NFSG Update: Resuming Data Collection with a New Multi-mode Design

NSFG

OMB Attachment F

OMB No. 0920-0314

National Center for Health Statistics

National Survey of Family Growth Update: Resuming
Data Collection with a New Multi-mode Design
Anjani Chandra, Ph.D.
NSFG Team Lead & Principal Investigator
NCHS Board of Scientific Counselors Meeting
February 10, 2022

NSFG Team
At NCHS:
NSFG Team within Division of Vital Statistics/Reproductive Statistics Branch:
Anjani Chandra, Team Lead/Principal Investigator
Joyce Abma, Contract Officer’s Representative
Gladys Martinez
Kim Daniels
Colleen Nugent
Jennifer Sayers

Hee-Choon Shin, Mathematical Statistician, Division of Research & Methodology

At RTI:
Andy Peytchev, Principal Investigator
Susan Kinsey, Project Director
Emilia Peytcheva, Instrumentation and Methodology Reports Task Leader
Taylor Lewis, Sample Design, Weighting, and Imputation Task Leader
Lilia Filippenko, Systems Development, Data Processing, Documentation, and Delivery Task Leader
William Savage, Information Security and Confidentiality Task Leader
Milton Cahoon, Preparation for Data Collection and Data Collection Task Leader

Overview of presentation
 NSFG background
 Progress since last BSC presentation in January 2020
–
–
–
–

Closeout of previous contract and final data release (2017-2019 NSFG)
Award of new contract September 2020 and preparation for data collection launch
Instrument development work – including adaptations for online mode
Multi-mode survey design envisioned for Year 1 data collection

 Resumption of data collection in January 2022
–
–

Revisions made for Year 1, Quarter 1 due to COVID-19
Plans under review/consideration for Quarter 2+

 Feedback and discussion

NSFG Background

NSFG Purpose and Uses
 Address Section 306 of the Public Health Service Act, stipulating “NCHS
shall collect statistics on… family formation, growth, and dissolution”
 Help explain variations in birth rates using intermediate or “proximate”
determinants of fertility
 Assess factors that affect timing and consequences of sexual activity and
pregnancy
 Describe relationships and families
 Measure receipt of family planning and other medical services
 Monitor risk of HIV and STIs in general household population

NSFG Background (1973-2019)
 Each survey provides a nationally representative, cross-sectional snapshot
of the U.S. household “reproductive-age” population
 Key changes since “Cycle 1” in 1973:
–
–
–
–
–
–

1982 – Inclusion of never-married women
1988 & 1995 - linked to NHIS sampling frame from several years prior
1995 – Conversion to CAPI and ACASI and 1st use of incentives
2002 – Inclusion of men (independent sample) & expansion of ACASI
2006 – Transition to continuous fieldwork design
2015 – Expansion of age range from 15-44 to 15-49

 Public-use files have been released with every periodic survey 1973-2002
(Cycles 1-6) and roughly every 2 years since 2006 (under continuous
fieldwork design)

Year

Scope

N

Over Samples

Response Rates

Incentive

OMB Approved
Intvw Length

1973

Ever-Married Women
15-44

9,797

Black women

90.2%

None

60 min

1976

Ever-Married Women
15-44

8,611

Black women

82.7%

None

60 min

1982

All Women 15-44

7,969

Black women & teens 15-19

79.4%

None

60 min

1988

All Women 15-44

8,450

Black women

82.5%

None

70 min

1995

All Women 15-44

10,847

Black & Hispanic women

78.7%

$20

100 min

2002

All Women & Men

12,571
W = 7,643
M = 4,928

Black people, Hispanic people,
People 15-19

All - 79%
W – 80%
M – 78%

$40

W – 85 min
M – 60 min

2006-2010

All Women & Men 1544

22,682
W=12,279
M=10,403

Black people, Hispanic people,
People 15-19

All - 77%
W – 78%
M – 75%

$40

W – 85 min
M – 60 min

2011-2013

All Women & Men 1544

10,416
W=5,601
M=5,815

Black people, Hispanic people,
People 15-19

All – 72.8%
W – 73.4%
M – 72.1%

$40

2013-2015

All Women & Men 1544

10,205
W=5,699
M=4,506

Black people, Hispanic people,
People 15-19

All – 69.3%
W – 71.2%
M – 67.1%

$40

2015-2017

10,094
All Women & Men now
W=5,554
15-49
M=4,540

Black people, Hispanic people,
People 15-19

All – 65.3%
W – 66.7%
M – 63.6%

$40

W – 80 min
M – 60 min

2017-2019

All Women & Men 1549

11,347
W=6,141
M=5,206

Black people, Hispanic people,
People 15-19

All – 63.4%
W – 65.2%
M – 61.4%

$40

W – 80 min
M – 60 min

W – 80 min
M – 60 min
W – 80 min
M – 60 min

Current NSFG Cosponsors (all within DHHS)













National Center for Health Statistics (NCHS)
Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD)
Office of Population Affairs (OPA)
Administration for Children & Families (ACF)/ Children’s Bureau
Administration for Children & Families (ACF)/ Office of Planning, Research, & Evaluation (OPRE)
Office on Women’s Health
CDC/NCHHSTP/Division of HIV Prevention
CDC/NCHHSTP/Division of STD Prevention
CDC/NCHHSTP/Division of Adolescent & School Health
CDC/NCCDPHP/Division of Reproductive Health
CDC/NCCDPHP/Division of Cancer Prevention & Control
CDC/NCIPC/Division of Violence Prevention

Progress since last BSC update
in January 2020

Closeout of previous contract and final data release
(2017-2019 NSFG)
 Previous 10-year contract with University of Michigan ended Dec 2020
–
–

Final public-use files for 2017-2019 NSFG released in October 2020
Restricted-use files made available through Research Data Center (RDC):
• Contextual, interviewer observations, and paradata files
• Additional data, formerly included on PUF, but now only available in RDC due to disclosure risk

 Analyses with latest NSFG data:
–
–
–

Data Brief on current contraceptive status released with 2017-2019 PUF in October 2020
NSFG Key Statistics page updated with 2017-2019 data in November 2021
Several QuickStats with 2017-2019 data published in 2021, but most other analyses delayed due to
staff size and competing demands of new contract and 3 staff COVID deployments
• Teen report using data for 2015-2017
• Menarche report using data through 2017

Award of new contract to RTI and preparations for
Year 1 (2022) data collection
 New 10-year contract awarded to RTI in September 2020, with plans to
resume data collection with a multi-mode survey design in January 2022
 Oversight, review, and collaboration with contractor on:
Sample design and establishing quarterly sample areas
Interviewer training materials and other preparation for multi-mode data collection
Onboarding of contractor staff, hiring interviewers
Establishing IT systems that meet CDC security requirements
Modification of contract to move web survey mode earlier (to 1st year of data
collection) due to COVID
– Obtaining clearances for OMB and ERB
–
–
–
–
–

Instrument development work


Producing programmer-ready full specifications based on “capi-lite” questionnaires in RFP
–
–
–
–
–

Streamlining and restructuring of content to reduce complexity and respondent burden
Improvements to accommodate respondents in same-sex marriages and cohabitations
Incorporation of new items from cosponsors and revisions prompted by CCQDER work***
Adapting survey instruments and interview aids to work for both FTF and online modes
Translating new or revised items into Spanish




Developing web screener to manage multiple survey tasks accomplished in FTF mode
Developing an electronic life history calendar for female survey



Extensive instrument testing in English and Spanish

***Collaboration with CCQDER on selected instrument issues:
–
–

Cognitive interviews to assess Life History Calendar, as well as selected question series
RANDS panel (in progress) to do split-panel testing of specific questions or series

Survey design envisioned for Year 1
 Continuing to interview 1 person aged 15-49 per sample household
 Building off success of continuous FTF survey design used 2006-2019:
–

Responsive survey design using paradata in real time with subsampling for nonresponse allowed for
nonresponse bias mitigation with cost control

Prior FTF design
•
•

Current Multi-mode design

Fieldwork organized into 4 12-week quarters/year •
2-Phase design:
•
– Phase 1 (weeks 1-10):
 All sample worked
 $0 screener, $40 main incentive
– Phase 2 (weeks 11-12):
 30% subsample of non-respondents
 $5 prepaid screener; $40 prepaid for main,
and $40 at end for main

Fieldwork organized into 4 16-week quarters/year
3-Phase design:
– Phase 1 (weeks 1-4): Web
 $2 screener, $40 main incentive
– Phase 2 (weeks 5-12): Web & FTF
 No change in incentive
– Phase 3 (weeks 13-16): Web & FTF
 Up to 50% subsample of non-respondents
 $5 prepaid screener; $40 prepaid for main,
and $40 at end for main

Multi-mode, multi-phase survey design for Year 1:
4 overlapping quarters

Multi-mode Comparison Study (originally planned for
Q1/2, but shifted to Q3/4)

New Multi-mode Design:

Web (CAWI) (Phase 1)
Face to Face (and Web continued) (Phase 2)
Face to Face, increased incentives (Phase 3)
Mail eligibility or NRFU survey (Phase 4)

Face to Face, as used 2006-2019:

Face to Face (Phase 1)

Face to Face, increased incentives (Phase 2)

Resumption of data collection
in January 2022

Quarter 1 changes and challenges
 The surge in COVID-19 rates, national staffing shortages, and contractor
policy precluded FTF interviewing in most NSFG sample areas at start of
Q1; policy changed as of 2/3/22 to allow some FTF during Q1.
 As a result, for Quarter 1:
– Only CAWI data collection in Phases 1 and 2
– Field staff will conduct Phase 3 interviews with higher incentives, for a subsample of
non-respondents from Phases 1 & 2, as originally planned, and will follow up 100% of
breakoffs.
– Seeking OMB and ERB approval for further changes in Quarter 2

Changes being considered for Quarter 2+ (1 of 2)
 Mailed paper screener to augment CAWI screener
– Most nonresponse in HH surveys relying on mailed invitations to a web survey is at
screener stage. Even if FTF interviewing can resume to originally intended levels in Q2,
mailed screeners could reduce reliance on in-person follow-up.
– Paper screener would be mailed around week 2 (on 3rd mailing, approx. 1 week after 1st
mailing)
– Would only be used to select adult respondents, given complexity of parental
permission and minor assent process for teens 15-17
– We would evaluate impact on response rates, demographic composition, & cost
indicators.

 Conducting telephone prompting calls to complete the web main survey

Changes being considered for Quarter 2+ (2 of 2)
 Experiment to test accelerated delivery of Phase 3’s higher incentives
– Phase 3 protocol already approved by OMB & ERB:
• $5 prepaid incentive for screener; $40 prepaid incentive before main and $40 upon
completion
– Experimental condition would be assigned 50-50 at sample address level – half receiving
higher incentives at standard Phase 3 timing (weeks 13-16) and half receiving up front in
Phase 1.
– Higher incentives have potential to increase response rates and decrease nonresponse
bias
– Incentivizing participation from the beginning could be cost-efficient overall:
• In areas where FTF is not possible, can reduce # of mailings needed per case
• In areas where FTF is possible, can reduce # of sample addresses needing FTF
– Some increased challenge for field staff juggling different incentive protocols

Feedback and Discussion