Incentive Memo

M EMORANDUM
DATE:

April 24, 2009

SUBJECT: 2009-2010 NS-CSHCN: Proposed Incentive Groups
Based on discussions at NORC and NCHS, the following revised groups are recommended for the incentive
experiment in quarters Q3/2009 and Q4/2009 of the National Survey of Children with Special Health Care
Needs (NS-CSHCN). These incentive groups (summarized in Table 1) have been designed to refine the
successful incentive effort fielded in the 2007 National Survey of Children’s Health (NSCH) and to answer
seven key questions about prepaid and promised incentives.
All proposed incentive groups require that cases be age-eligible (a child lives in the household) before an
incentive is offered. Age-eligibility can be determined during the NIS or NIS-Teen screening, or at the start
of the NS-CSHCN interview. In addition, several of the incentive groups are refusal-based. Due to the
complex relationship between the NIS, NIS-Teen, and NS-CSHCN surveys, prepaid incentives cannot be
mailed systematically after first refusals because some respondents refuse participation before reaching the
NS-CSHCN interview. As a result, refusal-based incentive groups will test the promise of a monetary
incentive for cases with one refusal and the use of prepaid incentives after cases’ second refusal.
Table 1. Incentive Experiment Groups
nd

2 Refusal
st

1 Refusal

Cases with
Mailing Address

Cases with
No Mailing Address

Group 1

No incentives offered at any point

Group 2

Offer of $5 directly after age-eligibility established

Group 3
3B
Group 4
5A
Group 5
5B
Group 6

Age-Eligibility Established

3A

Offer of $10
upon call back

$5 mailing
with promise of $10

Offer of $15
upon call back

No offer

$5 mailing
with promise of $10

Offer of $15
upon call back

No offer

$5 mailing
with no promise

Offer of $5
upon call back

Offer of $10
upon call back

$1 coin mailing
with promise of $10

Offer of $11
upon call back

No offer

$1 coin mailing
with promise of $10

Offer of $11
upon call back

No offer

$1 coin mailing
with no promise

Offer of $5
upon call back

Tests of
$5
prepaid

Tests of
$1 coin
prepaid

Group 1 serves as the control group and Group 2 serves as a key experimental group. In group 1, age-eligible
cases are offered no incentives, and in Group 2, age-eligible cases are all offered $5 as a token of appreciation.
The remaining experimental groups test refusal-based incentive models. Group 3A replicates the approach
adopted in 2007 NSCH (where age-eligible cases are offered $10 on call back after the first refusal; after the
second refusal, they are mailed $5 with a promise of $10) and is known to be successful, while Groups 3B, 5A
and 5B are variations on reducing the costs associated with this successful approach (see Table 1). In Groups
4 and 6, age-eligible cases are mailed prepaid incentives after the second refusal without a promised incentive
to measure the need to offer additional money.
This design allows NCHS and NORC to investigate the following seven questions:
Page 1

Question 1. Are monetary incentives necessary to increase participation?
Our first goal is to determine whether an incentive effort is needed for 2009-2010 NS-CSHCN to increase
response rates. Group 1, where no incentives are offered, will be compared to the experimental groups
(individually and collectively) to measure the effects that incentives have on participation.
Question 2. Are refusal-based monetary incentives more cost effective than eligibility-based
monetary incentives?
Our second goal is to determine whether a refusal-based incentive effort is needed for 2009-2010 NSCSHCN to increase response rates while maintaining reasonable costs. The cost of offering $5 to all ageeligible cases is likely to be more expensive than offering monetary incentives of equal or greater value to
cases that first refuse participation. Group 2, where an incentive is promised to everyone with age-eligible
children, will be compared to the Groups 3-6 (individually and collectively) to measure the effects that
refusal-based incentives have on participation and cost.
Question 3. Do promised incentives increase participation?
This question can be answered by comparing the final participation rates for Groups 1 and 2. It can also be
answered by comparing the interim participation rates for Groups 3A and 5A (collectively) with the same rate
for Groups 3B, 4, 5B, and 6 (collectively). This interim participation rate would be calculated as if no further
contact attempts were made following a second refusal.
Question 4. Do promised incentives with prepaid incentives increase participation relative to
prepaid incentives alone?
The design also permits us to determine the boost in response rates (if any) due to promised incentives that
accompany prepaid incentives. Participation rates for Groups 3B and 5B (collectively) will be compared to
participation rates for Groups 4 and 6 (collectively). The comparison will be limited to those cases with
known addresses, as these are the only cases that will be sent the prepaid incentive.
Question 5. Do prepaid incentives increase participation?
By comparing known-address cases in Groups 4 and 6 (collectively) to known-address cases in Group 1, we
can evaluate whether prepaid incentives after the second refusal increase participation. We would have
preferred to be able to examine this question by offering prepaid incentives following the first refusal, but as
noted earlier, prepaid incentives cannot be mailed systematically after first refusals because some respondents
refuse participation before reaching the NS-CSHCN interview.
Question 6. Is a promised incentive after the first refusal more cost-effective than offering a larger
incentive only after the second refusal?
By comparing participation rates and cost between Groups 3A and 5A (collectively) and Groups 3B and 5B
(collectively), we can evaluate whether including a promised offer of $10 after the first refusal is more costeffective than only offering $11/$15 ($1/$5 prepaid with $10 promise) after the second refusal.
Question 7. Does the value of the prepaid incentive matter?
Our final goal is to evaluate whether a prepaid $1 coin provides the same boost in participation rates as a
prepaid $5 bill. This can be evaluated by (individually and collectively) comparing known-address cases in
Groups 5A, 5B, and 6 (the $1 groups) with known-address cases in Groups 3A, 3B, and 4 (the $5 groups).
In conclusion, answering these seven questions will allow NCHS and NORC to determine the best incentive
model for the remainder of the 2009-2010 NS-CSHCN considering the impact on response rates and costs.
Page 2