Testing Incentive Use

Attachment 7

Testing Incentive Use in the

National Health Interview Survey

Purpose

To test the use of incentives to improve response rates to the National Health Interview Survey (NHIS).

Rationale/Policy Question

The National Health Interview Survey is considered the gold standard of health surveys. It provides benchmarks and is the sampling frame for other federal surveys. Estimates from the NHIS address many of the Department’s initiatives. Response rates on the NHIS, as with other surveys, have been gradually declining. Incentives are known to improve completion rates but have never been used by NHIS. This project would test the use of incentives on the NHIS to determine whether the erosion of response rates can be slowed or reversed

Approach

The NHIS has a complex design with multiple components and the 2010 interview is the lengthiest. This plan will address overall survey participation using different levels of financial incentives. The two Regional Offices with the lowest household response rate in Q3 of 2009 were chosen to develop estimates of the amount of incentives that would be paid. For the family core, Philadelphia had a response rate of 74.7%, while Seattle had a response rate of 77.5% based on Census field reports.

In one Regional Office, $10 would be sent to half the households with the advance letter and $20 would be sent to the other half. For the other Regional Office, half the letters would promise $20 and the other half would promise $40 upon completion of the interview. The evaluation of the project would include not only response rates but also the level of effort required to complete the interview, the rate of item nonresponse (focusing especially on difficult items like income and permission to link records using the last 4 digits of the Social Security Number), complete versus partial interview rates, and other qualitative factors. The test would occur during two calendar quarters to accumulate sufficient observations.

Proposed funding level or if an in-house effort an estimate of staff time

Funding is needed to pay for the incentives, the cost of handling incentives, and for changes to NHIS procedures. NCHS would contribute staff time to plan the project and analyze the results. It is estimated that the cost will be $275,000 for the test.

Priority

I

Proposed partners (financial and non-financial, including other ASPE offices) if any.

The US Census Bureau, the data collection agent for the NHIS, would be an active partner in this research.

Proposed mechanism - include specific contractor or researcher if not planning to compete

Project officer(s)

Marcie Cynamon, Survey Planning and Special Surveys Branch, Division of Health Interview Statistics, NCHS

Estimated Timeframe

The incentive distribution would occur in calendar quarters three and four of 2010.

Deliverables

A final report will be written which summarizes the findings of the study.