Counting missing values per case

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

Posts: 1
Joined: Thu May 24, 2012 12:41 pm

Counting missing values per case

Postby mattb » Thu May 24, 2012 12:56 pm


I've searched the forum and couldn't find anything that answered this question......

I have a dataset with about 200 variables and 180 cases. I'm writing up a methodology paper and I'd like to comment on the mean number of missing items for each case so that I can make a statement such as 75% of people completed 90% or more of the questions.

HOWEVER I can't work out how to do this.

I'd like to specify which variables I'd like to include as I've created quite a few new ones which I don't want to be in the total. One of the possible issues is that there's a mixture of user defined missing (for string variables) and system missing (for numeric variables).

I'd be really grateful of any suggestions you might have.

Best wishes,

Posts: 56
Joined: Mon May 02, 2011 6:07 pm

Re: Counting missing values per case

Postby apeape » Fri May 25, 2012 3:44 am

1. Use the autorecode function on all your string variables making sure you check the 'Use the same recoding scheme for all variables' option. Take note which value has been assigned to your missing designation for strings, and define this value as missing for your newly created variables.

2. Sort all the variables you are interested in determining the missing responses for, including the newly recoded ones and excluding the string variables, so that they are contiguous in your data file.

3. Run the following syntax (inserting the variable names of the first and last variables in your contiguous block):

Code: Select all

compute nummissing = nmiss(firstvar to lastvar).

Who is online

Users browsing this forum: No registered users and 2 guests