{"id":225301,"date":"2025-05-08T09:49:13","date_gmt":"2025-05-08T13:49:13","guid":{"rendered":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/?p=225301"},"modified":"2025-11-03T10:31:57","modified_gmt":"2025-11-03T15:31:57","slug":"methodology-social-trust","status":"publish","type":"post","link":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/2025\/05\/08\/methodology-social-trust\/","title":{"rendered":"Methodology"},"content":{"rendered":"\n

Much of the data in this essay is drawn from the Religious Landscape Study (RLS). We also draw upon data collected from surveys conducted on Pew Research Center\u2019s American Trends Panel (ATP) and data from the General Social Survey (GSS). We walk through the details of these data sources below. We also discuss how we measure trust and the models and analyses used in the report.  <\/p>\n\n\n\n

The Religious Landscape Study<\/h3>\n\n\n\n

The RLS is a national cross-sectional survey conducted for the Center by NORC at the University of Chicago. It was conducted in English and Spanish from July 17, 2023, to March 4, 2024, among a nationally representative sample of 36,908 respondents.  <\/p>\n\n\n\n

To recruit people to take the survey, we used a method known as address-based sampling (ABS). This involved mailing invitation letters to randomly sampled addresses from the U.S. Postal Service\u2019s Computerized Delivery Sequence File. This approach gave nearly all U.S. adults a chance of being selected to participate in the survey.  <\/p>\n\n\n\n

People who received our invitation had the option of completing the survey online, on paper, or by calling a toll-free number and completing the survey by telephone with an interviewer. In total, 25,250 respondents participated online, 10,733 completed the survey on paper, and 925 participated by phone. <\/p>\n\n\n\n

The survey is designed to be representative of all 50 states and the District of Columbia. The data is weighted (using benchmarks from the U.S. Census Bureau and other sources) to be representative of the adult population of each state (and D.C.) and the overall U.S. adult population by gender, age, race, ethnicity, education and other categories. The survey\u2019s margin of error for results based on the full sample is plus or minus 0.8 percentage points. The response rate is 20%.  <\/p>\n\n\n\n

Sample sizes and margins of error  <\/h4>\n\n\n\n

The following tables show the unweighted sample sizes and the error attributable to sampling that would be expected at the 95% level of confidence for different groups in the survey. <\/p>\n\n\n\n

\"A<\/a><\/figure>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n

For more details, refer to the Religious Landscape Study\u2019s Methodology<\/a>. <\/p>\n\n\n\n

Panel recruitment <\/h4>\n\n\n\n

Since 2018, the ATP has used address-based sampling (ABS) for recruitment. A study cover letter and a pre-incentive are mailed to a stratified, random sample of households selected from the U.S. Postal Service\u2019s Computerized Delivery Sequence File. This Postal Service file has been estimated to cover 90% to 98% of the population.1<\/sup> Within each sampled household, the adult with the next birthday is selected to participate. Other details of the ABS recruitment protocol have changed over time but are available upon request.2<\/sup> Prior to 2018, the ATP was recruited using landline and cellphone random-digit-dial surveys administered in English and Spanish. <\/p>\n\n\n\n

A national sample of U.S. adults has been recruited to the ATP approximately once per year since 2014. In some years, the recruitment has included additional efforts (known as an \u201coversample\u201d) to improve the accuracy of data for underrepresented groups. For example, Hispanic adults, Black adults and Asian adults were oversampled in 2019, 2022 and 2023, respectively. <\/p>\n\n\n\n

Sample design <\/h4>\n\n\n\n

The overall target population for this survey was noninstitutionalized persons ages 18 and older living in the United States. All active panel members were invited to participate in this wave. <\/p>\n\n\n\n

Questionnaire development and testing <\/h4>\n\n\n\n

The questionnaire was developed by Pew Research Center in consultation with SSRS. The web program used for online respondents was rigorously tested on both PC and mobile devices by the SSRS project team and Pew Research Center researchers. The SSRS project team also populated test data that was analyzed in SPSS to ensure the logic and randomizations were working as intended before launching the survey.  <\/p>\n\n\n\n

Incentives <\/h4>\n\n\n\n

All respondents were offered a post-paid incentive for their participation. Respondents could choose to receive the post-paid incentive in the form of a check or gift code to Amazon.com, Target.com or Walmart.com.  Incentive amounts ranged from $5 to $20 depending on whether the respondent belongs to a part of the population that is harder or easier to reach. Differential incentive amounts were designed to increase panel survey participation among groups that traditionally have low survey response propensities.  <\/p>\n\n\n\n

Data collection protocol <\/h4>\n\n\n\n

The data collection field period for this survey was March 10 to March 16, 2025. Surveys were conducted via self-administered web survey or by live telephone interviewing.   <\/p>\n\n\n\n

For panelists who take surveys online:<\/strong>3<\/sup> Postcard notifications were mailed to a subset on March 10.4<\/sup> Survey invitations were sent out in two separate launches: soft launch and full launch. Sixty panelists were included in the soft launch, which began with an initial invitation sent on March 10. All remaining English- and Spanish-speaking sampled online panelists were included in the full launch and were sent an invitation on March 11. <\/p>\n\n\n\n

\"A<\/a><\/figure>\n\n\n\n

Panelists participating online were sent an email invitation and up to two email reminders if they did not respond to the survey. ATP panelists who consented to SMS messages were sent an SMS invitation with a link to the survey and up to two SMS reminders.  <\/p>\n\n\n\n

For panelists who take surveys over the phone with a live interviewer:<\/strong> Prenotification postcards were mailed on March 7. Soft launch took place on March 11 and involved dialing until a total of four interviews had been completed. All remaining English- and Spanish-speaking sampled phone panelists\u2019 numbers were dialed throughout the remaining field period. Panelists who take surveys via phone can receive up to six calls from trained SSRS interviewers. <\/p>\n\n\n\n

Data quality checks <\/h4>\n\n\n\n

To ensure high-quality data, Center researchers performed data quality checks to identify any respondents showing patterns of satisficing. This includes checking for whether respondents left questions blank at very high rates or always selected the first or last answer presented. As a result of this checking, four ATP respondents were removed from the survey dataset prior to weighting and analysis.  <\/p>\n\n\n\n

Weighting <\/h4>\n\n\n\n

The ATP data is weighted in a process that accounts for multiple stages of sampling and nonresponse that occur at different points in the panel survey process. First, each panelist begins with a base weight that reflects their probability of recruitment into the panel. These weights are then calibrated to align with the population benchmarks in the accompanying table to correct for nonresponse to recruitment surveys and panel attrition. If only a subsample of panelists was invited to participate in the wave, this weight is adjusted to account for any differential probabilities of selection. <\/p>\n\n\n\n

Among the panelists who completed the survey, this weight is then calibrated again to align with the population benchmarks identified in the accompanying table and trimmed at the 1st and 99th percentiles to reduce the loss in precision stemming from variance in the weights. Sampling errors and tests of statistical significance take into account the effect of weighting.<\/p>\n\n\n\n

\"A<\/a><\/figure>\n\n\n\n

The following table shows the unweighted sample sizes and the error attributable to sampling that would be expected at the 95% level of confidence for different groups in the survey.  <\/p>\n\n\n\n

\"A<\/a><\/figure>\n\n\n\n

Sample sizes and sampling errors for other subgroups are available upon request. In addition to sampling error, one should bear in mind that question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of opinion polls.<\/p>\n\n\n\n

Dispositions and response rates <\/h4>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n

Note:<\/em><\/strong> Additional data in this essay<\/a> is drawn from ATP Wave 162<\/a>. For more about this survey, refer to <\/em>its methodology<\/em><\/a>.<\/em> <\/p>\n\n\n\n

General Social Survey<\/h3>\n\n\n\n

We report on the trend in social trust from the General Social Survey (GSS). The GSS is a national survey of American adults conducted by NORC at the University of Chicago since 1972. It generally has been conducted every two years since the 1990s. Through 2018, GSS data was collected mostly through in-person interviews. Due to the COVID-19 pandemic, the GSS was not fielded in 2020. The 2021 GSS was conducted mainly online, supplemented by telephone. <\/p>\n\n\n\n

The change in data collection mode in the GSS after 2018 has raised questions about the comparability of trend measures. As a result, we have chosen not to report on the 2021 GSS.  <\/p>\n\n\n\n

Measuring trust<\/h3>\n\n\n\n
\"A<\/a><\/figure>\n\n\n\n

There is no universally accepted way to measure social trust, and different measures obtain different results. In this report we focus largely on one particular measure of social trust: whether \u201cmost people can be trusted,\u201d or \u201cyou can\u2019t be too careful in dealing with people.\u201d We use this measure because it has a long history, both in our surveys and those conducted by the General Social Survey<\/a>. It was also the question fielded as part of the largest single survey we conduct in the United States: the Religious Landscape Study<\/a>. <\/p>\n\n\n\n

But alternative measures find different levels of trust. If we measure trust by asking whether most people can be trusted or most people cannot be trusted,<\/em> we find substantially higher levels of trust. According to a February 2025 survey, 55% of Americans think most people can be trusted if the alternative is that most people cannot be trusted<\/em>, rather than simply that one needs to be careful<\/em>.  <\/p>\n\n\n\n

\"A<\/a><\/figure>\n\n\n\n

On this alternative measure, social trust has not fallen significantly over the past six years; 52% of Americans said most people could be trusted in 2018, and 55% said the same in February 2025. Notably, though, the patterns of who is trusting and who is not tend to be similar across both measures.<\/strong> Women, racial and ethnic minorities, younger people, those with less education and those with lower incomes tend to be less trusting across both measures.  <\/p>\n\n\n\n

A third measure of trust that we have also asked twice \u2013 in February and March 2018 and March 2025 \u2013 is similarly \u201cflat\u201d \u2013 meaning that trust did not increase or decrease. This question asks people, \u201cGenerally speaking, would you say you can trust most people a great deal, a fair amount, not too much, or not at all?\u201d Results in both surveys, eight years apart, find that about six-in-ten Americans say you can trust people a great deal or fair amount. Demographic differences between men and women, age groups, racial and ethnic groups and more tend to largely mirror those found when we ask about trust in the ways described above. <\/p>\n\n\n\n

Because social trust is a subjective concept, how people answer questions about it may depend on very specific features of the question. But even with the variation in results, questions about social trust are useful for comparing levels of trust across groups, between countries and across time. <\/p>\n\n\n\n

Modeling trust<\/h3>\n\n\n\n

In this essay, we explored factors related to how much interpersonal trust people have. To do this, we conducted a logistic regression, modeling who says \u201cmost people can be trusted\u201d as a function of people\u2019s demographic characteristics, as well as characteristics about their neighborhood. <\/p>\n\n\n\n

Trust, the dependent variable, is a dichotomous dependent variable, where 1 denotes someone who says \u201cmost people can be trusted\u201d and 0 means someone who says \u201cyou can\u2019t be too careful\u201d or did not respond to the question.  <\/p>\n\n\n\n

The independent variables include: <\/p>\n\n\n\n