Questionnaire Dwelling Unit-Level and Person Pair-Level Sampling ...
Questionnaire Dwelling Unit-Level and Person Pair-Level Sampling ...
Questionnaire Dwelling Unit-Level and Person Pair-Level Sampling ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
4. The respondent had a child <strong>and</strong> a child-in-law living in the house <strong>and</strong> had<br />
gr<strong>and</strong>children younger than 18.<br />
6.4.2 Determining the Final Household-<strong>Level</strong> <strong>Person</strong> Count<br />
For a particular type of household-level person count, there are three types of households<br />
from a sample selection perspective. For the first type, a pair was selected <strong>and</strong> both pair members<br />
responded, where the pair relationship corresponded directly to the pair domain being counted.<br />
In this case, the household-level person count was usually easy to obtain using the multiplicity<br />
counts, although an adjustment was sometimes required if more than one family unit was in the<br />
household. For example, a parent-child pair was selected where the child was 12 years old, <strong>and</strong><br />
the household-level person count for the parent-focus parent-child (12 to 14) domain was<br />
required. In the second type of household, a pair also was selected <strong>and</strong> both pair members<br />
responded, but in this type the pair relationship did not correspond directly to the pair domain<br />
being counted. In this case, determining the final count was sometimes more difficult,<br />
particularly if one or more of the counts was a count of 0. A count of 0 from a roster with good<br />
data did not necessarily mean that the final count should be 0. For example, suppose a household<br />
consisted of a man, his wife, brother, <strong>and</strong> two sons, <strong>and</strong> suppose one of the sons <strong>and</strong> his uncle<br />
(the man's brother) were selected. If the uncle's roster had a count of 0 for all domains of<br />
interest––since all of the household members were "other relatives" from his perspective––then<br />
no nonzero parent-child count could be obtained. The final count would have to be determined<br />
from imputation. In the third type of household, only one respondent was selected. In this case, it<br />
was not possible to match counts from different pair members, but determining the final count<br />
could still be difficult if the count was 0 for a household where the value was not truly 0.<br />
For situations where a pair was selected <strong>and</strong> both pair members had good roster data, if<br />
the counts agreed between the pair members <strong>and</strong> were not 0, then an easy determination of the<br />
final household-level count was possible. Surprisingly, this occurred in a majority of cases. If<br />
one pair member had a bad roster with no information in it <strong>and</strong> the other had a good roster, this<br />
was treated in the same way as if a single respondent was selected with a good roster. In either of<br />
these cases, the final count could be determined, provided a considerable number of conditions<br />
were satisfied. The conditions used to accept a good roster's count, when either the other pair<br />
member's roster was bad or no pair was selected, are provided in Appendix S. If these conditions<br />
were not met, the final household-level person count was left to imputation. Imputation also was<br />
required if two pair members were selected, both with bad rosters.<br />
For the remainder of cases, some could be reconciled <strong>and</strong> some could not. In the cases<br />
where reconciliation was possible, some of the disagreements were caused by the pair members'<br />
rosters having different age <strong>and</strong> gender compositions. In these cases, many of the disagreements<br />
between the pair members were resolved by going to the screener. However, the screener did not<br />
provide much help if the age <strong>and</strong> gender composition of the pair members' rosters were identical,<br />
yet the counts still disagreed, as was the case with the nephew-uncle pair described above. In that<br />
example, one count was 0 <strong>and</strong> the other was nonzero. Under conditions set out in Appendix S, it<br />
was possible to determine that the disagreement in this case was due to the uncle not being able<br />
to identify the parent-child domains, <strong>and</strong> the nonzero count was used. More detailed rules for<br />
reconciling differences between pair members are described in Appendix S.<br />
57