the "-" there is the statement which is identical for all landscape types.
But there are also 2 factors that do not represent a single landscape. They represent a conglomerat of "bad items" like human impact etc. for urban landscapes in one factor and rural landscapes in the other one.

So I have never computed a factor analysis with such a result and I am wondering If this can be true. If there wouldn't be the 2 factors with the human impact I would doubt the correctness of this analysis. But since they are evident, I am thinking of interpretation. Could it be that people really have such specific and divers meanings about landscapes at the same time?

thanks a lot for your help and sorry for the long post, but I felt the need to explain my situation clearly.
all the best,
boris Bildschirmfoto 2019-01-04 um 13.56.10.png Bildschirmfoto 2019-01-04 um 13.56.10.png