Upload
spencer-cross
View
215
Download
0
Embed Size (px)
Citation preview
Slide 1
Incentives in Surveys with Incentives in Surveys with FarmersFarmers
Third International Conference on Establishment SurveysMontreal, Quebec, Canada
June 19, 2007
Slide
Kathy Ott and Dan Becklerpresented by Norman Bennett
USDA, National Agricultural Statistics Service
Background Information - NASS
The National Agricultural Statistics Service (NASS) surveys Agricultural Operations (farmers/ranchers)
Collects information from the US agricultural sector on acreage, production, yield, economics, labor, etc.
Conducts over 400 surveys per year
Technically, these are establishment surveys, but have some similarities to household surveys
Slide 2
Background Information - ARMS
Agricultural Resource Management Survey (ARMS) is a long data collection, typically conducted by face-to-face interview
Asks detailed information on what is produced, economic items (assets, debt, income, expenses), and operator characteristics
Average interview time = 90 minutes
Typical national response rate = 60 to 65%
Very complex survey design with multiple versions. For this project, just two questionnaire versions used
Slide 3
Incentive Use
Goals for ARMS:• Raise overall response rate• Contain data collection costs• “Shift” to mail completion
Two Incentive Studies completed at NASS; both on the ARMS.
• 2004 Mail-out/Mail-back Study• 2005 Face to Face Interview Study
Slide 4
2004 Mail-out/Mail-back Study
2004 Reference Year, conducted early 2005
“Core” Questionnaire Version
Administered in 15 states, sample: ≈16,000
Mail-out/Mail-back data collection with face-to-face interview follow-up for all nonrespondents
Pre-Survey Letter First Questionnaire Mailing Postcard Reminder/Thank You Second QuestionnaireFace-to-face Interview follow-up on all mail nonrespondents
Slide 6
2004 Mail-out/Mail-back Study
Used a “monetary” incentive in the form of a $20 ATM card
Administered by: 1st class and priority mail Prepaid and promised
Treatment Groups receiving incentives included cover letters that:
Explained the incentive as a “Thank You” Described uniqueness of ARMS Justified incentive by overall cost-savings to government
Slide 7
ATM Cards
Cards loaded with $20 plus $4 for transaction fees
ATM Cards came pre-activated
ATM Card was affixed to an instruction sheet
Slide 8
Treatment Group
1st Questionnaire Mailing
2nd Questionnaire Mailing
1 (Control) 1st Class, no incentive 1st Class, no incentive
2 (1st, Prepaid) 1st Class, $20 ATM 1st Class, no incentive
3 (Priority) Priority, no incentive Priority, no incentive
4 (Priority, Prepaid) Priority, $20 ATM Priority, no incentive
5 (1st, Promised) 1st Class, promise $20 ATM
1st Class, promise $20 ATM
All groups also received a pre-survey letter, a post-card reminder, and face-to-face follow-up
2004 Treatment Groups
Slide 9
2004 Response Rate Results
Treatment Group N Mail ReturnsFace-to-Face Completes
Overall Completes
1 (Control) 1,948 30.1% 33.3% 63.4%
2 (1st, Prepaid) 1,941 40.8% 29.6% 70.4%
3 (Priority) 1,935 32.9% 31.8% 64.7%
4 (Priority, Prepaid) 1,952 43.9% 28.5% 72.4%
5 (1st, Promised) 1,946 37.2% 31.2% 68.4%
1,3
1,3,5
1,3
1,3
1,3
1
Red numbers indicate Treatment Groups to which this percent was significantly greater (at α = 0.05 from t-tests with Bonferroni adjustment).
Slide 10
2004 ATM Card Usage
Treatment GroupMail
RespondentsOverall
RespondentsOverall
NonrespondentsOverall Card Recipients
2 (1st, Prepaid) 60.4% 41.1% 4.7% 30.3%
4 (Priority, Prepaid) 58.9% 40.8% 5.4% 31.0%
5 (1st, Promised) 71.4% 61.4% 3.1% 61.3%
These people initially returned questionnaires which were later deemed inadequately completed.
Slide 11
2004 Overall Costs
Treatment Group
Postage & NPC Costs
All ATM Costs
Face- to- Face Costs
TotalAverage
per Sample
Average per
Complete
1 (Control) $11,520 -0- $176,291 $187,811 $96.41 $152.07
2 (1st, Prepaid)
$13,128 $14,845 $154,154 $182,127 $93.83 $133.33
3 (Priority) $21,722 -0- $170,951 $192,673 $99.57 $154.02
4 (Priority, Prepaid)
$23,250 $15,342 $148,356 $186,948 $95.77 $132.21
5 (1st, Promised)
$13,939 $20,486 $160,475 $194,900 $100.15 $146.32
Face-to-face follow-up costs.
Includes all ATM/POS withdrawals, all fees, and administrative costs.
Includes postage, printing, and NPC administrative fees.
Slide 12
Summary of 2004 Mail-out/Mail-back Study
Offering $20 ATM card incentives for the ARMS Core improved both mail and overall response rates
The incentives reduced overall costs, the average cost per sample and the average cost per complete.
The prepaid incentives performed better than the promised.
Priority mail further increased response rates (albeit non-significantly) when combined with the incentive but alone was ineffective.
Slide 13
What to do next?
Implemented use of prepaid incentives for entire Core version mail-out/mail-back sample (approximately 16,000 records)
For 2005, extended incentive research for face-to-face interview component of the data collection using the Cost and Returns Report (long) version of the ARMS questionnaire
Slide 14
Slide 15
2005 Face-to-Face Incentive Study
“Long” (32 pages) questionnaire version only
All 48 states included
Face-to-face interview only, no mail data collection
Presurvey letter was mailed with the incentive, followed by face-to-face interview data collection
Stimuli:Prepaid Monetary Incentive ($20 ATM card)“Prepaid” Non-monetary Incentive (Wall clock)Promised Individual Farm Analysis (IFA)
Treatment Group
First Contact Interview Contact
Pre-survey letter, plus: Face-to-face interview, plus:
1 (Control) No incentive No Incentive
2 (ATM Card) $20 ATM Card No Incentive
3 (IFA) Mention IFA Promise IFA
4 (Clock) Non-monetary clock No Incentive
5 (ATM Card, IFA)$20 ATM Card
Mention IFAPromise IFA
2005 Treatment Groups
Slide 16
Slide 20
Summary of 2005 Face-to-Face Study
As implemented, Prepaid Monetary, “Prepaid” Non-monetary, and Individual Farm Analyses Incentives:
Had no significant impact on response rates, although response rates did go up slightly
Increased costs
It’s possible that a different method of distributing a non-monetary incentive or a different non-monetary incentive all together would have been more effective
Slide 21
Recommendations
Continue the use of monetary incentives for ARMS “Core” mail-out/mail-back data collection methodology and include a control group at least every other year
Do not use prepaid monetary or “prepaid” non-monetary incentives for face-to-face data collections as a means of increasing response rates
Track the card cashing rates of respondents and nonrespondents every year to maintain cost effectiveness of incentive methodology for mail data collection – part of studying the long-term effects on NASS’ entire survey program.
Contacts:
Dan Beckler: [email protected]
Kathy Ott: [email protected]
Norman Bennett: [email protected]
Slide 23
References:
Beckler, D. K. Ott, P. Horvath. (2005). Indirect Monetary Incentives for the 2004 ARMS Phase III Core. Research and Development Division Research Report RDD-05-05. National Agricultural Statistics Service, United States Department of Agriculture.
Beckler, D. K. Ott (2005). Indirect Monetary Incentives With a Complex Agricultural Establishment Survey. Proceedings of the Joint Statistical Meetings of the American Statistical Association.
McCarthy, J.S. D. Beckler, K. Ott (2006). The Effect of Incentives on Response in 2005 ARMS Phase III Interviews. Research and Development Division Research Report RDD-06-07. National Agricultural Statistics Service, United States Department of Agriculture.
Slide 24