But what this means? When computing the responses of the respondents percentage distribution for each alternative is of the total number of respondents to the question, regardless of the substantive values of the alternatives. So, the question on the nature of employment distribution of answers to alternatives depends on the total number of respondents to this question, i.e. those who have temporary or permanent job, and those who do not have work. Because the total number of respondents increases because of the latter, so far data on the nature of employment we obtain a not group of people having work, and for all respondents. It is clear that these data are not entirely correct (they are understated), and the distortion will be greater, the more these groups differ in number.
Here is another example. "How often do You read fiction?" (often, seldom, rarely, never read). As in the previous example, the question is built on two bases: whether the intensity of reading fiction including among those who do not read literature at all. Calculate what happens if you build a set of alternatives not one, but two reasons. Let's say that half of the respondents read the literature, and the other half reads. Then the distribution of responses for alternatives in the first and second cases will be as follows:
![]()
Here is another example. "How often do You read fiction?" (often, seldom, rarely, never read). As in the previous example, the question is built on two bases: whether the intensity of reading fiction including among those who do not read literature at all. Calculate what happens if you build a set of alternatives not one, but two reasons. Let's say that half of the respondents read the literature, and the other half reads. Then the distribution of responses for alternatives in the first and second cases will be as follows: