The “worksheet” treatment fielded in the 2005 test (see Figure 6-7) is an example of the question-based structure we envision for Question 1, breaking the complex cognitive task of determining a complete household count into smaller (and ideally more tractable) components, such as those persons staying at the household temporarily on Census Day. The intended effect is to replace dense instructions with a guided series of questions. The text of the questions should be concise and so harder to cursorily scan or skip.
It should be emphasized that the specific “worksheet” form tested in 2005 is not the only possible configuration of questions, nor may it be the optimal one. We also emphasize that the same warning that applied to the development of specific residence rules applies here: just as it is confusing and undesirable to carve out a new residence rule for every possible living situation and population group, so too is it unwise to divide the task of deriving a household count across too many questions and categories.
In addition to a revised structure of the basic household count and listing question at the beginning of the census questionnaire, a fully question-based strategy for gathering accurate residence information requires additional queries in the body of the questionnaire. Such additional questions are necessary to obtain enough information to make an individual-level “usual residence” determination and to make sure that a respondent is being counted at the correct place.
The assertion that “the short form is too short” is a strong one and is meant to draw attention to the need for some additional data collection; to be clear, though, we do not suggest a radical lengthening of the form. Followup interviews like the Accuracy and Coverage Evaluation interview used in 2000, or the planned coverage follow-up operation for 2010 (see Chapter 8), can include an entire battery of residence-related questions. It is proper that these highly detailed interviews ask numerous residence questions, while the question content of the census itself is kept more limited.
Two basic conceptual checks need to be made in order to determine whether a person’s enumeration location (that is, where the questionnaire finds them) is where they should be counted in the census:
Does this person have a residence elsewhere that could be considered their usual residence?
Is this census address correct, for this person and this building (physical structure)?
These two concerns prompt the questions we suggest adding to the census questionnaire. There is also a need for reinstating coverage probes to the cen-