NCPG designed the NGAGE 2.0 survey in consultation with the survey vendor Ipsos, the third-largest global market and opinion research group. The following points outline the research approach and methods used:
• An online survey was used to poll consumers throughout the continental United States, Alaska, and Hawaii. Survey respondents were drawn from panels recruited by Ipsos and supplemented with panels obtained from other members of the market research industry. Ipsos conducted all fieldwork, including sample sourcing, data collection and hosting the survey online.
• All surveys were collected between April 12 and April 29, 2021. The timing of the survey was deliberately selected to follow the NCAA basketball tournament (“March Madness” or the “Final Four”), an event that attracts billions of dollars in gambling handle.
• The survey was offered in English and Spanish.
• The survey was designed to allow for easy completion on any device type (e.g. computers, smartphone, and tablet).
A national sample of 2,000 adults was collected.
Quotas were established during fielding to ensure the sample was representative of key demographics such as gender, age, and Hispanic/non-Hispanic proportions. The data were further weighted post-fieldwork to ensure proper representation of differing levels of educational attainment. In some charts and tables, however, categories may not add up to 100 percent because not everyone answered all demographic questions—this most often happens with questions about household income. All demographic survey data were matched to data from the U.S. Census Bureau.
The NGAGE 1.0 (2018) survey was based on a national sample of 3,000 adults. While the smaller sample
size in 2021 had very little effect on overall results (the “credibility interval” as defined below only drops from +/- 2.0 percent to +/- 2.5 percent), it does mean that the sample sizes of certain demographic
subgroups drop below the level where statistical inference can be reliably made. This is most notably true for breakdowns of specific racial or ethnic groups, and as a result this report contains many fewer such comparisons than NGAGE 1.0.
Questions within each section of the survey were presented to each respondent in random order to eliminate any bias resulting from the order of the questions. For example, within the section on gambling participation, one participant might be asked about bingo first and sports betting second, while the next might be asked about roulette first, lottery second, with bingo being sixth and sports betting tenth.
For the most part, questions asked in 2021 were identical to those asked in 2018, and every question
asked in both surveys was worded identically. In 2021, however, new questions were asked about
financial investments, the effects of the COVID-19 pandemic, and other subjects which meant that
other questions had to be dropped so as not to increase the burden on respondents.
Ipsos measures the precision of their online surveys using Bayesian credibility intervals, an approach similar to classical confidence intervals or margin of error that attempts to account for the uncertain probability of an individual being included in an online panel.2 For this survey, the national sample (n=2,000) has a credibility interval of plus or minus 2.5 percentage points for all respondents. The credibility interval will be wider among subsets of the sample population, as follows:
• Sample of 1,000: credibility interval is +/- 3.5 percent
• Sample of 500: credibility interval is +/- 5.0 percent
• Sample of 200: credibility interval is +/- 7.9 percent
Given the size of the national sample, most comparisons will show differences that are statistically
significant. However, in many cases, these differences will be so small as to have little, if any, practical
significance. For this reason, levels of statistical significance will not be routinely cited in this report,
allowing a focus on differences large enough to be meaningful for public policy, public health, or
program development.
As with all sample surveys and polls regardless of sample methodology, this study is subject to other sources of error, including, but not limited to, coverage error and measurement error.
For more information about NCPG and the full range of programs and services in Advocacy, Awareness and Assistance in problem gambling and responsible gambling, please visit www.ncpgambling.org. NCPG is a not-for-profit, non-governmental organization funded by members and donors. NCPG receives no federal funding.