Methodological Report for the Dartmouth-UNH Survey Center NH COVID-19 Study
The findings of the Dartmouth-UNH NH COVID-19 Study are based on responses from members of the Granite State Panel, a probability-based web panel constructed and maintained by the University of New Hampshire Survey Center.
Granite State Panel
All data are weighted by respondent sex, age, education, and region of the state to targets from the most recent American Community Survey (ACS) conducted by the U.S. Census Bureau, as well as party registration levels provided by the New Hampshire Secretary of State. The Granite State Panel is part of an effort by the University of New Hampshire Survey Center to investigate new ways of gathering and understanding the opinion of New Hampshire residents. The Granite State Panel is a probability-based web panel recruited from randomly-selected landline and cell phone numbers across New Hampshire. Individual surveys are completed by panel members by computer, tablet, or smart phone.
For more information about the methodology used in the Granite State Panel, please visit our website or contact Dr. Andrew Smith at (603) 862-2226 or by email at andrew.smith@unh.edu.
For more information about the methodology used in the Dartmouth-UNH NH COVID-19 Study, contact Dr. Tracy Keirns at (603) 862-1060 or by email at tracy.keirns@unh.edu.
Baseline – Pre-Diary Survey
Respondents for the study were first recruited using the Granite State Panel. At the time of this study the panel consisted of 3,273 members. On Friday, April 3, 2020, panel members were invited to participate in a community symptom surveillance survey. Invited participants were told they would complete three longer, more detailed surveys over 2 months, the first being a monthly baseline survey where a household roster was built; the next would occur at the 4-week mark and another at the 8-week mark. For the baseline survey, 1,105 Panel members completed or partially completed the survey online. This represents a 34% response rate to this initial survey.
Daily Diary Surveys
Granite State Panel members who provided enough information to enroll in the study, whether or not they completed the full pre-diary survey (N=1,105), received an email once a day asking them to submit their daily diary responses. The daily diary survey was a much shorter instrument comprised of a single question asking if any member of the respondent’s household was experiencing any symptoms since their last response. Only respondents who answered “yes” were asked to indicate the symptoms and behaviors of those with reported symptoms. This daily survey took less than 1 minute for most and about 1-2 minutes for those with symptoms. After 8 weeks of daily diary surveys, only 5% had left the study and the response rate was robust at a daily average of 73%.
Monthly Surveys
The monthly surveys were designed to collect updated and new information about household behaviors and attitudes concerning COVID-19. Study participants were invited to complete the first monthly follow-up survey online between May 1 and May 5, 2020. Of the 1,076 remaining study members, 830 completed the survey, yielding a response rate of 77%. The second monthly follow-up survey was conducted of the 1,051 remaining study members between June 1 and June 4, 2020, and 791 completed the survey, yielding a response rate of 75%.
Weekly Diary Surveys
After the second monthly survey, the remaining study participants (N=1,048) were invited to continue their participation over the summer. The weekly diary survey is a short instrument comprised of two questions asking if any member of the respondent’s household has experienced any symptoms in the past 7 days and a question on testing. Only respondents with a household member who has experienced symptoms are asked to indicate the symptoms and behaviors of those with reported symptoms. This weekly survey takes approximately 1 minute for most and about 1-2 minutes for those with symptoms. After nine weeks of weekly diary surveys, the average response rate remains robust at 74% with minimal attrition (total attrition since April is 8%).
