Ravi Jasuja posted on Wednesday, June 07, 2006 - 5:40 am
I am currently using a dataset in which I have multiple items/indicators for several factors/latent constructs. I am debating whether to use these items separately as sub-factors or create an average or a sum of these items. Please advise as to what are the advantages and disadvantages of both approaches.
Also, if you could recommend any articles in this regard would be greatly appreciated.
I would not create sum scores without at least 15 factor indicators for each factor. The advantages of not using sum scores is that the items are not weighted using the value one but are weighted by the factor loadings. You might want to look at the item parceling literature for more information about this topic.
Ravi Jasuja posted on Thursday, June 08, 2006 - 3:32 am
Thank you so much for your prompt response. It has really helped me a lot. As advised by you, I did read the item parcel literature and it really cleared up a lot of concepts.