Hello, I am doing an Exploratory Factor Analysis on data from a 15-item questionnaire that contains binary (yes/no) response options. To perform this analysis, I first computed a tetrachoric correlation matrix. However, I get a message that says "matrix with tetrachoric correlations is not positive definite. It has 2 negative eigenvalues." After searching online, I found that including the "posdef" command at the end of the tetrachoric command fixed the issue, and let me run the exploratory factor analysis. However, I'm wondering if any of you know what the implications are of forcing a correlation matrix to be positive definite through the "posdef" command? Thank you!