Surveys are a bedrock of how we understand Americans’ political views. And that is actually the established demand for ensuring precise results.

But a brand new study by the Pew Research Center indicates the difficulty could be beat – and it is a finding that could have huge consequences for the usage of Internet-based surveys going.

The report is not quite great news for Internet polling; substantial typical blunders were discovered by it across 8 of the 9 web survey firms examined, in fact. Such samples are called “non-chance” samples since respondents are drawn through pools of volunteers who regularly receive benefits or other incentives for taking surveys. Because not every individual in the public has an opportunity of being chosen for the survey (volunteers self-select), they cannot be statistically projected to the population within a conventional margin of sampling error.

Most strikingly, Pew found its Internet-based American Tendencies Panel wasn’t particularly precise on total measures, despite the fact respondents were initially recruited through conventional “likelihood-based” phone sample. Equally astonishing is that one of the volunteer net panels outperformed all others.

The study also included some sobering results for web surveys’ ability to correctly represent African American and Hispanic respondents. That finding is important, since one of the largest dreams of net- surveys that are tried is correctly representing a smaller demographic of the population, like a racial minority, that’s prohibitively expensive to interview using conventional techniques.