I have a big data which contains 4787 Observations and almost 100 variables. Questionnaire has some nested questions like selected respondents are asked to Answer Q#2 if they have answers Q#1 as YES and Q#8 would be answered by those who answered Q#4 as YES, like that data is shrinking and missing values are increasing. So, how to handle this kind of missing data in R which are systematic missing not the user-missing data.
Firstly, if I am deleting all the observation with NA, it results in losing 75% of the data and losing good data points.
Secondly, Mice package in R is for user-missing data ( situation in which respondent failed to answer the question).
Kindly help in this regard
Hello Saeed,
This type of missing data is ignorable, since response-contingent questions (e.g., Do you own an automobile? If so, what brand?) that are properly skipped avoids having unqualified respondents give answers to questions. Survey researchers set up such "skip logic" instructions when constructing questionnaires.
So, if you are engaging in an analysis that involves one or more response-contingent questions, you are necessarily restricted to those for whom the question applies. This usually means a subset of the total sample.
It's not a flaw; it's nothing that would require imputation.
Good luck with your work.
Hello Saeed,
This type of missing data is ignorable, since response-contingent questions (e.g., Do you own an automobile? If so, what brand?) that are properly skipped avoids having unqualified respondents give answers to questions. Survey researchers set up such "skip logic" instructions when constructing questionnaires.
So, if you are engaging in an analysis that involves one or more response-contingent questions, you are necessarily restricted to those for whom the question applies. This usually means a subset of the total sample.
It's not a flaw; it's nothing that would require imputation.
Good luck with your work.
I agree with David Morse, it is normal that the number of observation would shrink at subgroup level. Imputation is redundant here.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Analysis
- Data Analysis
- Dataset
- Big Data