Generic ICR
Data Collection Through Web Based Surveys for Evaluating Act Against AIDS Social Marketing Campaign Phases Targeting Consumers
Supporting Statement B
Reinstatement with Change
OMB No. 0920-0920
October 29, 2018
Contact Person:
Euna August, PhD
Division for HIV/AIDS Prevention
Centers for Disease Control and Prevention
1600 Clifton Rd. NE
Mailstop E-49
Atlanta, GA 30329
Telephone: (404) 639-8297
Fax: (404) 639-2007
E-mail: [email protected]
Table of Contents
B. Collections of Information Employing Statistical Methods 1
B.1 Respondent Universe and Sampling Methods 1
B.2 Procedures for the Collection of Information 2
B.3 Methods to Maximize Response Rates and Deal with Nonresponse 6
B.4 Test of Procedures or Methods to Be Undertaken 7
B.5 Individuals Consulted on Statistical Aspects and Individuals Collecting and/or Analyzing Data 7
Exhibits
Number Page
The sample for each cross-sectional survey will include defined AAA target audiences recruited through multiple mechanisms. Although the eligibility criteria for each survey are unknown at this time, at a minimum, all respondents will be aged 18 to 64. We estimate conducting two to three surveys per year for an average of 850 to 1,075 respondents per survey. Most respondents will be sampled and recruited by online survey panel vendors.
The online survey panel vendors use a range of proprietary sources to obtain their samples. These samples may be biased because they are not randomly selected; people who opt to participate in a survey panel may have different characteristics than those who do not. The respondents for this study should represent some, but not all, of the campaign’s target audience because the sample is extremely targeted. Some target audiences may also have more limited Internet access or differ in other ways when compared with the populations of interest. Depending on the nature of the campaign phase or messages to be tested, we will utilize varied recruitment methods to mitigate some of these inherent biases. The screeners for each data collection will gather detailed information on respondent characteristics which will enable us to describe the sample and monitor the extent to which our recruitment methods are reaching the appropriate groups.
Obtaining a probability-based sample to reach the desired subpopulations of interest is cost prohibitive for this study. The target audiences for AAA campaign phases are varied and in most instances, will be narrowly defined (e.g., African Americans at high risk of HIV infection; gay, bisexual, and other men who have sex with men [collectively referred to as MSM]). Although the sample is not meant to be generalized to the entire target population, these information collections will enable CDC to gather information critical to ensuring the efficacy of the AAA initiative.
The sample size requested for this three-year Generic ICR is 6,445. Annualized over the 3-year period, this amounts to 2,148 respondents. The expected response rate is approximately 20%; thus, we will need to screen 10,740 respondents annually to achieve the desired sample size. Sample characteristics will vary depending on the audience for specific AAA campaign phases; therefore, the sampling strata for each collection are unknown. Prior data collections under this Generic ICR have stratified samples based on race and ethnicity, age, gender, HIV status, sexual orientation, and other demographic characteristics to ensure we gather input from a variety of subgroups within the target audience and that the sample size for each subgroup is sufficient to detect differences. It is anticipated that future data collections will be stratified similarly depending on the nature of the campaign. To reduce the effects of nonsampling error, nonresponse and post-stratification weighting adjustments will be applied to the sample when feasible.
For each data collection, we will conduct a power analysis to determine the optimal sample size for detecting statistically significant differences by campaign exposure level among targeted audiences. As an example, we conducted a power analysis to detect the percentage difference among people who were and were not exposed to a partner communication campaign where the outcome variable was communication with sexual partners about HIV. Based on prior data collections conducted under this Generic ICR, we anticipated that 30% of respondents were exposed to this campaign, and we assumed based on the literature that 65% of nonexposed respondents discussed their HIV status with their sex partners. Given a type I error rate of 0.05 (alpha = 0.05), we determined that we will achieve 80% power to detect that 74% of respondents in the exposed group discussed HIV status with their sexual partners with a sample size of 1,000.
Because these will be cross-sectional surveys, any differences in outcomes cannot be directly attributed to the campaign. We can, however, examine correlations between campaign exposure and the identified outcomes which will enable us to (1) investigate naturally occurring phenomena which would be impractical to assess experimentally and (2) determine whether a relationship exists between campaign exposure and the outcomes of interest.
This section describes procedures for recruitment, screening, and administering the survey.
Recruitment
Recruitment procedures will vary depending on the identified target audience. Recruitment through online survey panel vendors will be the primary method. They will identify individuals who fall into the target audience for the specific campaign being assessed using their proprietary research panels. This list will be supplemented with sample lists from external partners, including respondent lists from membership organizations (e.g., the National Urban League, the National Medical Association). The online survey panel vendor will send e-mail invitations (Attachment 5) to the sample list. Each invitation will contain the survey title, a brief description of the survey, the length of the survey, and token of appreciation amount provided for successful completion of the survey. To access the secure website for the study, they will be told to click on the URL provided in the invitation. The online survey panel vendors will send one email reminder about the survey to nonresponders requesting their participation (see reminder email in Attachment 5).
We will also recruit respondents through partnerships with Community Based Organizations (CBOs) that work with identified target audiences, such as AIDS service organizations, community health centers, and health departments. We will provide CBO partners with flyers and palm cards to advertise the study and ask them to place them in areas frequented by program clients (e.g., waiting areas, exam rooms). The recruitment materials will describe the survey topic, its length, and the token of appreciation amount and direct individuals to visit the study’s secure website if they would like to be screened for eligibility.
Finally, we will recruit subjects through the internet by posting online advertisements on targeted websites, including traditional banner and electronic bulletin board advertisements, and advertising the study on social media sites that cater to the targeted populations. Websites and social media outlets will be chosen based on the campaign’s focus. The advertisements will describe the survey topic, its length, and the token of appreciation amount. Individuals responding to electronic advertisements will be asked to click on the URL for the study’s website if they would like to be screened for eligibility.
Informed Consent
When individuals enter the study’s secure website, they will be presented with the consent form (Attachment 4). The consent for screening and the survey are combined and addresses the following:
The study’s topic and goals;
The procedures that will be involved, including the sensitive nature of some of the questions that will be asked;
Potential risks and discomforts associated with participation and the right to refuse or withdraw;
Benefits to participation;
Remuneration amount and form;
The measures that will be taken by CDC and the evaluation contractor to protect privacy as well as the measures respondents can take to protect their information (i.e., take the screener in a private location, such as their own home and/or in a room with a door, and close their browser window when they are finished or if they choose to withdraw); and
Contact information for the evaluation contractor’s project director if they have questions about the study and the contractor’s Office of Research Protection if they have questions or concerns about their rights as a study participant.
After reading this information, potential participants will be reminded of the voluntary nature of the study and that they can refuse to participate or stop participating without penalty. They will then be asked to select, “I have read this consent form and agree to participate in the survey” or “I have read this consent form and do not want to participate in the survey”. If an individual chooses not to participate, they will be thanked for their time and asked to close their browser window for privacy purposes. No additional contact will be made with individuals who select no. Individuals who do not meet the eligibility criteria for the survey will be told they are ineligible, thanked for their time, and asked to close their browser window for privacy purposes. The consent form as well as other study materials will be at an 8th grade reading level or below.
Screener and Survey
Potential participants will be asked to complete the two-minute screener (Attachment 6). If an individual is determined to be eligible, they will be routed directly to the survey. Individuals who are ineligible will be told they do not qualify for the survey, thanked for their time, and asked to close their browser window.
Study respondents will self-administer the 30-minute survey (Attachment 3) during a designated timeframe, making it feasible for respondents to complete the survey during their own time, in private. This data collection mode is well-suited to surveys covering sensitive topics (e.g., sexual behavior, HIV status); in an interviewer-administered survey, individuals may underreport perceived undesirable behaviors which could compromise the validity and reliability of the data. Respondents can complete each survey only once.
The survey questions will vary to some extent depending on the campaign phase and target audience. Survey questions will cover the following topics: Campaign exposure and receptivity; theoretical constructs, such as norms, self-efficacy, attitudes, behavioral beliefs, and behavioral intentions; communication behaviors, including information seeking and media use; HIV risk perceptions; and HIV prevention behaviors.
Respondents will have the option to take the survey in one or multiple sittings as long as they complete it during the designated timeframe. When a respondent reopens the survey using the original URL, they will reenter the survey where they left off; in such cases, they will be unable to go back to questions they already answered. This measure will help to protect privacy by ensuring that individuals who inadvertently or purposefully access the survey cannot see a respondents’ answers.
Individuals who complete the survey will receive a token of appreciation of points redeemable for purchase (value of $20-$40). Adults are difficult to engage in a survey about sensitive topics without the use of a small token of appreciation. The token of appreciation is intended to recognize the time burden placed on respondents, encourage their cooperation, and convey appreciation for contributing to this important study.
When the data collection is complete, the evaluation contractor will decrypt and download the data from the online survey panel vendor into an SPSS data file. Online survey panel vendors will have no logical access to the data or a master key to decrypt the data. The evaluation contractor will analyze these data and summarize the findings in a report.
Our sample design is based on conservative assumptions about survey response. Thus, our estimates of cooperation rates should be viewed as “worst case” scenarios that, if they hold true, would still ensure sufficient sample sizes to reasonably detect small effects with multivariate models. We estimate that at least 20% of respondents will complete the survey. See Exhibit B.2.1 expected cooperation rates annually and for the three-year generic ICR period.
Exhibit B.2.1. Annual and Three-Year Total Number of Respondents for the Survey
Numbers and Cooperation Rates |
Annual Total |
Three-Year Total |
Number of respondents to be contacted |
10,740 |
32,220 |
Expected cooperation rate |
20% |
20% |
Number of completed surveys |
2,148 |
6,445 |
The following procedures will be used to maximize cooperation and to achieve the desired high response rates:
A token of appreciation will be provided to respondents upon completion of the survey.
Online survey panel vendors will send one email reminder to nonresponders.
Online survey panel vendors will provide toll-free telephone numbers to all respondents so that they can call with any questions or concerns about any aspect of the study. The consent form also provides toll-free telephone numbers for the evaluation contractor’s project director and Office of Research Protection should respondents have any questions about the study or their rights as a study respondent.
Respondents will have the option to complete the survey in one or multiple sittings which will allow them to take it at their own pace.
The contractor will work closely with the online survey panel vendors to monitor progress and troubleshoot issues related to sample recruitment.
To estimate the time burden for screening and the survey, two survey specialists will be consulted for each data collection. The survey specialists will test the time burden by providing affirmative responses to most or all questions that branch to follow-up questions. By testing the screener and survey in this manner, the burden estimate will approximate the maximum average burden because almost all survey questions will be asked. In addition, the survey specialists will deliberately read each item slowly to ensure that the time burden accounts for the inclusion of low-literacy respondents.
The survey specialists tested the sample screener and estimated the maximum average burden to be two minutes for the study screener (Attachment 6). The surveys will be designed so that respondents can complete them in 30 minutes or less (Attachment 3). We will submit the phase-specific screener and survey with each information collection request under this generic ICR.
To ensure that response burden is minimized for each data collection, the evaluation contractor and CDC staff will test the process from end-to-end. This will enable us to pilot the screener and survey programming and logic and correct any potential problems prior to implementation.
Euna August
CDC Behavioral Scientist
1600 Clifton Rd, NE
Atlanta, GA 30333
404-639-8297
Jennifer D. Uhrig
RTI International - Project Director
3040 Cornwallis Road
Research Triangle Park, NC 27709
919-316-3311
Carla Bann
RTI International - Statistician
3040 Cornwallis Road
Research Triangle Park, NC 27709
919-485-2773
| File Type | application/vnd.openxmlformats-officedocument.wordprocessingml.document |
| Author | snaauw |
| File Modified | 0000-00-00 |
| File Created | 2021-01-20 |