I have some cases in my dataset where respondents have abandoned the survey half-way, so I coded the missing values as 9999 "Refused to answer". In addition, for likert-scale type of items I applied the code 8888 for "Don't know" in the data set. Finally, I have some skip patterns: some variables of these skip patterns are categorical and some are scale and ordinal. For the legitimately skipped answers, I have coded them 7777 "Legitimate skip" in the data set.

I've set up all these codes in the "Values code" column in SPSS.

My questions are:

• For the codes 8888 and 9999, should I define them as "Discrete missing values" in the "Missing" column in SPSS as well?
• For the code 7777, how should I consider it, as a missing value as well? (I don't think "Legitimate skips" are actually missing values in the proper sense of the word, but if I don't code it as Missing, then for scale variables it will be a problem, as it will be added up in means, etc. right?)
  

Thank you to anyone who can help!