I am currently working on tender data for my master dissertation. However this tender data is often not completed, f.e. the datum on which they decide who gets awarded the contract is often missing. However, the fact that this data is missing can be an important conclusion for me, so i don't want to delete this contract out of my sample. How can I solve this? Should I combine this variable with a dummyvariable stating that 0 is unknown, 1 is known?
Kind regards,
Luca
Your suggestion could be a viable strategy, yet it may require quite some work to add one more variable to the database of results. And, upon evaluation of the data, your queries might get a bit complicated.
I think an easier and quit robust solution is to substitute the missing value with some easily distinguished value of the same type, such as if the date of contract award is missing ("DATE" type variable), you might substitute the missing value with something like "01.01.1970". Likewise, if the contractor name ("STRING" type) is missing, you might use something like "missing", "not stated".
You may also opt for leaving the respective field in the record "BLANK" and query for records with "BLANK" value in the respective field. This strategy carries a risk that the "BLANK" value can get "nonblank" e.g. by inadvertent substitution (or just typing in) with some (invisible) non-blank value (such as empty - BLANK - string might get replaced with " " - space) which the query would no more recognise as "BLANK".
- Badges
- Science topic
- Similar topics
- Medical Science Nursing
- Dissertations