I am making a research and I have the following dilemma: My material contains some specific allergic diseases, e.g. asthma, birch and egg. I have one variable that checks if there is any allergic disease at all (1) or if all the individual values are zero (any allergic disease = 0). But there are some missing data and it might look like:
- Case1: Asthma=1, Birch=0, Egg=0, Any Allergy=1
- Case2: Asthma=0, Birch=0, Egg=0, Any Allergy=0
- Case3: Asthma=1, Birch=N/A, Egg=0, Any Allergy=1
- Case4: Asthma=0, Birch=N/A, Egg=0, Any Allergy=???
Cases 1 and 2 are straightforward. For me it seems logical to calculate Case 3 Any allergy to 1, since the unknown result for Birch wouldn’t matter, the Asthma is already 1. Is this right or should I omit this record even though I know that it contains sufficient information to set Any Allergy = 1? And what should I do with Case 4? I cannot know if it would be 1 or 0 and therefore, I cannot determine Any allergy to 0 or 1. If I leave it out my summary would be (in the above example) that 2/3 has any allergy (Given that Case 3 is 1). At the same time, in the same report I say that the Asthma is 2/4. The denominator (or N) is different, and this confuses me. I cannot claim that Any allergy would be zero either because the Birch might as well be 1. One solution might be to omit all records where any of the allergies = N/A but I would loose a lot of information if all those rows are unconsidered. Could anybody shed some light on this, I would appreciate it a lot. Thanks in advance.