How to actually write a multiple Imputation with Syntax?

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Erefin
Posts: 1
Joined: Fri May 26, 2017 3:47 pm

How to actually write a multiple Imputation with Syntax?

Postby Erefin » Fri May 26, 2017 4:09 pm

Hey guys and girls,

I'm currently visiting a seminar that's meant to simulate all the hardships in a researchers daily life, including the handling of Missing Data. As such, I'm working on an (arbitrary) SPSS-file I've received, which consists of 26 different variables (Q1 to Q26), which are items of a four-factor questionnaire and of these items some got missing values (Q13,Q16,Q18,Q21,Q25).

So the first thing I did was to select the MI (multiple Imputation) via the built-in SPSS-menu, setting Q13 as the Criteria (to be predicted) and Q8 Q9 Q12 Q13 Q14 Q15 Q23 Q24 Q25 as Predictors. That operation gave me five imputations for each participants. After deleting the non-imputed Q13's I wanted to create a mean of all the different Imputations (VarName: Imputation_) of Q13, i.e. a new Variable called Q13ImpMeans.

For that, I wrote the following Syntax:

compute Q13ImpMeans = -222.
formats Q13ImpMeans(F4).
execute.

but none of THE FOLLOWING worked:

compute Q13ImpMeans = means Imputation_ by Q13.
execute.

compute Q13ImpMeans = means (Q13) by Imputation_.
execute.

compute Q13ImpMeans = ((Q13 (Imputation_ = 1) + Q13 (Imputation_ = 2) + Q13 (Imputation_ = 3) + Q13 (Imputation_ = 4) + Q13 (Imputation_ = 5))/5).
execute. <--- SPSS complained about that command saying it'd be coded as a Vector - but why?

So, what did I do wrong? And how can I compute the Q13ImpMeans (each one for the 46 participants?

Cheers and thank you a lot for your expertise and patience
Felix

Who is online

Users browsing this forum: No registered users and 2 guests

cron