21.06.2014 Views

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 ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Table 6.2<br />

Measures of the Quality of Definitive Roster Matches<br />

Measure<br />

Number<br />

Description<br />

0 Age <strong>and</strong> gender matched exactly, with exactly one MBRSEL correctly identifying the other<br />

pair member<br />

1 Age <strong>and</strong> gender matched exactly, with MBRSEL correctly identifying the other pair<br />

member, but there was more than one MBRSEL 1<br />

2 Age within one, gender matched exactly, with exactly one MBRSEL correctly identifying<br />

the other pair member<br />

3 Age within two, gender matched exactly, with exactly one MBRSEL correctly identifying<br />

the other pair member<br />

4 Age <strong>and</strong> gender matched exactly, with MBRSEL missing for all roster members<br />

5 Age matched exactly, gender off, with exactly one MBRSEL correctly identifying the other<br />

pair member<br />

6 Age within one, gender matched exactly, with MBRSEL correctly identifying the other pair<br />

member, but there was more than one MBRSEL 1<br />

7 Age within two, gender matched exactly, with MBRSEL correctly identifying the other<br />

pair member, but there was more than one MBRSEL 1<br />

8 Age within one, gender matched exactly, with MBRSEL missing for all roster members<br />

9 Age within two, gender matched exactly, with MBRSEL missing for all roster members 2<br />

10 Age within 10, gender matched exactly, with exactly one MBRSEL correctly identifying<br />

the other pair member 2<br />

1<br />

Since the 2001 survey, it was technically impossible to identify more than one roster member as the "other pair member<br />

selected," resulting in either 0 or 1 MBRSEL for each responding pair. As a result, measures #1, #6, <strong>and</strong> #7 did not occur in the<br />

2006 survey.<br />

2<br />

For pairs where one pair member had a match corresponding to measures #9 or #10, if the other pair member had a match no<br />

better than measure #9, an additional requirement was implemented where the reported household sizes for both pair members<br />

had to be equal to 2.<br />

Given that at least one side had a match according to one of the measures provided in<br />

Table 6.2, the other side could have a match that was weaker (i.e., not definitive), using the<br />

measures in Table 6.3. Additional columns are provided in Table 6.3, showing the weakest<br />

match that was allowed (as denoted by the measure) for the other pair member. The column<br />

titled "In Code" shows the weakest measure allowed in the code, <strong>and</strong> the column titled<br />

"Observed" shows the weakest measure that was actually observed for the other pair member.<br />

In the cases where a single roster member had to be selected among duplicates (measures<br />

#14 through #18), where the duplicates had the same relationship code, it was necessary that the<br />

relationship codes be limited to child or sibling.<br />

In some cases, due to the poor quality of the rosters of the pair members, it was not<br />

possible to locate the listed household member in A's roster that corresponded to pair member B,<br />

<strong>and</strong> vice versa. The determination of the pair relationships for these cases was left to imputation.<br />

Even when a pair of roster members was successfully identified, it was not always possible to<br />

successfully determine the pair relationship, as is pointed out in the next section.<br />

30

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!