I have a data-set of 30000 population where i got large no. of missing values >10% and variables are taken from 3 different phases for outcome, exposure and mediator. When i tried multiple imputation (followed Youtube videos ), i couldn't treat missing values well, and still there are >10% missing values. So, could you please suggest me what should i do or how to treat Missing values properly???
Treating missing values need specific attention from the researcher.
first, you need to run descriptive statistics and to diagnose not only the amount but also to check if the missing is related to specific variable, or cases, or setting, to phase of of study.
then you need to run skewness test to check if the distribution is normal or skewed for the variables.
if the distribution is approximate to normal, then you can use the mean for replacing the missing, otherwise you need to use the median.
and if the missing is related to specific variable or setting,,etc...then it is preferred to replace with the mean or median for that setting.
There are many other issues related to missing, but the above ones can be useful
Regards
Treating missing values need specific attention from the researcher.
first, you need to run descriptive statistics and to diagnose not only the amount but also to check if the missing is related to specific variable, or cases, or setting, to phase of of study.
then you need to run skewness test to check if the distribution is normal or skewed for the variables.
if the distribution is approximate to normal, then you can use the mean for replacing the missing, otherwise you need to use the median.
and if the missing is related to specific variable or setting,,etc...then it is preferred to replace with the mean or median for that setting.
There are many other issues related to missing, but the above ones can be useful
Regards
In SPSS, you should run a missing values analysis (under the "analyze" tab) to see if the values are Missing Completely at Random (MCAR), or if there is some pattern among missing data. If there are no patterns detected, then pairwise or listwise deletion could be done to deal with missing data. However, if the missing values analysis detects a pattern, then imputation must be done.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS