The Kaizer-Mayer-Oklin Test of Sampling Adequacy is used to verify that the sample size is enough for conducting factor analysis. The value of the test should be greater than 0.5. The closer it is to 1.0, the better.
Another assumption to be considered is to make sure that there are a minimum amount of inter-correlations between the variables being tested. The Bartlett's test of sphericity is used for that purpose. A significant result (p<5%) indicates that the correlation matrix is not an identity matrix.
Yet another assumption to be tested is that there should not be too high inter-correlations between the variables, in other words, there should be no multicollinearity.
