how to deal with inapplicable values
Dear support team
a lot of the variables in the Understanding Society dataset has a high percentage of inapplicable values. For example, the highest qualification variable, w_qfhigh, has around 85-90% of inapplicable values. How should I deal with the inapplicable values? Is inapplicable values the same as missing values? why is a variable like the highest qualification has so many inapplicable values? I would assume most of the people in the sample has some kind of education. Any other inputs on the questions will be much appreciated.
#1 Updated by Gundi Knies over 5 years ago
- Category set to Data documentation
- Assignee set to Gundi Knies
- % Done changed from 0 to 90
we differentiate between a number of different sources of missingness. The full listing is reported in Table 19 of the Understanding Society User Guide (2014).
You will see in the questionnaire routing that _qfhigh is only asked when a respondent first enters the study, hence the large number of [-8 missing] from Wave 2 onward. Differences between _qfhigh and _qfhigh_dv are explained in the variable level descriptions of the respective variables. _qfhigh_dv is a derived variable and the variable level description includes references to which variables are used in its construction. See, for example:
Hope this helps,