PROC MI - Height & Weight of obs 128 still missing

[drkd]

Hello Gurus,

I'm trying to impute height & weight of row 128 in this dsn but not sure what I am missing.... I can't seem to impute these 2 columns in obs 128.   Any ideas?

DATA HEART;SET SASHELP.HEART;RUN;
PROC PRINT DATA=HEART(FIRSTOBS=126 OBS=130);TITLE 'BEFORE IMPUTING';RUN;
proc mi data=HEART seed=12345 nimpute=5 out=HEART_mi; var HEIGHT WEIGHT;RUN;
PROC PRINT DATA=HEART_MI(FIRSTOBS=126 OBS=130);TITLE 'AFTER IMPUTING';RUN;

Thanks.

[Rick_SAS · Accepted solution]

Look at the Missing Data Patterns table:

It shows that there are two observations for which BOTH Height and Weight are missing. The imputation does not impute those records. For other records, the procedure can estimate the joint distribution of Height and Weight and conditionally impute one missing value when the other variable is nonmissing.

One way to get imputed values when both Height and Weight are missing is to include a third variable. For example, if you think that joint distribution of heights and weights might be dependent on the SEX  variable, you can use

proc mi data=HEART seed=12345 nimpute=5 out=HEART_mi;
class SEX; 
var HEIGHT WEIGHT SEX;
FCS;
RUN;

[drkd]

Thanks Rick! That helped.

[vsandoval6]

Please help with this probelem

Tabulate the percent of observations in the SASHELP.HEART dataset that have non-missing values for all the predictor variables that you will use in later analyses: AgeAtStart, BP_Status, Chol_Status, Cholesterol, Diastolic, Height, MRW, Sex, Smoking, Smoking_Status, Systolic, Weight, and Weight_Status.

[Rick_SAS]

Since this sounds like a school assignment, I will give you a hint rather than a complete SAS program. You can get your answer by reading the article "Count the number of missing values for each variable."

The article shows how to get counts. To get percentages also, just remove the NOPERCENT option on the TABLES statement in PROC FREQ.