Hi,

There are chapters on item parceling in Barbara Byrne's book 

"Structural equation modeling with MPLUS" and in Kevin Kelloway's book "Using Mplus for Structural Equation Modeling"

And, of course, a paper by Todd Little is a "must" (doesn't deal with MPLUS though):

http://faculty.psy.ohio-state.edu/cunningham/pdf/little.sem.2002.pdf

When I use parceling, I just prepare a data file outside of MPLUS, with some other software like SPSS...

Thank you very much for your tips, Marina! Also for the tip of using SPSS to prepare the data file.

Dear Noud,

You can use the DEFINE command to create parcels within MPlus. 

For example with three items:

DEFINE:

parcel1 = (item1 + item2 + item3)/3

Than add this new variable in the USEVARIABLES ARE line behind all wihtim data set variables u use in your model and it should work.

Best

Dear Manuel,

Thank you very much for your reply. I understand your example, I wonder however what to do if I have 8 items for 1 parcel? Can I just combine them all in 1 parcel? I don't think so, but I'm not sure how to combine the items in multiple parcels? Do I have to look at correlation coefficients for example?

kind regards,

Noud

Dear Noud,

in general, you might use two parcels (or four), it depends on the quality of your data and the items you want to group and the nature of your analysis. In general I would not use one parcel since you have to fix the residual variance otherwise. And: if avoidable you might use all 8 indicators instead of parceling. 

For two parcels with eight items it would look like this:

DEFINE:

parcel1 = (item1 + item2 + item3 + item4)/4;

parcel2 = (item5 + item6 + item7 + item8/4;

Regarding your second question:

Like usual, it depends which is the optimal strategy to parcel. The dimensionality of your items is critical, especially when using multidimensional items parceling becomes difficult. Some approaches like correlational parceling (see the added paper of Little et al 2012) might be useful. However, there is no general one size fits all solution. There is a big amount of Literature dealing with this topic, and some very critical views (just to make the situation more difficult: https://www.statmodel.com/download/marsh_ludtke_2013.pdf). 

Here you can find the Paper of Little et al. (2012):

https://www.apa.org/pubs/journals/features/met-a0033266.pdf

Hope that helps,

Best

Dear Manuel,

Thank you very much for your clarification! I will read the paper soon!

Best wishes,

Noud

Dear Noud,

Another helpful resource about parceling:

Bandalos, D. L. (2002). The effects of item parceling on goodness-of-fit and parameter estimate bias in structural equation modeling. Structural Equation Modeling, 9(1), 78-102. doi:10.1207/S15328007SEM0901_5

Best wishes,

Tehila

Thank you very much for your suggestion, Tehila!

Best wishes,

Noud