NCT Methodology Brief

2012 National Census Test

Internet Content Methodology Brief

March 20, 2012

Treatments

A: Roster

A1: Modified 2005 NCT (Popcount box approach)

Allows us to see whether the presence of a question asking the number of people in the household, along with the residence rules instructions helps respondents create an accurate roster.

A2: Modified CQS (Multi-mode approach)

Allows us to begin a new approach – one where we ask guided questions to elicit residence data.

B: Undercount

B1: Modified CQS

A household level question that probes to determine if there are any additional people not listed originally that should be included for consideration as residents of the household. (similar to paper approach – several types of people and living situations are shown in a bulleted list.)

B2: Modified Multi-mode approach

Allows us to begin a new approach – one where we ask guided questions to elicit response data. Several short household level questions are asked types of people and living situations that might apply to someone in the household.

C: Race

C1: 2010 AQE X3 (combined Hispanic origin/race)

C2: 2010 AQE X4 (combined Hispanic origin/race)

D: Overcount

D1: 2010 MOMB approach

A person level question that asks if the person had another residence. Similar to the approach used across all modes in the 2010 Census.

D2: Mobile approach

Allows us to begin a new approach – one where we ask guided questions to elicit response data. Respondents are asked several short household level questions about types of people and living situations that might apply to someone in their household.

Path 1* Path 2

A2

A1

B2

B1

B1

B2

B2

C2

C1

C1

C1

C2

D1

D2

D2

D2

D1

D2

D2

D2

D2

D2

*Path 1 was chosen in an effort to ‘rule out’ A1 as a viable approach on the web. Treatments B2 and D2 were seen as the most likely to go forward, and therefore chosen to be tested with A1 (since we do not have the ability to test all combinations of B and D with A1).