New research reveals a paradox when it comes to public perception of polling practices: while many Americans suspect poll results may be manipulated to serve specific agendas, or even fabricated altogether, there remains a strong appetite for trustworthy data. This tension represents a critical opportunity for media organizations to boost public confidence by prioritizing accountability and ethical research standards.

The report, by AI-native quantitative research platform Outward Intelligence, sheds light on the main factors shaping perception of poll reliability, including concerns around bias, objectivity, and transparency. The findings, based on a September 2025 online survey of 775 U.S. adults balanced demographically, provide valuable insights for media leaders to utilize towards improving both polling and communications practices.

Audiences still value polling

First, some good news. While reservations abound, the majority of those studied value high quality data, feel represented at least somewhat in polling results, and view polling data with at least some level of trust.

• Despite widespread skepticism, 65% of participants express at least some level of confidence in poll accuracy.
• 85% of respondents view high quality of polling data as either extremely or very important.
• An overwhelming majority (86%) believe polls represent their views “a lot” or at least “somewhat.”

In addition, most participants (83%) believe it is very or extremely important for leaders of media organizations, government entities, and businesses to heed public opinion when making decisions. Therefore, organizations that demonstrate quality public opinion research as well as utilization of results in guiding their practices could have an edge going forward.

Audiences lose faith

However, this research brings to light a variety of doubts and concerns. Nearly half of those surveyed say they often or very often question the validity of polling data. Over one third think polling has declined in quality over time. Certain issues seem to provoke even greater levels of cynicism:

• Election stress. 70% of those surveyed feel that election-related polls are correct only occasionally, or not at all.
• Artificial Intelligence apprehension. 83% voice concern over how AI might affect trust in polling, highlighting growing awareness around data integrity issues.
• Bias suspicion. 87% believe that organizations “spin” data for their own purposes – either sometimes (53%) or very often (34%).
• Underrepresentation concerns. Almost 90% of participants believe that polls leave out or underrepresent certain groups of people.

Over a third of respondents believe that polls are completely lacking in transparency. Almost a quarter of respondents don’t even believe that polls are conducted with “real people actually taking a survey.” Other notable concerns include inadequate sample sizes and unclear or poorly communicated methodologies.

Polling possibilities for media leaders

These findings offer media leaders the opportunity to shore up their polling methods and communications to foster trust. More transparency about how data is gathered, the representative nature of the survey pool, and the responsible use of AI in data collection and analysis are all areas in which organizations can increase their oversight and improve trust.

Based on this data, actionable steps media leaders can take to boost public confidence in polling practices may include the following:

• Improve transparency in polling coverage. Disclose methodology such as sample size, margin of error, and how participants were selected. Acknowledge what polls can and cannot predict, especially around elections. Minimize framing poll results to fit narratives, as audiences are sensitive to perceived manipulation.
• Educate audiences on polling fundamentals. Provide explanations that demystify polling processes. Use interactive formats such as infographics or Q&A sessions to show how data is gathered and interpreted.
• Address AI concerns proactively. Be transparent about how AI tools are used in data analysis and results reporting. Highlight the human oversight integral to editorial decisions. Consider publishing AI ethics guidelines for polling and data use.
• Champion methodological rigor. Partner with reputable research firms that adhere to high standards and ensure that polls include diverse and representative samples.
• Foster interaction. Invite audience feedback on polling coverage through social media, newsletters, or live forums. Use skepticism as a springboard for dialogue, acknowledging doubts, and responding with clarity.
• Position polling as a tool, not absolute truth. Frame polls as snapshots of sentiment, not definitive forecasts. Balance polling data with qualitative insights, such as interviews with individuals or community discussions.

Finally, demonstrating how public opinion is being taken into consideration when making decisions can instill more confidence in audiences going forward. Media executives who embrace these strategies can strengthen their credibility and trustworthiness. In a landscape where skepticism is high but demand for quality data remains strong, publishers who lead with transparency and integrity can deepen audience loyalty and differentiate themselves in the market.