OMB Justification Memo

Memorandum

Date: October 24, 2016

To: Steph Tatham, OMB Desk Officer, Food and Nutrition Service

Through: Rachelle Ragland Green, Food and Nutrition Service, Information Clearance Officer; Ruth Brown OCIO Desk Officer

From: Jessica Larson, Public Affairs Specialist, Center for Nutrition Policy and Promotion (CNPP)

USDA – Food, Nutrition and Consumer Services

Re: Under Approved Generic OMB Clearance No. 0584-0523 Request for Approval to Perform Formative Research to Test Messages Developed to Help Communicate 2015 Dietary Guidelines Recommendations

Center for Nutrition Policy and Promotion (CNPP) of the USDA Food, Nutrition, and Consumer Services is requesting approval for formative research under Approved Generic OMB Clearance No. 0584-0523.

This request is to acquire clearance to conduct formative research (web-based survey) in order to further refine dissemination strategies to communicate recommendations stemming from the Dietary Guidelines for Americans 2015.

The following information is provided for your review:

1. Title of the Project: CNPP Segmentation Study

2. OMB Control Number and Expiration Date: 0584-0523, January 31,2016

3. Public Affected by this Project: Individuals and Households. The formative research will be conducted with Americans over the ages of 18 and will include both English- and Spanish-speaking individuals. See section 7, Project Purpose, Methodology & Formative Research Design for a description of the number of participants necessary for the research methodology. The definitions of the public affected by this project are identified in the screening instruments (see Attachments B-1 and B-2).

4. Number of Respondents:

CNPP, along with our contractors, will contact approximately 3,625 individuals and households to recruit:

• 2,700 participants completing the survey in English or Spanish.

  • 300 will take a web-based survey in Spanish

  • 2400 will take a web-based survey in English

Among the 3,625 individuals initially contacted, we estimate that 600 (19%) will not participate in any phase of research, whereas 2,950 (80%) will go on to participate in the screener. We further estimate that among 2,950 participants who would respond to a screener, 250 (8%) will not be eligible to complete the survey and 2,700 (92%) will be eligible to complete the survey. Since the consent form and the Privacy Act statement will be presented prior to the screener, the participants who would respond to the screener (2,950) would also submit a consent form and privacy statement.

Table 1: Number of respondents

Instrument	Total Number of Participants	Number of respondents	Proportion of Responders	Number of non-respondents	Proportion of non-respondents
Invitation
Invitation for English Language Survey	3200	2600	81%	600	19%
Invitation for Spanish Language Survey	425	350	80%	75	20%
Consent
English Language Survey	2600	2400	100%	0	0%
Spanish Language Survey	350	350	100%	0	0%
Privacy Statement
English Language Survey	2600	2400	100%	0	0%
Spanish Language Survey	350	350	100%	0	0%
Screener
English Language Survey	2600	2400	100%	0	0%
Spanish Language Survey	350	350	100%	0	0%
Survey
English Language Survey	2600	2400	92%	200	8%
Spanish Language Survey	350	300	86%	50	14%
	3625	2,950		675

5. Time Needed Per Response:

Participants will engage in a 15 minute web-based survey addressing topics outlined in the survey instrument (Attachments C-1 and C-2). We estimate that it will take respondents 2 minutes to read the invitation (Attachment A-1 and A-2) to the survey. Additional 5 minutes will be required for the screening instrument (Attachments B-1 and B-2), as well as 2 minutes for acknowledging consent (Attachment D-1 and D-2) and 2 minutes for reading the Privacy Act statement (Attachment F-1 and F-2). The total burden per respondent, including the invitation, is 26 minutes.

The table summary of estimated burden hours is presented below.

Table 2: Burden Hours

6. Total Burden Hours on Public: 1,075.71

7. Project Purpose, Methodology, and Formative Research Design:

Background

The U.S. Departments of Agriculture and of Health and Human Services recently released the 2015-2020 Dietary Guidelines for Americans policy document, which is the cornerstone of Federal nutrition programs. Updated every five years, the Dietary Guidelines for Americans (DGA) serve as the nation’s authoritative recommendations regarding making informed eating choices, and being physically active in order to maintain a healthy weight, reduce risk of chronic disease, and promote overall health. Recommendations from the DGA are intended for Americans who are aged two years and older, including those at risk of chronic disease.

The USDA Center for Nutritional Policy and Promotion (CNPP) is collaborating with HHS’s Office of Disease Prevention and Health Promotion (ODPHP) to develop a communication strategy to promote healthy eating and physical activity among Americans based on the 2015 DGA.

In 2015, CNPP and ODPHP determined that, in addition to communicating DGA recommendations, messaging must also communicate the notion of holistic dietary patterns—the idea that consumers should use the information from all of the DGA recommendations to help them make better nutrition decisions now, and into the future. To this end, CNPP developed messages encouraging “Quick wins” in order to achieve a healthier eating routine, and provided online tools and resources to help Americans to reach these goals.

The success of this communication strategy, however, depends on reaching the right audience with the right messages and the right tools. Traditionally market research to understand audiences has considered factors other than simple demographics, including psychographics, attitudinal and behavioral characteristics of the audience. The CNPP wants to explore these elements in order to more effectively and efficiently tailor messages to the American public. Specifically they want to understand how the conscious and subconscious Influences that impact decision making around food differ by audience.

This information collection is necessary for CNPP to meet the following objectives:

• Develop more nuanced and detailed audience profiles that are based on the tangible (i.e. price, taste, health content, “food rules”) and intangible (emotional and physiologic responses) attributes that shape nutritional choices

• As with previous Dietary Guidelines-related research, this information collection must also identify how best to communicate new message concepts to encourage the adoption of healthy food and beverage consumption decisions in view of the existing decision making profiles.

Purpose

The body of literature around behavioral economics and nutritional decision making suggests that people rely on both habit and heuristics and are subject to environmental, emotional, physiological, and social cues when making decisions around what foods to purchase and to eat. Understanding how individuals may differ in terms of the different influences or attributes that shape nutritional choices will help CNPP better target its communication strategies and tools to help Americans eat healthier. Specifically, the purpose of this survey is to understand how individual Americans potentially differ on the tangible (i.e. price, taste, health content, “food rules”), intangible (emotional and physiologic responses) and contextual (social cues, characteristics of the decision space) attributes that impact decisions around which foods to purchase and consume.

This project began with a qualitative exploration, through online journaling and focus groups, of eating habits and decision making around food. These previous formative research activities were approved under generic clearance 0584-0523. In aggregate, these findings show that food decisions are generally the result of the interplay among entrenched preferences, as well our emotional and physiological states, social cues, our reliance on certain heuristics, and the “rules” we put in place around eating. These factors differ person to person, depending on who we are, and our often heuristic driven perceptions around healthy heating, and are highly context-dependent. For example where we make these decisions and with whom we make these decisions are highly influential.

Although our qualitative research yielded rich data on the various factors that shape food decisions, the limitation of qualitative research is that it cannot accurately capture covariation between the different attributes that shape food decisions. A web-based study (attachment C-1 & C-2) is therefore necessary to investigate how the specific attributes that shape food decisions are related to each other and what additional behavioral, attitudinal, demographic and psychographic characteristics, like health orientation, may be related to decision making profiles.

The information collected during this final phase of the research will provide CNPP’s research and communication teams with an understanding of:

• Clusters of attributes that shape food decisions

• Relationships of these clusters to psychographic and demographic characteristics of the audience

These research activities directly meet the Secretary of Agriculture’s directive to develop innovative approaches to inform Americans about 2015 Dietary Guidelines and positively impact the health of all Americans.

These research activities are also aligned with President Obama’s 2015 executive order mandating that the Federal Government use behavioral science insights to better support a range of national priorities, including “enabling Americans to lead longer, healthier lives.” As stipulated under Section 1(b), USDA’s CNPP will use these behavioral insights to “improve how information is presented to consumers.”

Methodology/Research Design

Sampling

Participants will be recruited from an existing online panel. For the purposes of this study we will use Gfk Knowledge Panel. The Gfk Knowledge Panel was selected because it utilizes probability-based sampling to recruit participants (adults 18 and over), and thus better reflects the American population than opt in panels. KnowledgePanel does not accept self-selected volunteers as part of the KnowledgePanel. Instead, KnowledgePanel is based on a household sampling frame which recruits households: With unlisted telephone numbers, without landline telephones, that are cell phone only, without current Internet access, and/or without devices to access the Internet. To ensure participation of lower income and elderly individuals, individuals who do not have a computer and internet access to participate in research are provided with a laptop and monthly subscription to internet services in their home. This determination is made by Gfk at the time of recruitment using internal proprietary guidelines. Provision of the equipment and internet access are part of panel recruitment process and does not directly pertain to this or any other data collection that utilizes Gfk Knowledge Panel.

Two samples will be used:

• English-speaking participants

• Oversample of Spanish-speaking participants (both Spanish-dominant and bilingual)

The primary sampling criteria is the age of participants (over 18). The English language sample will be balanced on age, geographic region, gender, income, and race for inclusiveness. The Spanish language sample will be balanced on age, geographic region, gender and income.

The sampling strategy intends to obtain a sample of participants to achieve sufficient statistical power to:

• Assess covariation between attributes influencing food choices to develop audience profiles

• Relate obtained profiles to psychographic and demographic characteristics.

It is not the intent of the sampling strategy to be able to estimate any population parameters. Therefore, the sampling approach represents the need to understand the covariance between different influences on food choices and related characteristics as well as ensure inclusiveness of different demographic and sociographic groups

Recruitment

We anticipate to begin this study within 4 weeks of final OMB approval. Invitations to participate in the survey (see Attachment A1 and A2) will be sent only to participants who fit the study criteria:

• Adults above age 18 for general population

• Adults over 18 and who indicate Hispanic ethnicity and at least conversational competence in Spanish for Hispanic oversample.

This initial pre-screening of participants will be based on panel member characteristics (demographics such as age and race/ethnicity) kept by the Gfk that were collected at the time participants joined the panel. The pre-screening of the participants is conducted using these existing electronic records and therefore does not impose additional burden on the respondents.

Screening Criteria

After participants agree to participate and electronically sign the consent form (Attachment D1 and D2) for the study (see section 8). A consent form and privacy statement will be presented as a first screen for the study and will coincide with the time frame for the initiation of the research project (within 4 weeks following final OMB approval. A brief 5 minute screener will be administered (Attachments B1 and B2) online to participants who consent to the study to determine if the respondents meet the eligibility criteria. The screener will:

• Confirm age criteria for study participation

• Screen out participants who work in nutrition-related fields (e.g., health care providers, dietitians, those working in the wellness industry) and participants who work in marketing, advertising or market research.

• For Hispanic participants, measure the level of acculturation. Hispanic participants who were recruited as part of the English Language sample and were determined to be bicultural or Spanish-dominant will be reassigned to the Spanish language sample. Hispanic participants who were recruited as a part of the Spanish language sample and were determined to be acculturated will be reassigned to the English language sample.

Web experiment

After completing the screener, participants who qualify will proceed to the main survey (Attachment C-1 and C-2) in the appropriate language. The survey will contain several modules described below:

Attitudes and Knowledge

This module will assess participant attitudes and knowledge around food and healthy and unhealthy eating, as well as explore the “food rules” that participants may or may not have in place around eating and mealtimes.

Food Decision Making

Participants will be asked to reflect on recent food purchases as well as healthy and unhealthy foods or meals they have recently consumed in order to explore the heuristics and emotional, physiological and environmental factors that shape these decisions.

Psychographics

The module will administer the measures of attitudes and behaviors that may impact receptivity to CNPP messages, including:

• Assessment of healthy habits and role in food purchasing decisions

• Assessment of their own eating patterns

• Self-efficacy to engage in healthier eating behaviors

Demographics

The final module will collect demographic measures including income, educational attainment and household size. Participants’ height and weight information will also be collected in order to calculate BMI.

Incentives

It is customary for panel providers to offer points for participation in the surveys. Typically, the value of the points offered to participate in the survey are less than one dollar. Accumulated points over many surveys can typically be traded for small monetary gifts, merchandise, or entry into a sweepstakes. Panel providers typically have two types of sweepstakes: those for which all panel member are eligible for and have equal chance of winning, and lottery sweepstakes for which panel members can trade in their points for extra chance to win (similar to buying lottery tickets for State lotteries). Point systems are part of the numerous other panel management activities conducted by the panel providers in order to keep panelists engaged. However, the price of the sample mostly depends on incidence rates and therefore does directly correlate to the panel management cost incurred by the sample provider. Thus, when using an online sample, the Federal government does not directly provide funding for the incentives to the participants, rather it provides funding for access to the sample that may or may not be incentivized by the panel provider.

8. Privacy:

CNPP and contractors will follow procedures for maintaining safeguards consistent with the Privacy Act during all stages of data collection. A system of record notice (SORN) titled FNS-8 USDA/FNS Studies and Reports in the Federal Register on March 31, 2000, Volume 65, Number 63, and is located on pages 17251-17252 discusses the terms of protections that will be provided to respondents.

Using the consent forms in Attachments D-1 and D-2, all participants will receive a privacy statement Appendix F-1 and F-2 along with the consent form. The privacy statement will inform the participants that all opinions they provide will be private and none of their responses will be associated with personally identifiable information (PII) or release by anyone, except as otherwise required by law. Participants will also be informed that findings will be presented as a summary. The consent form and Privacy Act statement will be the first screen on the online survey and require participants’ electronic acknowledgement prior to proceeding to the rest of the survey.

Additional steps employed to notify the respondents about their privacy are as follows:

• As part of the conditions of employment, all trained staff working on the project have signed a non-disclosure agreement (NDA) (Attachment E-1) prohibiting them from discussing any information about the research studies they work on with anyone not involved with the project, especially discussing or releasing personally identifiable information (PII).

• At the completion of the study, study files stripped of PII information will be transmitted to CNPP. The contractor will keep the files for the period agreed on with CNPP. Following that period, files will be purged from the contractor IT systems. The procedure employed includes multiple over writes of the study file both on the server and any backup media to assure that the files are not recoverable.

9. Federal Costs: $177,706  for data collection and analyses/reporting

10. Research Tools/Instruments:

• Attachment A-1: Survey Invitation (English)

• Attachment A-2: Survey Invitation (Spanish)

• Attachment B-1: USDA Segmentation Survey Screener (English)

• Attachment B-2: USDA Segmentation Survey Screener (Spanish)

• Attachment C-1: USDA Segmentation Survey Instrument (English)

• Attachment C-2: USDA Segmentation Survey Instrument (Spanish)

• Attachment D-1: Consent form (English)

• Attachment D-2: Consent form (Spanish)

• Attachment E-1: Trained Contractor Privacy Agreement

• Attachment F-1: Privacy Act Statement (English)

• Attachment F-2: Privacy Act Statement (Spanish)

3