What is the American Trends Panel (ATP)?
The American Trends Panel is Pew Research Center’s primary source of survey data for U.S. public opinion research. It is a multimode, probability-based survey panel made up of roughly 10,000 adults who are selected at random from across the entire United States. All surveys are conducted in English and Spanish.
Since its creation in 2014, the ATP has grown and changed in many ways. This page provides an overview of the ATP, how it works, and how the Center’s researchers use it to study American public opinion.
Jump to:
Why we use a survey panel
A survey panel is made up of a group of people who have agreed to regularly take surveys.
A major advantage of survey panels is that they provide a relatively efficient way of collecting data compared with fresh samples. With a fresh sample, polling organizations make an initial contact, persuade respondents to take a survey and ask the necessary demographic information for weighting (more on weighting later). Researchers then repeat the process for every survey, selecting a new group of people every time. But for a panel, where people have joined and agreed to take additional surveys, most of this work does not need to be repeated.
Another benefit of using survey panels is that considerable information about panelists’ views and experiences can be accumulated over time. Because panelists may respond to multiple surveys on different topics, we can build a much richer portrait of the public than is feasible in a single survey interview, which is often limited in length to prevent respondent fatigue.
One possible downside of a panel is that participating in multiple surveys has the potential to make people behave differently compared with survey respondents from a fresh sample. Panelists may become more interested and engaged with the topics we ask about simply because we are asking about them – a phenomenon called “panel conditioning.” But most studies (including one of our own) suggest that it does not seriously affect the quality of data we obtain.
How people join the ATP
To create a scientific poll that accurately represents the views of all Americans, we select people at random to join the American Trends Panel. Only those selected can join – you can’t volunteer. But nearly every adult living in the U.S. – including you – has a chance of being selected.
When the ATP was created in 2014, the first group of panelists was invited to join the panel at the end of a large, national survey offered in English and Spanish that was conducted by randomly calling both landline and cellphone numbers. Two additional recruitments were conducted using the same method in 2015 and 2017, respectively.
In 2018, the ATP switched from telephone to address-based sampling (ABS) recruitment. This means that invitations are sent to a random sample of households selected from the U.S. Postal Service’s master list of all residential addresses, the Computerized Delivery Sequence File or CDS.
Randomization in sampling is carried through right down to the household level to maintain representativeness of sample. The adult (age 18 or older) with the next birthday in each selected household is asked to complete a survey, at the end of which they are invited to join the panel.
Every year we recruit a new national sample of adults to join the ATP. In some years, we design recruitments to boost the sample size for small or underrepresented groups. For example, larger-than-usual samples of Hispanic adults, Black adults and Asian adults were included in 2019, 2022 and 2023, respectively. This allows us to provide a much more detailed portrait of smaller groups in the population.
How panelists take ATP surveys
American Trends Panel panelists now take surveys in one of two ways: either online (on a computer, tablet or smartphone) or over the phone with a live interviewer. Starting in mid-2024, panelists were given a choice between taking their surveys online or over the phone when they join the panel. This ensures that U.S. adults who don’t use the internet are represented on the panel.
Each ATP survey is typically available to panelists for six to 12 days. Exactly how long depends on the survey. Panelists are invited to take surveys in a few different ways:
For online panelists …
- All online panelists receive an initial invitation via email.
- Online panelists who agree to receive SMS messages also get a text invitation.
- Those who don’t respond initially are sent email and text reminders throughout the week.
- Some online panelists are also mailed a reminder postcard.
For panelists who take surveys over the phone …
- All phone panelists are sent a prenotification postcard letting them know to expect a call about a new survey soon.
- A live interviewer calls these panelists to complete the survey.
- If panelists don’t pick up or they tell us they’re busy, the interviewer will try to call them on different days or at different times.
- Or, phone panelists can call in to take the survey at their own convenience.
The ATP didn’t always include the option to take surveys with an interviewer by phone. From 2016 to early 2024, the ATP operated as an online-only panel where all panelists took surveys on the web. People in households without internet service were provided with a data plan and tablets that they could use to take the surveys online.
From 2014 to 2016, the ATP conducted interviews with internet non-users by mailing them paper questionnaires.
Drawing samples for ATP surveys
One of the benefits of a large panel like the American Trends Panel is that there are more panelists than a typical survey requires. We typically send out one or two surveys each month, but not every panelist is asked to take every survey.
Rather than selecting all 10,000 panelists each time, many ATP surveys interview only a subset (e.g., 5,000) of the panelists. This reduces the cost of each survey and reduces the burden on individual panel members, sparing them from having to respond every time the Center fields a survey. It may also reduce the likelihood of panel conditioning.
Drawing subsamples (rather than interviewing everyone on the panel) also allows our researchers to tailor the samples to the needs of each survey. For example, a study with a focus on parenting and the transition to adulthood used two groups from the panel: parents of young adults, and young adults ages 18 to 34 who have a living parent they are in contact with.
Incentives
We offer all American Trends Panel members a small amount of money for each survey they take. These are known as incentive payments and have been shown to help increase the willingness to participate in surveys. We typically offer a range of incentive amounts, and the amount can also depend on the length of the survey.
How we quality-check ATP surveys
At Pew Research Center, it’s important to us that our data is accurate. So we’ve instituted checks at every stage of the survey process.
After a survey is programmed online, researchers test it to make sure that questions render on screen – mobile and computer – the way they ought to, and that respondents see all the questions that are intended for them. We also test on different devices and browsers.
Before sending the survey to the whole sample, we start slowly, sending survey invitations to about 60 ATP panelists who often answer surveys quickly. This gives us a chance to check for any errors in the data and fix them, if necessary, before researchers give the green light to send the survey to the rest of the sampled panelists.
Once we receive the final data, researchers check for questionable data quality – whether respondents left questions blank at very high rates or always selected the first or last answer presented. Fortunately, very few panelists act this way. But if they do, the answers from that panelist are removed from the study. If someone has a track record of acting poorly in our surveys, we may remove them from the panel.
Weighting
In a perfect world, everyone who we sample for a survey would respond. It would certainly make our surveys easier for us pollsters (and cheaper)! Unfortunately, we live in a world with declining response rates, where some people are more likely than others to respond to surveys.
In addition, we sometimes “oversample” people in certain groups of interest in order to have enough of them for reporting on separately. Because their share in the sample is larger than their share in the population, we need to make a statistical adjustment to the data so that, in terms of their impact on the results of the survey, their share of the sample is comparable to their share of the population.
Between the problems of nonresponse (skipping a survey) and the special oversampling that we sometimes do, the raw survey samples we obtain are imperfect models of our population of interest. But they can be adjusted to better match the population through the process of weighting, which aligns characteristics of the sample to known population parameters. To explain weighting in detail, we need to get into the technical side of statistics.
American Trends Panel data is weighted in a multistep process that accounts for multiple stages of sampling and nonresponse that occur at different points in the survey process. First, each panelist begins with a base weight that reflects their probability of selection into the panel. Then, the weights are calibrated to align with the population benchmarks. This corrects for nonresponse to recruitment surveys and panel attrition (i.e., leaving the ATP). To understand exactly which benchmarks were used for a particular survey, read the survey’s methodology statement.
Maintaining the ATP
Pew Research Center currently works with the survey vendor SSRS to recruit panelists, manage the panel and conduct the surveys. Previously, the panel was managed by Abt (2014-2017), GfK (2017-2018) and Ipsos (2018-2024).
Every year, we add new panelists to the panel. We do this for a few reasons:
- Some panelists stop taking our surveys, so we need to add new people to replace them and maintain our ideal panel size.
- New people become part of our target population. This includes new cohorts of 18-year-olds and people who move to the U.S.
- Comparing new panelists with those who have been in the panel a long time helps us determine whether being in the panel has actually led people to respond differently.
- We also sometimes want to add additional people from certain small or underrepresented groups.
Panelists also take an annual profile survey where they are given the opportunity to update certain demographic characteristics, such as their income, household size, level of education, etc. If a panelist does not participate in the annual profile survey, they are no longer invited to participate in ATP surveys.
Occasionally, we also retire members that are demographically overrepresented on the panel, such as those who are more educated. We do this to ensure representation on the panel that resembles the general population and to avoid having to weight certain panelists too heavily.
How we protect our panelists’ privacy
We make a promise to our panelists to protect their identity. Several processes are in place to make sure that Pew Research Center remains true to its word. Personally identifiable information (PII) such as a panelist’s name or address is maintained solely by the core panel administration team and is never released to the public. We also do not include detailed geographic information such as ZIP code in any survey data we release to the public. (We use additional statistical techniques to help avoid the identification of individuals with certain rare characteristics. For details on such techniques, see https://www.niss.org/research/confidentiality-data-swapping.)
Read more about our privacy policy.