Doesn’t an online participation framework favor the tech-savvy?

Yes, an online-only version of My Vote Forward would tilt toward people with reliable devices, home internet, and comfort using digital tools. Pew Research Center’s 2026 data show that 97% of United States adults use the internet but only 78% have home broadband, falling to 57% among adults in households earning under $30,000 a year. The framework’s answer is to report who participated, never to present participants as the public, and to require offline and assisted channels in any real deployment; whether those channels close the gap is an open question.

The objection in its strongest form

The objection to My Vote Forward on digital skew is that every participation channel has a gatekeeper, and for an online tool the gatekeeper is technology. Near-universal internet use hides deep differences in how people use it. Someone who reaches the internet only through an older phone with a limited data plan, who shares a device with family, or who struggles with passwords and verification codes will participate less often and less easily than someone with a laptop and fast broadband. Those differences track income, age, disability, language, and rural residence. So an online system will systematically overrepresent the comfortable and connected, and its “public priorities” will be their priorities.

A sharper version says the skew compounds an existing one. People with more education and income already vote, contact officials, and attend meetings at higher rates. A new channel that is easiest for the same people widens the gap it claims to narrow, while giving their views a fresh appearance of democratic legitimacy.

Why the objection matters

Digital skew matters for My Vote Forward because the framework’s purpose is to make participation between elections more informative for institutions. If the people who participate are systematically different from the people who do not, the information is biased in a predictable direction. Institutions that act on it would be responding to an unrepresentative subset while believing they had heard from the public. See unequal participation and representation gap.

Defined terms

• Digital divide: differences in access to, and effective use of, digital technology across groups.
• Smartphone-only user: in Pew’s usage, someone who owns a smartphone but does not subscribe to home broadband.
• Self-selection: the process by which people choose whether to participate, so that participants differ from nonparticipants. See opt-in sample.
• Weighting: adjusting results so that groups count in proportion to their share of a target population. See weighting.
• Assisted participation: taking part with help from a trained person, at a library, community center, or by phone.

Evidence

Internet use is close to universal in the United States, but home broadband is not. Pew Research Center’s 2026 fact sheet, based on a survey of 5,511 adults conducted January 30 to June 18, 2026, reports that 97% of adults use the internet and 78% have home broadband.⁠¹ Among adults in households earning under $30,000, internet use is 91% but home broadband is 57%, and 29% are smartphone-only.⁠¹ Among adults 65 and older, 89% use the internet and 71% have home broadband; among rural adults, 95% use the internet and 72% have home broadband.⁠¹ Pew reports that “16% of U.S. adults are ‘smartphone-only’ internet users”.⁠¹ These numbers measure access, not skill or comfort, and they do not show how many people would participate.

Estonia offers evidence on who uses a mature online channel. In the 2023 parliamentary election, 313,514 people voted online.⁠² By age, the counts were 24,569 aged 18 to 24, 55,005 aged 25 to 34, 66,269 aged 35 to 44, 63,043 aged 45 to 54, 48,947 aged 55 to 64, 32,930 aged 65 to 74, and 22,751 aged 75 and over.⁠² These are raw counts, not rates; age groups differ in size, so the figures do not by themselves show which ages were more likely to vote online. They do show that online voting in Estonia, after nearly two decades of use since 2005, is not confined to the young.⁠² Estonia also has a national electronic identity system that most countries lack.

Evidence from New York City’s participatory budgeting shows that outreach design can change who participates. Celina Su reports that in District 8, East Harlem, the very poor “constituted 4% of voters in the 2009 city council elections” but rose to 22% of participatory budgeting voters, and that half of 2014 voters “had never worked with others on a community issue before”.⁠³ Lower-income and foreign-born residents were “more likely to learn about PB through word-of-mouth or targeted campaigns” than through online or government channels, and mobile or pop-up voting sites were more likely to reach lower-income and undocumented residents.⁠³ These are findings from one program with in-person assemblies and paper ballots, not an online tool.

How comparable systems have handled it

Programs that care about reach usually combine channels.

• In-person and paper alongside digital. New York City’s participatory budgeting used neighborhood assemblies, paper ballots in several languages, and pop-up voting sites; Su reports that in one district “over two thirds of distributed ballots were in languages other than English”.⁠³ See the New York City case.
• Digital as one channel among others. Estonia’s online voting is an alternative to polling places, not a replacement.⁠² See the Estonia case.
• Random selection with support. Citizens’ assemblies draw members by lot and pay or support them to attend, which removes self-selection by comfort. See citizens’ assembly.

Competing interpretations

One view holds that the digital divide is closing fast enough that the objection is temporary. Internet use near 97% suggests that access to some connection is nearly universal,⁠¹ and smartphone-first design can serve smartphone-only users.

A second view holds that the important divide is no longer access but confidence, time, and trust. People may have a phone and still not use it for civic tasks, especially if they distrust institutions or worry about data. On this view, outreach and assistance matter more than device access, which is consistent with the New York City findings.⁠³

A third view holds that any voluntary channel, online or offline, will be skewed, so the honest response is to stop treating voluntary participation as representative and to use random samples when representativeness matters. See what if most people never participate?.

What the framework answers

My Vote Forward’s responses are about reporting and design, not about technology fixing itself.

• Participants are not the public. Every release describes who participated and never presents results as public opinion. See representation.
• Report the skew. Releases compare participants with available population figures where group sizes allow, and say plainly where groups are underrepresented.
• Offline and assisted channels in any real deployment. Paper, phone, and in-person assisted participation are required design elements, not afterthoughts. See what about people without internet?.
• Low-bandwidth, accessible design. The demo is meant to work on modest phones and with assistive technology; see disability access and language access.
• No weighting without disclosure. If results are weighted, the method and variables are published, and unweighted results are shown alongside.

Where the objection still wins

The objection wins today. The framework exists as websites and a demo, with no offline channel, no assisted participation program, and no evidence about who would actually take part. Its design responses are stated intentions. Weighting can correct for measured characteristics like age and income but cannot correct for differences that are not measured, such as interest in politics, so a weighted online sample may still overrepresent the engaged. And outreach that works, like the targeted campaigns Su describes,⁠³ costs money and staff that the project does not have.

Unresolved questions

• Which offline channels are affordable enough to run alongside an online system without becoming token gestures?
• Should results from different channels be reported separately, so readers can see whether they differ?
• How much does verification friction, such as one-time codes or identity checks, deter participation by older or lower-income people?
• At what point, if any, should a voluntary system give way to random samples for questions where representativeness matters most?

Related

• Unequal participation
• Representation
• What about people without internet?
• Weighting
• Hard questions