Sampling Plan

Foodborne Outbreak Prevention

Attachment A: Focus Group and Hybrid Group Sample and Recruitment

1. Overview

Each year, fourteen foodborne pathogens cost the American public 61,000 quality-adjusted life years and $14 billion dollars in economic loss (Hoffmann, Batz, and Morris, 2012). Once a foodborne outbreak of enteric bacteria has occurred, consumers need the ability to find and understand reliable information to avoid joining one of the 48 million Americans that become sick from a foodborne illness each year (CDC, 2018).

The Outbreak Response and Prevention Branch (ORPB) reduces additional incidences of foodborne illnesses by educating the public via web announcements with current and accurate information on the outbreak so that consumers, retailers, media, and medical professionals can act accordingly. Though the webpage metrics indicate high viewership of these messages since their inception in 2006, the format and content of these webpages has not undergone rigorous message testing to determine the level of understanding and behavioral responses from the general public.

EurekaFacts will employ a mixed-mode approach that combines one-on-one cognitive interviewing with focus groups to derive qualitative data at the individual and group level that will help direct ORPB’s communication efforts.

As part of this work, CDC has contracted Eureka Facts to conduct four focus groups and four groups of a cognitive interviewing-focus group hybrid to investigate the knowledge, attitudes, and behaviors of the general population as well as older adults regarding food safety and foodborne illness. These efforts will support the refinement and improvement of appropriate messaging to strengthen efficacy of public service announcements for foodborne illness.

2. Study Population

For this effort, EurekaFacts will recruit an array of diverse adults of all ages for the general population groups, including a mix of:

• Socioeconomic status (Low, Medium, High SES)

• Educational attainment levels

• Gender

• Ages (18-75 years old)

• Urbanicity (urban, rural, suburban residency)

• Adults with and without children

EurekaFacts will also recruit a subset focusing on adults aged 65 years and older to include a population at greater risk of foodborne illness.

3. Sample Frame

This information collection will employ a non-probability sample. With the limitation of generalizability that comes with qualitative research we propose a quota sampling method to make sure various demographics are represented in the sample of respondents.

In the United States, 2017 census data estimates indicate 6,216,589 people living in the Washington-Arlington-Alexandria, DC-Virginia-Maryland-West Virginia Metro Area (US Census Bureau, 2017). The EurekaFacts participant database includes tens of thousands of individuals in the DC Metro area and is constantly refreshed and updated through our independent participant outreach methods in both English and Spanish. These efforts ensure that the research participants recruited by EurekaFacts are not “professional focus group respondents,” instead they are individuals with limited to no experience with qualitative research who can provide fresh and actionable information for the CDC effort.

The sample selection will include members of EurekaFacts’ database of individuals and households in the desired geographic location. The EurekaFacts team will recruit a sufficient number of participants to complete four focus groups as well as four hybrid groups combining the cognitive interviewing and focus group methodology. General population participants will be recruited to achieve a mix of age, race/ethnicity, education, socioeconomic background across all groups, while the groups of older adult participants will include a subset of the general population sample with participants aged 65 years and above. The participants will be recruited from the Washington-Arlington-Alexandria, DC-Virginia-Maryland-West Virginia Metro Area. Exclusion criteria include the following:

• Having participated in a focus group within the past six months

• Currently employed in any of the following fields: Risk communication, market research, advertising, website design, food safety, restaurant management, public health, or healthcare practitioner, cheesemaker or cheese monger, or a leafy green produce grower, distributor, or seller

• Having experienced a severe illness requiring hospitalization that was attributed by a healthcare provider to consumption of contaminated food within the past six months

Additionally, participants who are related to or in a close relationship with another participant will be placed in separate groups.

4. Sample Size Calculations

The information collection targeting consumers will involve only qualitative methods, specifically of cognitive interviews and focus groups, which are not intended to yield results which can be generalized to the overall population. Results of this research will not be used to make statements representative of the universe of study, to produce statistical descriptions, or to generalize the information beyond the scope of the sample

This information collection will employ a non-probability sample. The participants will be selected using quota sampling to ensure that the sample includes enough individuals with particular characteristics to ensure representation of the target audience.

The four focus groups and four hybrid groups will include participants from the general population as well as older adults. The 54 participants will be divided into these groups such that general population and older adults will be included in both the focus groups and the hybrid groups.

	Area of investigation 1: Foodborne Illness Knowledge and Practices	Area of Investigation 2: Risk Communication Preferences
Focus groups
General population	1 group of 9 participants	1 group of 9 participants
General population	1 group of 9 participants	1 group of 9 participants
Hybrid groups
General population	1 hybrid group (4-5 participants)	1 hybrid group (4-5 participants)
Older adults	1 hybrid group (4-5 participants)	1 hybrid group (4-5 participants)
Total number of participants	27	27

5. Sampling Methods

In order to meet the sample requirements identified by the client and maximize effectiveness of the recruitment, EurekaFacts will utilize a multi-pronged recruitment approach which will include three main methods: 1) recruitment using our EurekaFacts Panel in the desired geographical region, 2) general advertisement method, and 3) individual/direct recruitment with multiple modes such as individual emails and telephone recruiting. More specific methods and resources will be used are as follows:

1. EurekaFacts panel recruitment:

1. Contact members of our database of individuals and households in the desired geographic location, offering a brief online screener to compel interest in participating;

2. General advertisements method:

1. Use of classified advertisements or more general announcements to appear in online and print versions of newspapers.

3. Direct recruitment via multiple contact methods:

3. Referrals from personal networks at the time of responding to general appeals and after completing interviews.

To maximize the effectiveness of recruitment and data collection, EurekaFacts has carefully designed screening questions that identify the characteristics of interest (e.g. age, gender, race and ethnicity, regions, etc.), and at minimum, contact information including first name, phone number, email address, and zip code. The screener will establish targets for demographic groupings based on age, race and ethnicity, and education level by regions.

The recruitment plan will address the strategies for ensuring that recruiting procedures are implemented as consistently as possible across all entities involved in recruiting. This will involve project specific training of the recruitment staff, consistent screening procedure across all location and recruiting organization, and use of such tools as programmed online screener, and Interview Progress Tracking tool and other tracking resources together with daily or more frequent communication between EurekaFacts and each of the recruiting locations.

Potential participants who have initially qualified via online screener will be further screened using a pre-approved screener script and questionnaire to be programmed into a computer-assisted telephone interview (CATI) software (Verint) to ensure that the screening procedure is uniformly conducted, instantly quantifiable throughout the recruitment effort, and that qualified and screened individuals who fit the criteria are then scheduled for the focus group or hybrid group. The recruiting team will send out a confirmation email and letter that includes the date, time, and location of the focus group or hybrid group, along with a map anddirections. Participants will receive a telephone reminder at least 24 hours prior to their appointment to confirm participation as well as to field any questions participant may have regarding their participation, and if they opt to receive text messages, we will send a text message reminder of their appointment on the day of their scheduled group.

6. Incentives

EurekaFacts anticipates providing an honorarium to each participant of $75. These incentives will be provided by EurekaFacts interviewers through the provision of gift cards upon completion of the focus group. Provision of the honorarium will take place after the completion of the 100-minute group session. The sessions are not expected to exceed 100 minutes.

7. Strategies for Sample Retention and Nonresponse Bias Analysis

To reduce the number of no-shows, scheduled group participants will be sent a reminder letter and/or telephone call providing the time of the group and directions to the location. As described in Part A of the supporting statement, participants will receive $75 incentive which will be provided to participants in appreciation for their completion of the interview. As previously noted, the proposed information collected will be qualitative; these methods to deal with non-response are adequate for this purpose.

Response rate is not applicable to quota sampling because this type of sampling results in a non-probability sample which is not representative of the population. Over-recruiting will be used to compensate for non-respondents.

References

U.S. Census Bureau (2017). American Community Survey 1-year estimates. Retrieved from Census Reporter Profile page for United States https://censusreporter.org/profiles/01000US-united-states/. Retrieved on November 14, 2018.