Factor rotation after exploratory factor analysis

Рет қаралды 23,575

Күн бұрын

Пікірлер

@justinlross438 4 жыл бұрын

I've been trying to understand this concept after reading about it and watching lectures, you explained it in 7 minutes perfectly. Subscribed.

@edougg1 4 жыл бұрын

I really like the explanation but you misspoke: oblimin is an OBLIQUE rotation does the factors end up CORRELATED and Varimax was supposed to be described as ORTHOGONAL so the result is UNCORRELATED. Otherwise thank you very clear explanation

@davidgleaves5175 4 жыл бұрын

@@mronkko when you first defined oblique and orthogonal around 3:50 you said it backwards (you said oblique maintains uncorrelated factors).

@mikorees5853 5 ай бұрын

Thanks! the rule of thumb is really helpful

@mronkko 5 ай бұрын

You are welcome!

@yuripavlov5091 4 жыл бұрын

very well-explained, thank you!

@mrr1491 4 жыл бұрын

Thanks for doing this video! it was helpful for me.

@jannomeeuwessen4886 8 күн бұрын

I thought I understood it until he drew the Orange and Black axis lol Why are there multiple axis? The first one he drew (red) should be enough right

@mronkko 8 күн бұрын

Each line reprsents a factor. The idea of the factors is to represent a multidimensional space (variables) with a smaller number of dimensions (the factors). In this example the general factor tells which direction the observatiosn are from the origin and the second which direction they are related to the first factor. Does this help?

@jannomeeuwessen4886 8 күн бұрын

@@mronkkoso If I understand it correctly: - The red line shows which observations / items are above the line and which are below (seperating them into 2 factors) - The black line is to point to the 2 seperate clumps of items that each represent a factor - and the Orange line is to make that pointing more specific Is that correct? Are those the steps you need to follow or something?

@mronkko 8 күн бұрын

@@jannomeeuwessen4886 The red coordinate system is the unrotated factor solution. The black coordinate system is a rotated factor analysis with orthogonal rotation. The orange coordinate system is a rotated factor analysis with oblique rotation. You would generally first estimate the unrotated solution. Whether it is printed out or not depends on statistical software. Then you would do either orthogonal or oblique rotation. The oblique rotation is typically a better choice (at least for the kinds of problems I use factor analysis for.)