I am currently working on constructing a scale. The questions on the scale are measured on a scale from 0-100 in increments of 10, similar to the DES. I ran a factor analysis with my pool of items and eventually got to a scale I liked the psychometrics of. At this point, I realized I didn't reverse-score a subset of items on my scale. I went back and fixed the items that weren't properly scored and ran my factor analysis again, same items included, same method, etc. To my surprise, the results are exactly the same as before (excluding increased reliability for the overall scale because the error in scoring was messing with the reliability before).
Does this make sense/is this possible? I can't understand how the factor analysis results can be the same (down to the factor loadings) even though the values for certain items have reversed. My PI wasn't sure either so I was hoping to crowdsource some knowledge : )
Does this make sense/is this possible? I can't understand how the factor analysis results can be the same (down to the factor loadings) even though the values for certain items have reversed. My PI wasn't sure either so I was hoping to crowdsource some knowledge : )
