The Center for Campaign Innovation returned to the field with its post-election survey in New Jersey and Virginia last month to better understand how voters consumed news and media, how they made their political decisions, and how they experienced campaign outreach during the 2025 gubernatorial elections.

The findings reveal a landscape where highly engaged partisans continue to receive the bulk of campaign communication, while the voters who remain genuinely persuadable are harder to reach and slower to decide. A meaningful share of the electorate consists of passive news consumers or outright news avoiders. These voters encounter fewer political messages, receive less direct contact, and yet represent roughly 20 percent of the electorate. They also remain the most open to persuasion.

Media habits deepen this divide. Streaming has become the dominant way voters access live content, with half of Virginians and two in five New Jersey voters reporting that they are streaming-only viewers. Nearly a third of Virginia voters and one in five New Jersey voters say they do not watch live TV at all. Their viewing time is spread across an average of four streaming services, a fragmentation that also appears in social media usage, where voters rely on a broad mix of platforms rather than a single network.

Decision timing further underscores these differences. While about half of voters in both states say they chose their preferred candidate by the summer, the late deciders are concentrated among the cord nevers and other low-engagement voters who receive less outreach and inhabit media environments that traditional campaigns struggle to penetrate.

News Habits

News engagement continues to be one of the clearest signals of political behavior in both states. Voters who actively seek out news form a stable, strongly Democratic bloc, while passive news consumers remain far more competitive. In both Virginia and New Jersey, roughly one in five voters sits in this passive category that does not avoid news entirely but does not track it closely either. These voters are less anchored in partisanship and hold more fluid attitudes toward the gubernatorial candidates.

This distinction is important because passive news consumers are also less likely to encounter political information on their own. They see fewer ads, consume less political content organically, and tend to make decisions later. Campaigns that assume high-engagement voters define the broader electorate risk missing the softer middle where persuasion is still possible. The passive segment represents one of the few remaining groups where message reach and frequency materially affect outcomes.

For campaigns, the implication is that news-seeking behavior is not just a media-planning variable but a target definition tool. In general elections, high-engagement voters require less information, while passive news voters require clearer, earlier, and more repeated messaging across channels they actually use. Treating news engagement as a behavioral marker helps campaigns allocate attention toward voters who are still movable rather than overserving those who already have fixed preferences.

Media Consumption

The media environment in both states continues to fragment in ways that break old campaign assumptions. Linear TV is no longer the dominant channel for most voters. In Virginia, roughly half of voters watch live content only through streaming services, and nearly a third do not watch live TV at all. New Jersey shows a similar pattern, with streaming outpacing traditional cable and satellite by a wide margin. This shift is not simply a matter of cord-cutting. It reflects a structural change in how voters access video content across services, devices, and formats.

Streaming has also become a crowded and complex space. Voters in both states report using an average of about four different streaming platforms, and most of them watch some level of ad-supported content. Even accounting for fully premium platforms, a larger share of voters are reachable through streaming ads than through linear television. The implication is that consistent video reach is no longer possible through a single buy. Campaigns must plan for distribution across multiple environments, each with its own inventory, targeting rules, and audience patterns.

Crucially, when it comes to advertising on these streaming platforms, 39% of Virginians and 42% of New Jerseyans said they watch a mix of both with and without ads. Roughly a third (30% in Virginia, 29% in New Jersey) say they watch only or mostly with ads. Twenty eight percent (28%) of Virginia voters and 26% of New Jersey voters said they watch only or mostly without advertising. 

Even accounting for those who watch streaming services like Netflix and Amazon Prime that do not allow political advertising and voters who watch ad-free streaming, more voters watch streaming services with ads than only watch traditional TV via antenna, cable, or satellite.

Fewer than one in three voters are only reachable via linear TV advertising, while a growing share are not reachable via streaming ads or are unreachable.

The differences between high-propensity and low-propensity voters make this shift even more significant. Lower-history voters are far more likely to have no access to linear TV, meaning a linear-heavy program systematically misses the audience campaigns most need to reach. These voters are not avoiding political information. They are simply not present in the media environments where most traditional campaigns spend the bulk of their budgets.

The unaffiliated and independent voters that campaigns are targeting are much more likely to be reached via streaming ads only or unreachable via advertising. Campaign outreach is still disproportionately focused on high‑propensity, live‑TV‑watching voters, even though the growth segments are off‑TV and lower‑history. High propensity (4 of 4) voters skew more “linear” and “live-TV reachable” while low propensity (0 or 1 of 4) voters are more likely to not watch any live TV at all. 

The real drop-off is among “no live TV” voters – the 29% of Virginia voters and 20% of New Jersey voters who don’t watch live TV at all. This group is less likely to recall any candidate contact and leans more toward low-propensity voter history meaning that media-off, low-history voters are the most likely to be untouched by campaigns. 

Spending patterns in the Virginia governor’s race illustrate how budget allocations can shape which voters campaigns are most likely to reach. Republicans, for example, invested more in cable than in Google, which includes YouTube, and more was spent on radio than on Meta, which includes Facebook and Instagram. These choices direct resources toward traditional media channels rather than the digital platforms that large segments of the electorate now use.

Social Media Usage

Social media remains a major part of the information environment for voters in both states, though the platforms they use vary widely. Meta’s properties continue to dominate overall reach. Facebook is the most widely used platform, followed closely by Instagram, and WhatsApp maintains a stable presence as a communication tool. This cluster gives Meta a broader footprint than any other single company in the social space.

YouTube also plays a central role in voters’ online behavior. Roughly one third of voters in both Virginia and New Jersey say they use it as a social platform, and many more encounter it through connected TV. Its reach extends across age groups and partisanship, making it one of the few digital platforms with strong penetration across demographic groups.

The data show continued growth and normalization of TikTok. About one quarter of voters in each state report using it, placing it ahead of X (formerly Twitter) and close to parity with platforms like Snapchat and Reddit. TikTok’s adoption among older voters remains low, but its presence among younger and middle-aged adults is now consistent.

These patterns underscore how fragmented the social landscape has become. Few voters rely on only one platform, and most participate in a mix of social and video environments. This diversification contributes to the broader shift toward video-first content and reinforces that voters encounter political information across many different channels rather than through a single dominant source.

Voter Contact

Voters in both states reported seeing more television advertising from the Democratic gubernatorial candidates than from the Republicans. This perception gap was large in Virginia and narrower in New Jersey, but the pattern was consistent. A small share of voters said they did not recall TV ads from either candidate.

Text messaging has overtaken phone calls as the dominant form of direct outreach. More than two thirds of voters in both states recall receiving political text messages, while fewer than one quarter recall receiving calls. Fundraising solicitations delivered through mail, text, email, and digital ads remain one of the most widely remembered forms of contact, especially among higher-engagement voters.

When accounting for the fact that only a portion of voters receive in-person outreach, Democrats held a clear advantage in Virginia on canvassing, postcards, and other forms of human contact. Republican advantages were more concentrated in relational organizing, particularly in rural areas, but at smaller scale. Early voters generally reported higher levels of contact than Election Day voters, and voters with heavier streaming or no-live-TV habits were less likely to recall any contact at all.

Decision Making

Most voters reported making their choice for governor well before the final stretch of the campaign. More than half in both states say they had settled on a candidate by the summer or earlier, indicating that early impressions continue to shape the race long before fall advertising ramps up. This early decision pattern is strongest among high-propensity and high-engagement voters, who tend to follow politics more closely and form stable preferences earlier in the cycle.

Recommendations

Our findings lend themselves to several practical recommendations for campaigners preparing for the 2026 midterms. 

Run A Dual-Track Campaign

High-engagement voters already seek out political information and usually have firm views. They do not require heavy persuasion and often do not benefit from it. Their needs are clarity, reminders, and reasons to stay aligned with the ticket. Campaigns should treat this group as an activation base, not a persuasion or turnout universe.

Campaigns should incorporate news engagement as a segmentation tool, not just a media insight. Active news voters can absorb fewer but more targeted touches, freeing resources to concentrate messaging on passive consumers who require more frequency. This approach helps campaigns correct the common imbalance where they over-communicate with voters whose views are fixed and under-communicate with the softer middle that is still movable. By treating news habits as a behavioral marker, campaigns can better align message cadence, channel mix, and resource allocation with how voters actually take in information.

The decision-making patterns in the data point toward a dual-track electorate. Reliable voters who actively follow news form their preferences early and move through the cycle with stable, predictable behavior. Disengaged and infrequent voters follow a very different timeline. They decide later, encounter far less political information, and overlap heavily with streaming-only and no-live-TV households. These two groups operate on separate clocks. Campaigns must recognize that early deciders and late deciders are not simply at different points in the same process but are, functionally, two distinct electorates that require different communication rhythms and media approaches.

Find And Reach Cord Nevers With New Media And Personal Outreach

When the summary of voter contact is layered with media and technology habits, a clear gap emerges. The voters who campaigns struggled the most to reach share a common profile. They are low-propensity individuals with little or no exposure to linear television. They spend most of their attention on streaming platforms and social media, especially Meta properties and YouTube, yet they do not watch live TV at all, which means they never encounter traditional TV ads. They also tend to make their decisions later in the cycle and are more likely to vote on Election Day if they participate at all.

This group represents the digitally native, off-TV, late-deciding slice of the electorate that continues to grow. When campaigns rely heavily on linear buys, they end up reinforcing an imbalance. Older and habitual voters are overserved, while younger, lower-history, urban, and often more diverse voters remain under-contacted. 

The path to reaching these “cord nevers” runs through social channels, streaming video, and relational organizing. These are the same mechanisms that powered the Trump coalition’s mobilization in 2024, and they remain the most efficient way to reach voters who never touch traditional broadcast.

Treat Streaming As The Primary Paid Media Platform

Campaigns face a media environment that is fundamentally different from the one that shaped earlier strategies. Linear-dominant plans no longer align with where persuadable or low-propensity voters actually spend their time. Only about a quarter of voters are reachable through linear alone, while roughly half watch exclusively through streaming, and close to a third do not watch live TV at all. The voters most resistant to linear are also the ones campaigns most need to persuade or mobilize. If they never encounter a TV ad, a linear-heavy budget fails by design.

Streaming should function as the primary video infrastructure. Nearly nine in ten voters use at least one streaming service, and on average they use about four. Most of these environments still contain political advertising, whether through fully ad-supported platforms or mixes of ad-supported and premium viewing, which gives campaigns ample opportunity to build reach if they place inventory correctly. YouTube stands out across every major voter segment. Roughly half of all voters use it either as a social platform or a connected TV destination, making it one of the few channels that cuts across age, partisanship, and vote method.

This media landscape demands a planning model that starts with streaming and social rather than treating them as secondary add-ons. Campaigns that fail to adapt will continue to miss the voters who are most open to persuasion and the voters who are most in need of mobilization.

Conclusion

Historical trends and political headwinds mean conservatives face a challenging landscape heading into the 2026 midterm elections. While the broader political environment will remain the strongest determinant of outcomes, effective campaigns will recognize what this data makes clear. The electorate is increasingly split between highly engaged partisans and a disengaged segment – “cord nevers” – that moves through the cycle on a completely different timeline.

Streaming and new media are no longer emerging trends or gradual transitions. They are the dominant realities shaping how voters encounter information, and their effects are already fully visible in the data. Campaigns that continue to rely on a media strategy built for an earlier era risk missing a growing share of the electorate entirely.

The voters who will decide close races in 2026 are the disengaged, low-propensity, “cord never” voters who do not respond to the traditional playbook. Campaigns that adapt to this bifurcated media environment will have a meaningful advantage. Those that ignore it will fail to reach the voters who matter most.

Methodology

The Center for Campaign Innovation commissioned 3D Strategic Research to conduct a post-election survey in New Jersey and Virginia.

The survey was conducted November 6-9, 2025 via a mix of live calls to landlines and cell phones, along with text messages inviting voters to take the survey via a secure web link. Respondents were matched to the correct individual in the household based on their gender and age. Party registration/modeled partisanship and vote history were pulled from the file for the correct respondent within the household.

A total of n=1,278 registered voters in New Jersey and n=1,318 registered voters in Virginia completed the survey. Among all registered voters, the data was weighted to reflect the registered voter population for each state based on the voter file and the 2024 U.S. Census American Community survey based on gender, age, education, race/ethnicity, region, party registration/modeled partisanship, and vote history. The data was also weighted to reflect the proportion of 2025 voters who turned out in the November 2025 election out of all registered voters and the 2025 voter sample was weighted to reflect the actual election results and demographics for gender, age, education, race/ethnicity, and region based on the exit polls.

This results in a total of n=589 respondents in New Jersey and n=677 respondents in Virginia who report casting a ballot in the November 2025 election. The margin of error for the 2025 voter survey data is +4.04% in New Jersey and +3.77% in Virginia in 95 out of 100 cases.