Multiple imputation: Reporting frequency and percentages of categorical variables

[saba.haq]

Hi,

I have very dumb question to ask. This is the first time I have imputed data using multiple imputation [MI] in SPSS.  I have imputed some of my demographic categorical variables for my thesis. I want to report the frequency and percentages of these variables now. I am not sure how to go about it. I know I have to report the pooled results but the pooled results report some of the categories in decimal points not in integers?? for example like for categories of age it is as follows:

Age:
imputation 1: 1 year   582
imputation 1: 1 year   585
imputation 1: 1 year   586
imputation 1: 1 year   585
imputation 1: 1 year   586
Polled result: 1 years 585.2

It is not possible to have number of participants in decimals in a category and does not really make sense to report it in this way. And I am not even sure if I am supposed to report this before doing MI or I can do it after or if there is any other way to do it.

It will be a huge help if someone could tell me ASAP what is it that I am not missing here.

[Maguin, Eugene]

You may have already gotten useful off-list replies.

The results make perfectly plausible sense. There is missing data for the age category variable group and what you're showing us is the number of persons in one of the age categories. You're imputing value; therefore, some of the persons with missing data for this variable group are going to be imputed to be in this specific variable, i.e., they are estimated to be age 1 year. My opinion is that for each variable that had missing data four quantities are of interest: the observed data value, i.e., the proportion in this case; the missing data percentage, the mean value across the imputed samples, the range of the sample values across the imputed samples.

Gene Maguin


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of saba.haq
Sent: Monday, February 25, 2013 8:32 PM
To: [hidden email]
Subject: Multiple imputation: Reporting frequency and percentages of categorical variables

Hi,

I have very dumb question to ask. This is the first time I have imputed data using multiple imputation [MI] in SPSS.  I have imputed some of my demographic categorical variables for my thesis. I want to report the frequency and percentages of these variables now. I am not sure how to go about it. I know I have to report the pooled results but the pooled results report some of the categories in decimal points not in integers?? for example like for categories of age it is as follows:

Age:
imputation 1: 1 year   582
imputation 1: 1 year   585
imputation 1: 1 year   586
imputation 1: 1 year   585
imputation 1: 1 year   586
Polled result: 1 years 585.2

It is not possible to have number of participants in decimals in a category and does not really make sense to report it in this way. And I am not even sure if I am supposed to report this before doing MI or I can do it after or if there is any other way to do it.

It will be a huge help if someone could tell me ASAP what is it that I am not missing here.




--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Multiple-imputation-Reporting-frequency-and-percentages-of-categorical-variables-tp5718247.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD