Recode multinomial variable into new binary variable in SPSS

In this video I show how to create a new variable in SPSS that converts a multinomial variable into a binary variable for use in multigroup moderation in AMOS.

[@angelaquaye7122]

Wow this was very helpful. I am not so good in spss and I would like to find out, suppose I want to create a new binary variable and I am interested in the widowed. When I recode into different variables, what value would I then give the widowed who are coded 3, and what value would the rest get? Hoping for a quick response. Thanks

[@Gaskination]

Whatever you want to be the value of interest, make that = 1 and make the rest = 0.

[@alirezashabanishojaei6331]

first of all, I would like to convey my regards to you Dr.James Gaskin
Could you please mention that what we must report on our thesis form output of CFA and SEM.
which item is compulsory and which items it is optional. thank you in advance :)

[@Gaskination]

+Alireza Shabani Shojaei I have an example analysis on my statwiki in the general guidelines section. This will help you know what to report.

[@alirezashabanishojaei6331]

+James Gaskin thank you so much

[@yuren3267]

Dear Dr. I just want to analysis the "multinomial variable", without converting it into "numeric", in SEM, may I how to do that? Please

[@Gaskination]

Can't. SEM software expects numeric values. The numeric values can represent multinomial categories. For example, you could have a variable for industry that codes manufacturing=1, retail=2, service=3. However, unless you are using this as a multigroup variable, you still must break it up into binary dummy variables if you want to include these as predictors or outcomes in your model. The reason for the reason for this is that the numbers (1,2,3) represent categories that are not intrinsically "more" or "less" than each other. For example, a 3 on the scale does not mean "more industry". It just means service industry.

[@yuren3267]

@@Gaskination Thank you very much. it gives me a lot of help.