GratisAI civic · 4 min read · free tool
sample ballot stub

Lay a Ballot Measure's Arguments Side by Side

Ballot pamphlets put the arguments one after the other, which is exactly why the more confident one tends to win. This puts them opposite each other on the same question and marks whether each claim carries a source — because a claim with a citation and a claim without one are not the same kind of statement, and the layout usually hides that. Local only; nothing leaves the tab.

Scan report — what this replaces
Grid
Pro and con claims paired against the same question
local
What it replaces
Reading two confident pages sequentially
$0.00
Input
Claims from both sides, one per line
no card
Output
A paired grid with sources marked
recomputed

What you pay otherwise

RouteWhat you getCost
Voting on the pamphlet aloneWhichever side wrote better copya guess
Researching each claim yourselfAn evening per measurehours
Skipping the measureSomeone else decides it for youthe outcome
This pageBoth sides paired, sources marked$0

The workflow

  1. Paste each side's claims exactly as written

    Don't paraphrase — the wording is part of what you're evaluating. One claim per line. If a claim contains a number, a date, or a named study, keep it verbatim so you can check it later.

  2. Mark a claim as sourced only if it names its source

    Append ' [source]' to any claim that cites a study, an office, or a document. The tool doesn't judge whether the source is good; it marks whether one was offered, which is the first filter that matters.

  3. Read the unsourced tally before the arguments

    A side with eight confident claims and no sources is making a rhetorical case. A side with three claims and three sources is making an evidentiary one. Neither wins automatically, but you should know which you're reading.

  4. Pick one contested number and check it yourself

    You don't have to verify everything — that's the trap. Pick the single figure the decision turns on and check that one against the primary source. One verified number does more than ten unverified ones.

Worked example

Sample: measure on a transit levy
  Pro claims         5 (2 sourced)
  Con claims         4 (1 sourced)
  Contested figure   $42 annual cost per household
  Unsourced total    6 of 9 claims
  Uploaded           nothing — paired in the tab

Numbers above are sample data produced by the tool's own pre-loaded example, run in the browser — not averages or measured benchmarks.

What breaks this

Paraphrasing the claims loses the wording you're meant to be evaluating. Marking a claim sourced because it sounds official defeats the only filter here — mark it only if it names a study, office or document. Trying to verify everything is how people give up; check the one number the decision turns on.

Try it now

A transit levy with five pro and four con claims is loaded, six of them unsourced. Add a source tag to one and watch the tally move.

Open the comparison grid →

Go deeper: the printable stub

The grid is free forever. The $4 one-time tier adds:

See the paid tier inside the tool →

What this is not. This organises and marks the claims you paste. It does not verify them, judge their truth, or recommend how to vote — the sourced mark only records whether a source was named in the text you entered. Check any figure the decision turns on against its primary source.